Nokobit 2004 Siste

The 11th Norwegian Conference on Information Systems
Stavanger, 29. November 1.December, 2004
Program Chair and Editor Knut H. R. Rolland

Norwegian University of Science and Technology
Organizing Chair Bjarte Ravndal

Stavanger University College
Welcome to NOKOBIT 2004!

First of all, I would like to welcome you all to the 11th Norwegian Conference on Information Systems (NOKOBIT 2004). According to tradition, this years conference is a mixture of newcomers to the IS field as well as more experienced IS researchers. For many, NOKOBIT thus serves as a first arena to explore the Norwegian IS community and to learn the tricks of the trade of IS research. In this regard, it is important that NOKOBIT continues to be a relatively safe and friendly arena, where the more experienced researchers take the role of advisors and mentors. A quick browse through this years proceedings reveals papers focusing on such diverse issues as empirical software engineering, IT standards, IT and organizational change, and mobile informatics. Regarding research methods and approaches one finds a similar diversity, ranging from traditional positivist studies to interpretive case studies and highly experimental approaches. Obviously, one can always discuss whether this wide range of different styles or genres of research is favourable. In a small community such as Norway, would it not be better to concentrate on a few issues and approaches to research? On the one hand, this diversity is good because it means that the Norwegian IS community can relate to larger communities in Scandinavia, Europe, and America. To some extent, diversity is also good for knowledge creation, since we often can get new ideas from chatting with researchers specializing in other fields. On the other hand, however, diversity can create problems related to communication and understanding. For example, it is not always easy for researchers using qualitative approaches to understand and evaluate results from a study done from a positivist point of view, and vice versa. I hope this, and many other interesting issues will be discussed during the conference, and that we together can cultivate a strong Norwegian IS community that has a voice also beyond the borders of Norway. A heartfelt thanks goes to Bjarte Ravndal who has done all the work in Stavanger and to the reviewers who did all the hard work for selecting the best papers to be published in the proceedings. This could not have been done without you!
And, last but not least, I would like to take the opportunity to wish you all both a pleasant and interesting conference!
Knut H. Rolland, Program Chair
Reviewers
Reviewers
Jennifer Blechar Bendik Bygstad Torgeir Dingsyr Gunnar Ellingsen Arild Jansen Tor J. Larsen Carl Erik Moe Judith Molka-Danielsen Glenn Munkvold Petter Nielsen Maung K. Sein Ingjerd Skogseid Thomas sterlie Margunn Aanestad
Contents
Contents
NOKOBIT PANEL
Diskusjonspanel om forskningsprogram i offentlig sektor
PAPER SESSION 1: Alternative Approaches to IS Development

Paper 1: "A workshop-oriented approach for defining electronic process guides A case study" by Dingsyr, Moe, Dyb and Conradi
10
Paper 2: "Skapende teater og utvikling av informasjonssystemer - Likheter og forskjeller i prosess og produkt" by Omland and Seljeseth
26
Paper 3: "Developing e-Government Portals: From Life-Events through Genres to Requirements" by Haraldsen, Stray, Pivrinta and Sein
44
PAPER SESSION 2: Information Systems Innovation and Adaptation

Paper 1: "What is the relationship between software development processes and IT based business innovation? An exploratory study in Norway" by Bygstad, Fagerstrm and stensen Paper 2: "The Role of Actual IS Utilization Level: Extending the ExpecationConfirmation Model of IS Continuance Intention" by Larsen and Sreb
72
93
PAPER SESSION 3: IS Research and Issues of Collaboration

Paper 1: "Collaborative Research Network: Case Study at Molde University College" by Molka-Danielsen Paper 2: "Mobile work - Mobile ICT Supporting Secondary Work" by Skattr, Hasvold, Berntzen, and Engvig
108 132
Contents
Paper 3: "ICT-related change in complex organisations: The role of infrastructural and regulatory constraints" by Aanestad and Boulus
150
PAPER SESSION 4: IT Infrastructures and Standards

Paper 1: "An empirical investigation of factors influencing success in broadband implementation" by Elevoll and Graumann Paper 2: "Local actors bridging the broadband divides" by Skogseid Paper 3: "The emergence of functional standards: Digital Rights Management standards in the mobile telecommunications industry" by Blechar
173 194 210
PAPER SESSION 5: IT/IS Learning and Education

Paper 1: "IKT skaper bde variasjon og lring - De frste tanker fra et prosjekt om elevers oppfatning av IKT som lringsverkty" by Godejord Paper 2: "Empirical Software Engineering Education" by Jaccheri and sterlie
229 242
NOKOBIT PANEL
NOKOBIT Panel
Diskusjonspanel om forskningsprogram i offentlig sektor p NOKOBIT 2004
Forskningsrdet har igangsatt et prosjekt VIOS - Verdiskapende innovasjon i offentlig sektor som tar sikte p mobilisere virksomheter i offentlig sektor til formulere og gjennomfre forskningsbaserte innovasjonsprosjekter og utviklingsoppgaver, i samarbeid med forskere og andre virksomheter, offentlige og private. Mlet med programmet er medverke til utvikle ein smartare og meir effektiv offentleg sektor, med betre kvalitet i tenestetilbodet og i forvaltninga. (s. 50, KRDs St.prp.nr. 1). Begrunnelsen for et slikt program er blant annet: Suksessfulle og robuste offentlige tjenester er avhengig av innovativ forskning og en mer koordinert satsing p FoU Det er behov for mle produktivitetsendringer for tjenester i det offentlige En hper derfor at VIOS skal bidra til at de offentlige blir mer innovativt og kan fungere som hjelpeapparat for fremme innovasjon i samfunnet. Videre m offentlig sektor gir mulighet til utnytte kompetanse- og kunnskapsressursene, samt de mulighetene som ligger i IKT-sttte for tjenesteproduksjon for tjenesteproduksjon. Programmet er n forankret i forslaget til Statsbudsjett under Kommunal og regionaldepartementet og Nrings- og handelsdepartementet. Mer om programmet p http://www.program.forskningsradet.no/vios/nyheter/visnyhet.html?id=5 I diskusjonspanelet nsker vi drfte det foreliggende programforslaget og blant annet hvilke typer prosjekter det kan vre aktuelt igangsette. Innleder vil vre Trond Knudsen, fra sekretariatet for programmet i Forskningsrdet. Andre
deltagere er fra forskningsmiljene og forvaltningen.
Panelansvarlig er Arild Jansen, Avdeling for forvaltningsinformatikk, UiO.
Alternative Approaches to Information Systems Development
SESSION 1 Alternative Approaches to Information Systems Development

This track presents research concerning different alternative methods for the development of information systems. Papers by Dingsyr et al. and Omland & Seljeseth both discuss methods that focus more on informal techniques, creativity and participation than traditional approaches. In the paper by Haraldsen et al. a specific approach for the development of "new" technologies like e-Government Portals is described.
Papers are presented in the following order:
1. "A workshop-oriented approach for defining electronic process guides - A case study" by Dingsyr, Moe, Dyb and Conradi 2. "Skapende teater og utvikling av informasjonssystemer - Likheter og forskjeller i prosess og produkt" by Omland and Seljeseth 3. "Developing e-Government Portals: From Life-Events through Genres to Requirements" by Haraldsen, Stray, Pivrinta and Sein
A Workshop-Oriented Approach for Defining Electronic Process Guides A Case Study

Torgeir Dingsyr1, Nils B. Moe1, Tore Dyb1 and Reidar Conradi2
1
SINTEF Information and Communication Technology, 2Norwegian University of Science and Technology. E-mail: {Torgeir.Dingsoyr, Nils.B.Moe, Tore.Dyba}@sintef.no; Reidar.Conradi@idi.ntnu.no
Abstract: We introduce electronic process guides, and discuss their role in software engineering projects. We then present existing methods for constructing electronic process guides by defining a set of common processes for a company. Different approaches from the software engineering and management science are presented. We then go on to propose a new way of dealing with process description in software engineering: using process workshops as a tool to reach consensus on work practice. The main reason for this is to get realistic descriptions with accurate detail as well as company commitment in an efficient manner. We describe our workshop-oriented method to define processes, which we have used in small software companies, and show examples of results.
Introduction
The way we develop and maintain software, or the software process, has long been regarded as crucial for software quality and productivity (Lehman & Belady, 1985). Most quality systems and software process improvement initiatives prescribe recommended processes for the developers and organization to follow. We therefore need to describe the relevant processes. In the 1990s there was a lot of work on defining formal and rather sophisticated process modeling languages, and associated tools for process execution and evolution. However, in spite of substantial efforts by academia and partly industry (Derniame et al., 1999) and creation of several conference series (Oquendo, 2003), the attitude was too formal to have a practical impact. In fact, most companies prefer rather simple process models - such as IDEF0 (National Institute of Standards and Technology, 1993), proprietary ad-hoc formalisms (e.g. the one used for Rational Unified Process), or even quasi-formal diagrams using a document-producing tool like Word (Becker-Kornstaedt et al., 2001). We can draw two lessons from this: formal modeling of processes may easily be overdone and is anyhow not enough to ensure developer motivation and hence process conformance. Second, automated enactment should be used with great care. To our knowledge, there are no success stories of enactment in an industrial context, apart from stable and mature domains like inspections, configuration management and testing. 10
Knowledge-work, like software development seems to be extremely difficult to support with enactment. A more practical approach to process work for companies, is to make such process descriptions available as electronic process guides (EPGs) on the company Intranet. Our recommendation is that the developers should be involved in such processes, both to work as recommended and to contribute to the process models. Otherwise, there will easily be a too large gap between the official process model and the actual process, leading to poor process conformance. This has happened in many organizations with elaborate quality systems, that are hardly respected by (or applicable for) the rank and file (Conradi & Dyb, 2001). A balance must therefore be found between discipline (obeying formal routines) and creativity (Glass, 1995) (actual development with much improvisation (Dyb, 2000)). This chapter reports on the experience with developing of an electronic process guide in a Norwegian medium-size company with rather strict requirements on their software processes. To increase process awareness by the developers, process workshops were run to collect experience that could lead to better process descriptions. This kind of participatory design has a strong Scandinavian work and research tradition. The issue we would like to discuss in this chapter is our suggested method for organizing process workshops. Interesting questions are which organizing elements make a wellworking process, and how the process can be designed to increase process guide usage in the future. We will describe how this was done in an example company, and discuss experiences from using this method, compare it to other possible approaches, and conclude with advice for organizing similar workshops. Now, we present electronic process guides in further detail and then describe important issues in employee participation which we build on in designing process workshops. The rest of the chapter is organized as follows: we introduce the research method, describe our workshop-oriented method to define software processes, which we have used in several small and medium-sized software companies. We present a case study of results from conducting process workshops in a satellite software company. We discuss findings from the case study in relation to existing theory, and finally conclude the chapter.
ELECTRONIC PROCESS GUIDES

Effectively disseminating process knowledge to process participants is crucial in any software process improvement effort. Process participants need effective guidance when process conformance is important, when a process changes frequently, and when new personnel join a project. Traditionally, this has been the realm of large organizations, and the way of describing and communicating processes has focused on printed standards and handbooks. However, such handbooks are more often seen as dust collectors than software process improvement facilitators, and especially so in small and medium-sized companies. For process guides to be useful, increasingly more software companies not only tailor their process guides to the specific needs of the company, but also make them available on the companys intranet. This way the traditional process handbook shifts from a bulky
11
pile of paper to a flexible on-line structure allowing easy access to all relevant information by means of an electronic process guide (Scott et al., 2002). A process guide can be seen as a structured, workflow-oriented, reference document for a particular process, and exists to support participants in carrying out the intended process (Kellner et al., 1998). Whether in the form of a printed handbook or an electronic version, a process guide should include the following basic elements: Activities: descriptions of how things are done, including an overview of the activities and details regarding each individual activity. Artifacts: details regarding the products created or modified by an activity, either as a final or intermediate result of the activity or as a temporary result created by one of the steps. Roles: details regarding the roles and agents involved in performing the activities. Tools and Techniques: details regarding the tools and techniques used to support or automate the performance of an activity. It can also include access to: Lessons learned and frequently asked questions for relevant topics. Access to informal resources such as human experts, discussion boards. Metrics tools for instant data collection and display. Project-level information, including task allocation, updated plans and personal time sheets. This means that an electronic process guide becomes a low-key, effective and flexible integrator and access channel towards a company experience base. A common way to describe processes is to describe process entry, tasks, verification and exit, where entry and exit are criteria needed to be fulfilled and the tasks describe activities, roles, artifacts, tools and techniques. This is commonly referred to as the ETVX model. Based on these elements, Kellner et al. (1998) have proposed a set of basic requirements and design principles for EPGs. Most importantly, an EPG should provide all the information elements and relationships contained in a good paper-based process guide. In addition, it should capitalize on diagrams, tables, and narrative to provide an effective user interface. Also, it should make extensive use of hyper-links to support flexible navigation and direct access to supporting information such as examples and templates. However, the potential of EPGs can only be realized when key capabilities are not only adopted, but also infused across the organization. This is complicated by the fact that there is considerable scepticism among software developers to learn from and adhere to prescribed process models, which are often perceived as overly structured or implying too much control (Conradi & Dyb, 2001). Therefore, we cannot expect such infusion of EPGs unless they are perceived as useful and easy to use in daily practice and consistent with the existing values, past experience, and needs of the software developers (Davis, 1989; Venkatesh & Davis, 2000).
EMPLOYEE PARTICIPATION
12
Conradi and Dyb (2001) showed the importance of employee participation during the development and introduction of formal software routines and that such routines must be supplemented by collaborative, social processes to promote effective infusion and organizational learning. This insight is not new. Employee participation and the way people are treated, has been noted to be a crucial factor in organizational management and development ever since the famous productivity studies at Western Electrics Hawthorne plant in the 1920s (Mayo, 1933; Mayo, 1945). The results of these studies started a revolution in management thinking, showing that even routine jobs can be improved if the workers are treated with respect. Since then, participation and involvement has been one of the most important foundations of organization development and change (Cummings & Worley, 2001; French & Bell, 1999). Participation is also one of the fundamental ideas of Total Quality Management (Crosby, 1979; Deming, 2000; Juran, 1992). Similarly, participation has always been a central goal and one of the pillars of organizational learning. For example, autonomous work groups (Trist, 1981), quality circles (Ishikawa, 1990), survey feedback (Baumgartel, 1959; Neff, 1966), quality of work life programs (Davis, 1977), search conferences (Emery & Purser, 1996), and cultural analysis (Denison & Spreitzer, 1991; Schein, 1992) are all predicated on the belief that increased participation will lead to better solutions and enhanced organizational problem-solving capability. What can be learned from these prior studies is that people tend to support what they have participated in creating, or to use Berger and Luckmanns (Berger & Luckmann, 1966) words: it is more likely that one will deviate from programmes set up for one by others than from programmes that one has helped establish oneself. An important aspect of participation is co-determination, i.e. the direct participation of workers in decisions about what should best be done at their own level. Within the context of software development, no one is more expert in the realities of a software companys business with respect to the day-to-day details of particular technologies, products, and markets than the software developers and their first-line managers are. Hence, it is important to involve all those who are part of the software process, and have decisions made regarding the development of EPGs by those who are closest to the problem. Consequently, and in order to get realistic descriptions with accurate detail as well as company commitment in an efficient manner, we involve all relevant employee groups in defining processes by using process workshops as a tool to reach consensus on work practice.
Method
The research reported in this chapter is from a large industrial research project, Software Process Improvement through Knowledge and Experience (SPIKE), where many companies cooperate with research institutions and universities in improvement activities. The collaboration is based on finding common improvement and learning goals, and working together to obtain the goals. The communication between contact persons in the companies and researchers (and data collection) is through meetings (minutes, 13
observation, and pictures), telephone calls, and e-mail communication. The researchers usually stay two-day visits in the participating companies in order to also get into the informal arena in the company, and not just collaborate in official meetings. This research method is a form of action research (Greenwood & Levin, 1998), where the researchers and participants from the companies had common goals: to improve software development, and learn from that experience. Together with the company, we discuss how improvement activities can be organized, and try it out in a cogenerative learning process. That the process is cogenerative means that both company insiders and researcher outsiders are able to reflect on actions performed. A communication arena is established with regular meetings between researchers and people from the quality department in the company. In this case, the process workshops were a solution suggested by researchers for a problem the quality department had: to improve documentation of the core processes of the company. We organized feedback-sessions after performing the process workshops for cogenerative learning. Potential problems with this type of research are that it can easily be biased, in that everyone is interested in reaching the goals that are set up. Thus, we do not know if the same results would be achieved with another set of researchers, with other people from the company, or with another company in the same situation. But action research is a way to get interaction with companies in a way that would not be possible if it was not so much in the company's interest. The case company was selected because they were putting much effort in software process improvement, and was thus a candidate for participation in the SPIKE project.
Defining Processes in a Medium-Size Software Company

We first describe the company where we carried out research, and then present our work with process workshops in this company.
A SATELLITE SOFTWARE COMPANY

Since the company was founded in 1984, they have delivered turnkey ground station systems, consultancy, feasibility studies, system engineering, training, and support. The company has been working with large development projects, both as a prime contractor and as a subcontractor. Customers range from universities to companies like Lockheed Martin and Alcatel to governmental institutions like the European Space Agency and the Norwegian Meteorological Institute. Most of the software systems that are developed are running on Unix, many on the Linux operating system. The company possesses a stable and highly skilled staff, many with masters degrees in computer science, mathematics or physics, and have what we can describe as an engineering culture. Approximately 60 people are working in the company, and the majority is working with software development. Projects are managed in accordance with quality routines fulfilling the European Space Agency PSS-05 standards and ISO 90012000.
14
The company had an extensive quality system, but the system was cumbersome to use because of the size and because it existed partly on file and partly on paper. As a part of being certified according to ISO 9001-2000, the company decided to document all main processes in the company. We worked with the company in defining the processes for software development.
DEFINING REQUIREMENTS FOR AN EPG

We started out with an initial workshop. The goal of this workshop was to define the different existing project types in the company, and to decide the format and most important requirements for the process guide. The company defined four main project types, and they chose the most common one as a starting point for the following workshops. Product development was the most common project type, and the size of this project type was typically 1000-4000 work hours. Other project types was customer controlled development projects, delivery projects (integration of existing components, and configuration), maintenance projects, and studies. Typical activities for product development projects were either customizing an existing product for a customer, developing a new system for a customer, or an internal project with a mixture of new development and integration of existing products. After the project types were defined and product development was chosen as a starting point, the most important requirements were defined. The process guide should provide: Description of tasks for the most important roles in a project Checklists for each main process Templates for all documents produced Descriptions of best practice Access to all tools needed in the project (e.g. a requirement and a bug track system). In addition to these functional requirements a few non-functional requirements were defined during the first workshop. The most important such requirements were that it should be: easy accessible, as simple as possible, and up to date.
DISCUSSING PROCESSES: THE PROCESS WORKSHOP

We ran a total of six process workshops focusing on different parts of the development process. The workshops involved people from the market and quality department as well as the development unit. In the first process workshop for product development, initiation was the one the company wanted to start with. The initiation process was defined to include offer, follow-up and blast off. We followed the same pattern for each workshop, which we describe below with examples of output from the first workshop. See (Ahonen et al., 2002), for a discussion of a similar group process technique. The workshops differed in length, but would usually last half a day. The researchers acted as moderators and secretaries. In addition to a meeting room, the workshop required a collection of stickers in different colors, and walls that were covered with paper, where we could attach stickers and draw figures. A digital camera was useful to document the
15
results of the workshop. We also found it useful to bring large process worksheets, based on the ETVX model: a sheet with boxes for input, activities, output, roles and related documents involved in the process (see Figure 2). We defined process(es) in six steps and five sub-steps as shown in Figure 1. As the initiation of projects is an interface between different parts of the organization, it was important to bring together people from marketing, quality assurance and the development department. We started the workshop by giving a 15-minute presentation of what we were going to do, and put a large sheet with a figure of the process worksheet (as in Figure 2) on the wall one for each process that would be discussed in the meeting. For each sub-process we wanted to define, offer, follow-up and blast-off, we went through the sub steps:
Decide on process(es) to define
Define sequence
Invite participants
Identify activities
Process workshop
Define input and output
Find related documents Delegate responsibility for implementation
Define roles
Role-based reading of resulting process
Implement the process in EPG
Figure 1: Steps to define a process in a workshop Identified activities. We brainstormed on the main activities of the process by using the KJ process (Scupin, 1997) (after Japanese ethnologist Jiro Kawakita) and documented the result. The KJ is a creative group technique to organize and find relations between seemingly unrelated ideas. We did this as follows: We gave each participant a set of yellow stickers and a thick pen. We asked them to write suggestions for activities on each yellow sticker in large letters. People got time to document 5-10 ideas. 16
We asked each participant to present her suggestions: attach each sticker on a wall, and describe the activity. No-one was allowed to criticize or discuss the ideas at this point. Grouped the suggestions: the participants came forward to the wall and organized the yellow stickers into groups. We asked them to state why they chose to move the stickers. Formulated headings: we found new suitable headers that described the stickers in each group. The headings were formulated to make sense to people who have not participated in the workshop. We documented the diagram on the wall with groups and supporting activities on stickers. During this work, several interesting discussions came up, and several important problems and misunderstandings were solved. Especially marketing and project managers had different views on initiation, but were able to agree on a common process during the workshop. Because we wanted to get through three sub-processes in half a day, we used time boxing which limited discussion. However, we were able to produce an extensive material in the time slot for each sub-process. The main activities identified in this step for the blast-off sub-process were: Appoint project manager Organize Handover meeting First project analysis Allocate resources Prepare for kick-off meeting Internal kick-off. Defined the sequence of the activities. We took the activities from the previous phase, made a sticker for each. Then, we placed them on the activities-field of the process worksheet, where time goes from left to the right. We found a suitable workflow between the activities. Defined input and output. We found documents or artifacts that must be available to start the sub-process, and which documents that mark the end of the sub-process. We used stickers with other colors than for the activities to mark input and output, and attached them on the process worksheet on the wall together with the activities. Conditions that must be satisfied to begin or exit the sub-process can be described in checklists.
17
Figure 2: A process worksheet with input, activities, output, roles and related documents defined. Defined roles. We brainstormed on which roles should contribute in each activity and found the following roles for the blast off phase: project manager, quality assurance, development responsible, technical responsible, product committee, bid manager, purchasing manager, logistics expert. Related documents. We identified documents that either already existed in the company, or new documents that would be helpful in carrying out the activities. Such documents were templates, checklists and good examples of input or output documents. The researchers documented the process workshop by taking notes of stickers in different categories, and by the use of pictures (as in Figure 3).
18
Figure 3: A workshop participant adds an activity to a process worksheet. We found it helpful to ask the people who participated in the process workshop to read the result and comment on it (see (Shull et al., 2000) for an example of such a technique in requirements inspection). We assigned the most typical roles that were involved in the processes to people and asked them to find if there was information that was lacking or irrelevant for this role in the description. This reading resulted in a number of modifications and clarifications on the process description. Finally, two people in the company were responsible for making a draft process guide, based on the overall description of the processes which are developed in the workshop. Each activity was then described in much more detail than what appeared in the workshop minutes the participants gave feedback on these before the processes were implemented in the process guide, as shown in Figure 4.
19
Figure 4: A screenshot of a part of the resulting electronic process guide on the company Intranet.
FOLLOWING WORK
After the first version of initiation was accepted and implemented in the process guide, the company was ready for the next workshop. After initiation it was natural to focus on product development. This process was defined to include the sub-processes: specification, elaboration, component construction, and system integration. Also for these processes, input, activities, output, roles, and related documents involved in the process were defined. After the two main processes, product development and initiation were defined, the company was ready to release the first version of the process guide. The enthusiasm was high after the workshops. It was therefore important to give the workshop participants feedback through a running system even if it was not complete. Waiting for the perfect and complete process guide would take too long and could kill the enthusiasm. While implementing and releasing the process guide, the company conducted process workshops on project closure, product release, delivery and competence registration. These seven first workshops had from 4-6 participants (researchers not included), and 20 persons (1/3 of the employees) from the company participated in one or more workshops. The workshops lasted from 2 hours (workshop on format and requirements of the process guide) to 6.5 hours. The participants did not need to prepare themselves before the workshops. The company used: 168 work hours for seven workshops 40 work hours on supplementary work after workshops 208 work hours for implementing the process guide 223 work hours for implementing project tracking tools in addition to the guide 20
38 work hours on documentation. The total cost of developing the first version of the process guide was 1049 work hours. The two researchers used 10 work hours each including preparation and supplementary work for each workshop.
Discussion
In this section, we would like to discuss our experience with conducting process workshops, and elaborate on strengths and weaknesses of applying such an approach. We believe that participation and involvement is critical to achieve improvement in any organization, and see the process workshop as an arena which is open for many of the employees to take part in. Further, we see the process workshops as an arena where representatives from various departments can meet and discuss which will give participants a broader view of how work is conducted in the organization. Finally, we see the process workshop as an arena for collective reflection and learning, where employees can share experience on how they usually solve tasks, and discuss efforts to help them solve the tasks more efficiently. It is not the intention in this paper to prove that process workshops are more suitable than other techniques in eliciting process descriptions. We do not yet have sufficient experience with the resulting process descriptions to investigate that issue. We will rather point out some elements that we noted when conducting the workshops which can be useful for other approaches in the future. However, we note the findings of Ahonen et al. (2002), who report that a similar workshop-technique for modeling software processes both increased the knowledge of the real process and identified points of improvement. First, we noted that the people who participated in the workshops were contributing with many new perspectives on the processes. For example, one of the people in the quality department in the company had already made a draft version of a process description before organizing a workshop. He found that the workshop produced a number of activities, roles, and also input and output-documents that he did not think of himself. The brainstorming sessions with yellow stickers worked well to get all participants involved in the process. We have experienced that software developers often can be quite introvert people; and the workshops gave them the opportunity to participate more actively in discussions. Using the stickers gives each participant approximately the same time to present experience. The workshop provided an arena for cross-functional discussion in the company, and there were several discussions between for example the market and software development departments on how issues were to be handled. We think many clarifications were made that would not have appeared if it had not been for these workshops. We were satisfied with using the simplified version of the ETVX process worksheets in the brainstorming sessions. Using the worksheet gave an easily understandable visual presentation of the results and the connection between different elements of the result. None of the participants in the workshops we organized said they found the ETVX sheets inappropriate. During the sessions we used time boxing in order to generate ideas for all sub-processes and sub-process elements. Because of limited time, we had to stop some discussions to 21
move to the next process element. In an organizational learning sense, one could argue that we should have had more space for free dialogue, which would elicit more of the tacit knowledge from the people involved. However, using time boxing generated a flow in the workshop. We had the impression that none of the participants got bored or stopped engaging in discussions because the topic was irrelevant, which might have happened if we had allowed for more time. Another aspect that gave a lot of feedback on the results was the role-based reading of the results of the workshop. Assigning roles to people was a good tool in discovering inconsistencies, for example that a role was missing in one sub-process description or that a document relevant to a role appeared in one sub-process as output and not as input in another sub-process later. It also gave us general feedback of the wording of the names of roles, documents and activities. We claim that the workshops provided an arena for participation which was consistent with existing values, past experience and also with the needs of the company employees. Further, the process workshops were fairly efficient in terms of resources spent to design the process guide. We do not think using other approaches such as process experts conducting interviews or purchasing existing canned processes would have come out cheaper for the company. Other approaches would also probably require more tailoring, and would not involve the employees to such a large degree. It would also put less focus on the learning aspects through reflection on own practice, which are evident in groupwork. On the basis of the workshops conducted, we can recommend other companies wanting to develop electronic process guides to organize a set of workshops using the brainstorming techniques, the ETVX sheets and the role-based review.
Conclusion and Further Work

From the previous discussion of how process workshops worked in the case study of the satellite software company we can conclude: Process workshops conducted in the way described provides an open forum for reflection and learning about own work methods. Process workshops are an efficient method for discussing and agreeing on a set of work processes. Further work in this area will be to follow the acceptance, usage and impact of this process guide in the satellite company. We will study the usage of the process guide over time using a modified version of the Technology Acceptance Model as well as studying usage logs and performing qualitative semi-structured interviews. We intend to study the relation between involvement in process workshops and actual use of the process guide over time.
Acknowledgements
This work was conducted as a part of the SPIKE research project, supported by the Research Council of Norway. We are very grateful to our contact persons in the satellite
22
company for providing a stimulating environment in the project and to the participants in the process workshops for a positive attitude towards new work methods.
Bibliography
Torgeir Dingsyr wrote his doctoral thesis on Knowledge Management in MediumSized Software Consulting Companites at the software engineering group at the Department of Computer and Information Science, Norwegian University of Science and Technology. He has published papers on knowledge management in software engineering, software process improvement, case-based reasoning and on software engineering education. In 2001, he spent half a year as a guest researcher at the Fraunhofer Institute for Experimental Software Engineering in Kaiserslautern, Germany. He is now employed at the Sintef ICT research foundation, working on software process improvement projects. Nils Brede Moe is a research scientist at SINTEF ICT in Trondheim, Norway. He joined the department in 1998. He holds a M.Sc. degree in Informatics from the Norwegian University of Science and Technology in 1998. He has published papers on health informatics (telemedicine and homecare), user-cantered design and process improvement. Tore Dyb received his M.Sc. degree in Computer Science and Telematics from the Norwegian Institute of Technology in 1986 and his Ph.D. degree in Computer Science from the Norwegian Univer-sity of Science and Technology in 2001. He has worked as a consultant and a researcher both in Norway and in Saudi Arabia. His research interests include empirical software engineering, software process improvement, knowledge management, and organizational learning. Dr. Dyb is currently a senior scientist at SINTEF ICT and a visiting research scientist at Department of Software Engineering at the SIMULA Research Laboratory. Reidar Conradi received his MS and his PhD in computer science from the Norwegian University of Science and Technology in Trondheim (NTNU), and has been there since 1975. He is now a professor at the Department of Computer and Information Science at NTNU. Conradi was a visiting scientist at Fraunhofer Center for Experimental Software Engineering in Maryland and at Politecnico di Milano in 1999/2000. His interests are process modeling, software process improvement, knowledge and software engineering databases, versioning, object-orientation, component-based development, and programming languages.
References
Ahonen, J. J., Forsell, M., & Taskinen, S.-K. (2002). A Modest but Practical Software Process Modeling Technique for Software Process Improvement. Software Process Improvement and Practice, 7(1), 33-44. Baumgartel, H. (1959). Using Employee Questionnaire Results for Improving Organizations: The Survey 'Feedback' Experiment. Kansas Business Review, 12, 2-6. Becker-Kornstaedt, U., Neu, H., & Hirche, G. (2001). Software Process Technology Transfer: Using a Formal Process Notation to Capture a Software Process in Industry. In
23
V. Ambriola (Ed.), Proceedings from the Eight European Workshop on Software Process Technology (EWSPT2001) (pp. 63-76): Springer LNCS 2077. Berger, P. L., & Luckmann, T. (1966). The Social Construction of Reality: A Treatise in the Sociology of Knowledge. Harmondsworth: Penguin Books. Conradi, R., & Dyb, T. (2001). An Empirical Study on the Utility of Formal Routines to Transfer Knowledge and Experience. In V. Gruhn (Ed.), Proceedings of the European Software Engineering Conference 2001 (ESEC'2001) (pp. 268-276): ACM/IEEE CS Press. Crosby, P. B. (1979). Quality is Free: The Art of Making Quality Certain. New York: McGraw-Hill. Cummings, T. G., & Worley, C. G. (2001). Organization Development and Change. Cincinnati, Ohio: South-Western College Publishing. Davis, F. (1989). Perceived Usefulness, Perceived Ease of Use, and User Acceptance of Information Technology. MIS Quarterly, 13(3), 318-339. Davis, L. (1977). Enhancing the Quality of Work Life: Developments in the United States. International Labour Review, 116 (July-August), 53-65. Deming, E. W. (2000). Out of the Crisis. Cambridge, Massachusetts: The MIT Press (first published in 1982 by MIT Center for Advanced Educational Services). Denison, D., & Spreitzer, G. (1991). Organizational Culture and Organizational Development: A Competing Values Approach. In R. Woodman & W. Posmore (Eds.), Research in Organizational Change and Development (Vol. 5, pp. 1-22.). Greenwich, Connecticut: JAI Press. Derniame, J.-C., Kaba, B. A., & Wastell, D. (1999). Software Process: Principles, Methodology, and Technology. Springer Verlag LNCS 1500. Dyb, T. (2000). Improvisation in Small Software Organizations. IEEE Software, 17 (September/October), 82-87. Emery, M., & Purser, R. E. (1996). The Search Conference. San Francisco: Jossey-Bass. French, W. L., & Bell, C. H. J. (1999). Organization Development: Behavioral Science Interventions for Organization Improvement. Upper Saddle River, New Jersey: PrenticeHall. Glass, R. L. (1995). Software Creativity. Prentice Hall. Greenwood, D. J., & Levin, M. (1998). Introduction to Action Research. Sage Publications. Ishikawa, K. (1990). Introduction to Quality Control. London: Chapman & Hall. Juran, J. M. (1992). Juran on Quality by Design: The New Steps for Planning Quality into Goods and Services. New York: Free Press. Kellner, M. I., Becker-Kornstaedt, U., Riddle, W. E., Tomal, J., & Verlag, M. (1998). Process Guides: Effective Guidance for Process Participants. Proceedings of the 5th International Conference on the Software Process: Computer Supported Organizational Work, Lisle, Illinois, USA.
24
Lehman, M. M., & Belady, L. A. (1985). Program Evolution - Processes of Software Change. Academic Press. Mayo, E. (1933). The Human Problems of an Industrial Civilization. Boston: Harvard University Press. Mayo, E. (1945). The Social Problems of an Industrial Civilization. Boston: Harvard University Press. National Institute of Standards and Technology. (1993). The Standard for Integration Definition for Function Modelling (IDEF0). Neff, F. W. (1966). Survey Research: A Tool for Problem Diagnosis and Improvement in Organizations. In A. W. Gouldner & S. M. Miller (Eds.), Applied Sociology (pp. 23-38). New York: Free Press. Oquendo, F. (2003). Software Process Technology. Proceedings of the Ninth International Workshop, EWSPT2003, Helsinki, Finland. Schein, E. H. (1992). Organizational Culture and Leadership. San Francisco: JosseyBass. Scott, L., Carvalho, L., Jeffery, R., DAmbra, J., & Becker-Koernstaedt, U. (2002). Understanding the Use of an Electronic Process Guide. Information and Software Technology, 44, 601-616. Scupin, R. (1997). The KJ Method: A Technique for Analyzing Data Derived from Japanese Ethnology. Human Organization, 56(2), 233-237. Shull, F., Rus, I., & Basili, V. R. (2000). How Perspective-Based Reading Can Improve Requirements Inspections. IEEE Computer, 33(7), 73-79. Trist, E. (1981). The Evolution of Socio-Technical Systems: A Conceptual Framework and an Action Research Program. Toronto, Ontario: Ontario Quality of Working Life Center. Venkatesh, V., & Davis, F. (2000). A Theoretical Extension of the Technology Acceptance Model: Four Longitudinal Field Studies. Management Science, 46(2), 186204.
25
Skapende teater og utvikling av informasjonssystemer Likheter og forskjeller i prosess og produkt

Hans Olav Omland Jorunn Seljeseth
Hgskolen i Agder Serviceboks 422, 4604 Kristiansand Email: Hans.O.Omland@hia.no, Jorunn.Seljeseth@hia.no
Sammendrag
Skapende teater og systemutvikling kan synes som svrt forskjellige fagomrder. Produktene fra de to prosessene henvender seg til forskjellige brukergrupper og kan tilsynelatende ha forskjellig hensikt og bruk. Likevel er det to begreper som finnes i begge fagomrdene, nemlig prosess og produkt. I denne artikkelen sammenligner vi bde prosessene som brukes og produktene som utvikles innen de nevnte fagomrdene. Vi forsker f tak p grunnleggende temaer som ligger under prosessene for sammenligne dem. Vi finner mange likheter med det som foregr i utviklingen av produktene. Produktene, en teateropptreden og et informasjonssystem kan umiddelbart synes svrt forskjellige. Vr diskusjon viser at det ogs her er mange likheter. Konklusjonene i denne artikkelen kan vre nyttige i utdanning av studenter innen begge fagomrdene og at fagomrdene kan bde lre av hverandre og samarbeide for gjre utdanningen bedre.
Innledning
Bde innen skapende teater og utvikling av informasjonssystemer (IS-systemer) foregr det en prosess som leder til et produkt. Selv om skapende teater og utvikling av ISsystemer kan virke veldig forskjellig p overflaten brukes begrepene produkt og prosess
26
innen begge fagmiljene. Denne tilsynelatende store forskjellen mellom de to fagfeltene koblet med bruk av de samme grunnleggende begreper gjr at vi mener det er interessant se p hva som legges i begrepene og f fram likheter og forskjeller i dem innen de to fagfeltene. En slik klargjring vil kunne ha betydning for fagfeltene generelt. I denne artikkelen vil vi anvende resultatene fra den generelle sammenligningen i en undervisnings/lringssituasjon. Vi mener at en slik sammenligning vil bevisstgjre oss som undervisere og derigjennom kan vi lage et bedre undervisningsmilj for studentene innen de to fagomrdene. I utdanning av studenter er det viktig f fram disse to elementene og forskjellene p dem. Vi tar utgangspunkt i flgende problemstilling: Hva kan undervisningsmiljene innen skapende teater og utvikling av informasjonssystemer lre av hverandre i utforming av bde prosess og produkt? Vi sker svar p problemstillingen ved starte med en generell diskusjon og sammenligning av prosess og produkt innen de to fagfeltene. Deretter diskuterer vi konsekvensene for oppbygging av lringsmiljene innen de to fagfeltene. Artikkelen fortsetter med en litteraturgjennomgang. Deretter sammenligner vi utviklingsprosessene og produktene for skapende teater og systemutvikling. S flger et kapittel der vi diskuterer likheter og forskjeller mellom prosessene og produktene. Vi vil ogs diskutere mulige konsekvenser for lringsmiljet i de to fagene. Artikkelen avsluttes med en konklusjon.
Litteraturgjennomgang
Dramafagomrdet har hatt en utvikling hvor en kan se vektleggingen p prosess og produkt som et gjennomgende tema. Pendelen har svingt fra produktets viktighet og betydning uten tanke p prosess til total vektleggingen av prosess hvor produkt er utelukket. Faget har to grunnpilarer: Skoleteatertradisjonen og reformpedagogikken. Braanaas (1999) skriver at skoleteatertradisjonen vektlegger produktet mens reformpedagogikken vektlegger prosesstanken og utviklingen av naturlige og medfdte uttrykksevner. Slade (1954) og Way (1967) vektlegger prosessen. Heatcote og Bolton (1994) utarbeidet en metode for lre som blir kalt prosessdrama hvor vektleggingen er drama som metode for lre andre fag eller mellommenneskelige forhold. Hovedintensjonen med denne metoden er lre av vre i prosess, og produkt er utelukket. Szatkowski (1991) skriver om teaterprven som modell for dramafaget. Dette var et av anslagene i retning av utvikle betydningen av produkt i faget. Han viser til Ross (1978) sin modell for skapende arbeid p teateromrdet. Ross konkretiserer den generelle modellen for all skapende virksomhet. Han hevder at det estetiske innebrer en egen erkjennelsesform. For Ross er skapende virksomhet og lek p det nrmeste knyttet sammen. For kunne utvikle den kreative prosess m disse forutsetningene vre tilstede: Teatergruppens impuls til en forestilling bearbeides som flger for til slutt bli et teaterprodukt: 27
sanser (materiale til ideer: observasjon, lytting, fornemme bevegelser) uttrykksmidlene (kropp, stemme, teaterkonvensjonene) hndverk (ferdigheter, teknikker, teatertekst, scenografi, rombevissthet, lys) bildedannelse og fantasi (assosiasjoner, kontraster)
Pendelen har de siste rene klart svingt i retning av se p prosess og produkt som likeverdige og med klare rtter i teaterkunsten. Vektleggingen gr i retning av kunne utvikle evne til uttrykke seg gjennom form. Samtidig fr prosessen og produktet stor betydning som erkjennelsesform. Det er mange forskjellige systemutviklingsmetoder tilgjengelig i dag og de differ greatly, often addressing different objectives (Avison og Fitzgerald 1988). Det er ogs mange implisitte og eksplisitte antagelser og perspektiv bak hver metode (Iivari and Hirschheim 1996). Metoder har mange komponenter som spesifiserer: faser i et prosjekt, oppgaver som skal utfres, resultater fra prosjektet, begrensninger som foreligger, hvem skal vre involvert, verkty som skal brukes og hvordan man skal lede og kontrollere prosjekter (Avison og Fitzgerald 1995). Metoder blir laget for gjre systemutviklingsprosessen enklere og mer kontrollerbar. I noen r har det vrt en bevissthet om at metoder ikke lser alle problemer i utvikling av informasjonssystemer. Huisman og Iivari (2002) skriver at den utstrakte tro p at bruk av utviklingsmetoder I systemutvikling fremdeles er kontroversiell. Nyere underskelser rapporterer at mange organisasjoner sier at de enten ikke bruker metoder eller at de bruker deres egen inhouse metode (Huisman og Iivari 2002; Kiely og Fitzgerald 2003). Metoder kan ha mange forskjellige funksjoner fra det tekniske gjre det enklere kontrollere prosessen til den mer politiske siden hvor metoden kan fungere som ritualer (Robey og Markus 1984). Metoder m bli forsttt og brukt forskjellig avhengig av konteksten og den aktuelle utvikler. rvik el al. (1999) beskriver fire forskjellige forstelser av den samme utviklingsmetode relatert til en aktuell utvikling. Den frste er en formell beskrivelse av systemet. De tre andre oppstr alle i den aktuelle utviklingsprosessen; nemlig det utvikleren tolker, metoden forsttt av utvikleren; hvordan metoden blir brukt i organisasjonen, the adopted method; og hvordan den faktisk brukes, metoden i bruk. Iflge rvik et al. (1999) vil disse forskjellige forstelsene av metoden influere bruken av den. En del av metodelitteraturens beskriver prosessene i utvikling av informasjonssystemer som mer eller mindre mekaniske prosesser som frer til gitte resultater. Ciborra (1999) tar ett oppgjr med Business Process Re-engineering og dets forsk p utrydde improvisasjon fra konomiske organisasjoner. Han mener at improvisasjon er a well grounded process that can be leveraged to face those situations where rules and methods fail. Systemutvikling er p mange mter forandring. Forandring i organisasjonen kan vre mer eller mindre planlagt, og er langt mer vanlig i dag enn tidligere (Orlikowski 1996). Dersom systemutviklingsmetoder brukes uten tanke p forandring vil viktige elementer i systemutviklingen g tapt. Orlikowski (1996) beskriver sm forandringer, 28
som improvisasjoner relatert til ny teknologi kan fre til store forandringer i organisasjonen. Mathiassen et al (2000) skriver at metoder er viktige, men at gode systemutviklere kjennetegnes mer for erfaring og kompetanse de fr gjennom praktisk erfaring som den metodiske kunnskap de oppnr gjennom teoretiske studier. Cockburn (2003) diskuterer en-metode framfor flere-metode bruk i systemutvikling. Noen mener at en metode lser alt, mens Cockburn hevder at det kan vre ndvendig bruke flere metoder for utvikle et system. Han er en eksponent for agile systemutvikling som legger mindre stringente krav til metode i sin utviklingsprosess. Her vil det vre pent for den enkelte utviklers improvisasjon.
Prosessene
I dette kapitlet vil vi beskrive prosessene innen henholdsvis skapende teater og systemutvikling. Frst gir vi en oversikt tekstmessig over hvert av omrdene. Deretter lager vi en tabell med stikkord for de enkelte aktivitetene som inngr i prosessene.
PROSESS FOR SKAPENDE TEATER

Grunnpilaren i det skapende teater er improvisasjonen. Szatkowski (1991) sier at Mlet er, at s mange som muligt er s dybt som muligt involveret i en skabende tilstand, s lenge som muligt. I en skapende teaterprosess vil arbeidsrollene deltagerne har kunne skifte fra dramaturg til scenograf til skuespiller. Det er et ml at alle som deltar skal kunne bidra med ideer i alle ledd i prosessen. Dette er en vanlig arbeidsmte i skalt frie teatergrupper (ikke tilknyttet institusjonsteatre). Szatkowski (1991) har innfrt begrepet lre av lre lage teater. Det vil si at det er selve den skapende teaterprosessen og teaterforestillingen (produktet) som er essensielt. Det er der den viktige lringen foregr og ikke i hvilke tema som blir tatt opp i forestillingen. Utgangspunktet for en slik prosess kan vre alt fra et konkret tema til en abstrakt bevegelse (f. eks en gest). En forunderskelse eller research kan g i forskjellige retninger. Det kan vre en ytre research hvor en undersker temaet/ideen/impulsen eller mlgruppen en har valgt, og p den mten gr dypere og bredere inn i stoffet for tilegne seg kunnskap. Det kan ogs vre en indre research for huske assosiative hendelser, bevegelser eller bilder til den valgte impuls for bruke dette som en ressurs i arbeidet fram mot forestilling. Indre eller ytre research kan ogs anvendes som igangsetting av teaterprosessen dvs. ideen til forestilling kommer etter research og ikke fr. Dette grunnlagsmaterialet bearbeides s gjennom teaterimprovisasjoner. Foruten kunnskap og erfaring med improvisasjonsteknikkenes virkemidler, er fantasi og idrikdom grunnelementer i dette skapende teaterarbeidet. For Jacques Lecoq (2002) (startet egen teaterskole i Paris 1956) var det essensielt frigjre den enkelte skuespillerstudents skaperevner. Et sentralt begrep p hans skole er lauteur-acteur
29
(forfatterskuespilleren). Med dette mener han at skuespillere kan skape sin egen forestilling, og st inne for hele teateruttrykket, ikke bare som skuespiller. I hans system m studentene fra frste stund prve ut egne ideer, og fr dermed erfaring med en helhetlig arbeidsprosess. I improvisasjonsarbeidet har det vrt fokusert p muligheter og lsninger. Innenfor hvert omrde i en prosess er det forskjellige faser hvor det er vesentlig ha evnen til kunne velge eller velge bort, evnen til trre vre i kaos, evnen til takle motstand og tvil; og se p dette som viktige ressurser i en skapende prosess. Det vil ogs vre en dialektisk prosess mellom pne (opplsning av faste forestillinger og spredning av energi og konsentrasjon) og lukke (strukturering og samling av ideer) (Szatkowski 1991). ha mange ideer og muligheter kan fre til gode ideer. En god ide skaper ofte flere gode ideer. Som basis for arbeidet er det en fordel med energi og entusiasme, men motstand og konflikter kan ogs fre til nye og kvalitativt gode lsninger. jobbe med skapende teater innbefatter stor ide-produksjon. Det er formlstjenelig lete etter lsninger og ikke dvele ved problemer. Det blir viktig prve ut mange ideer (pne) for underske hvilke som fungerer. Dette vil oppleves som en kaoslignende tilstand. Det er en fase med uforutsigbarhet og overraskelser hvor sensureringslyst og kontroll ikke br vre tilstede. Nr prosessen er mlstyrt av et produkt, m en s gjre valg (lukke) og foredle ideer for f framdrift i prosessen. Denne vekslingen mellom vre i en skapende tilstand og strukturere/gjre valg er gjeldende i all skapende virksomhet som har et produkt som ml. Frst nr en kommer til innvingen av forestillingen br denne vekslingen i stor grad vre over. Men ingen regel uten unntak. Det kan ogs vre slik at valgene som taes under prosessen ikke kan sees p som 100% endelige. Et eksempel er hvis en skal lage en god slutt p forestillingen seint i prosessen kan dette f betydning for starten p sceneuttrykket som muligens m gjres om. Det er essensielt velge en struktur p dramaturgien i stykket som vil tjene teaterideen best. Ut fra improvisasjonsarbeidet m det skapes et konsept som det kan regisseres ut ifra. Etter dette kommer en rekke vinger med prving og feiling. Det vil s bli vist et mer eller mindre ferdig resultat for prvepublikum, som gir tilbakemelding p det de har sett og opplevd. Tilbakemeldingene brukes som ressurs i en avsluttende fase. Forestillingen vises s for valgte mlgrupper. Selv om forestillingen n kan betraktes som ferdig, vil den allikevel leve sitt eget liv, og sm endringer og justeringer gjres underveis ut ifra responsen til mlgruppene. Nr en er vant med jobbe p improvisatorisk basis opparbeider en kompetanse til raskt kunne endre og justere.
PROSESS FOR SYSTEMUTVIKLING

Det finnes mange forskjellige systemutviklingsprosesser innen systemutviklingsfagfeltet. Fossefallsmodellen er vel etter hvert blitt brukt mer som en pedagogisk framstilling av systemutvikling. Vi har iterative metoder og n ogs Agile systemutviklingsmetoder. I 30
denne sammenheng vil vi frst kort beskrive fossefallsmodellen og s iterativ og agile systemtviklingsmetoder. Men frst har vi en kort beskrivelse av aktrer i systemutvikling. De tre viktigste aktrene i en systemutviklingsprosess er styringsgruppen som tar beslutninger om rammer bde i forhold til hva som skal utvikles rent teknisk og hvilke ressurser som skal brukes. Systemutviklerne er aktive i hele prosessen. Brukerne er normalt aktive i de analyse- og designaktivitetene samt innfring og bruk av systemene. De er mindre aktive i selve programmeringen. En sammenfatning av systemutviklingsmetodene i Whitten, Bentley, and Barlow (1994) og Kendall og Kendall (2002) gir en standard utviklingsprosess kunne se ut som flger: forunderskelse, analyse, design, programmering, implementering, bruk, videreutvikling og avhending. Vi vil i det flgende beskrive de enkelte fasene i denne utviklingsprosessen. En forunderskelse kommer gjerne via en av flgende tre situasjoner: eget initiativ, ptrykk utenfra, for eksempel klager fra kunder eller p grunn av nye eller endringer av eksisterende offentlige reguleringer. Man prver finne ut av n-situasjonen, problemet og mulig framtidig informasjonssystem. Dersom det blir satt i gang et prosjekt gr man i gang med en analyse som er videre og dypere enn forunderskelsen. Man sker lage en kravsspesifikasjon for systemet. I denne fasen er framtidige brukere av systemet sentrale sammen med systemanalytikere/utviklere. Styringsgruppas ansvar er ta overordnede avgjrelser, srge for ressurser og gi tilbakemeldinger p forslag som mtte komme fra prosjektgruppa. Selve arbeidet utfres i prosjektgruppa og eventuelle undergrupper. Man gr s over i en designfase for gi svar p hvordan behovene/kravene oppfylles. I denne fasen er systemutviklerne sentrale siden de lager mye av designet. Brukerne gir tilbakemelding om de synes kravene oppfylles. Det er ikke sjeldent at denne fasen preges av skalte creeping requirements. Brukerne lrer mer om hva som er mulig gjre med systemet og vil derfor ha oppfylt flere og mer utfyllende ting enn de opprinnelig nsket. Dette kan fre til at man m g tilbake til analysefasen igjen for spesifisere kravene p nytt. Konsekvensene av dette er ofte forsinkelser og overskridelser av budsjetter. Designet gjelder bde funksjoner som skal utfres, brensesnitt som skal brukes og design av databasen som er lagringsdelen av systemet. Denne baseres p en modell av virkeligheten som tar hensyn til de oppgaver som skal lses av systemet og hvordan de skal lses. I denne fasen er det rom for kreativitet hvordan man kan lse oppgavene. Nr designet er ferdig overlates det til programmererne som lager applikasjonen. Denne m forholde seg til brukerne og deres rutiner slik at den letter arbeidet for de som skal bruke systemet. Her er det stort sett programmererne som rr grunnen. Det er likevel ikke uvanlig at de m snakke med designerne eller brukerne for forst hva de skal
31
programmere. Dette avhenger selvsagt av hvor godt designet er dokumentert og om det er konsistent. I implementeringen blir systemet satt i drift i organisasjonen som skal bruke det. Her er det brukerne som igjen overtar hovedrollen bruker systemet for gjre sin jobb bedre enn de kunne uten et slikt system. The proof of the pudding is in the eating er sant i denne sammenhengen. I denne fasen blir det klart om systemet svarer til hva brukerne nsket. Det er nemlig ikke sikkert at det de nsket er det de har kommunisert. Ved bruk vil det bli enda klarere om systemet lser de oppgavene det er konstruert for. Ettersom systemet brukes kan det oppst behov for videreutvikling som m gjres i samarbeid mellom brukerne og utviklerne. Dette kan vre prosesser som flger samme mnster som diskutert i det overstende, men for deler av systemet. Til slutt kan systemet bli avhendet og man gr over til et nytt system for lse de samme oppgavene. Iterative systemutviklingsprosesser har gjerne elementer av alle de ovennevnte aktivitetene. Utviklingen gr i flere iterasjoner som til slutt utgjr et ferdig produkt. Agile utvikling er enda lsere i kravet til struktur og system under utviklingen. Metoden er det som utviklerne blir enige om gjre i utviklingen av systemene (Cockburn 2003). De forskjellige metodene har ulike grader av frihet avhengig av hvor utviklerne er i prosessforlpet. I fossefallsmetoden nsker man gjerne fastlegge kravene tidlig for f en totaloversikt over det endelige systemet. Det kan by p vanskeligheter siden verden forandrer seg over tid og det kan vre vanskelig modellere systemet slik fremtidige brukere ser det (Mathiassen 2000). Iterative metoder tillater definering av kravspesifikasjoner gjennom prosessens gang. Samtidig begrenser friheten seg etter hvert etter som avgjrelser tatt tidlig i prosessen vil kunne innvirke p moduler som lages senere i prosessen (Omland 2004). Selv om utviklingen skjer iterativt eller ved bruk av en agile metode kommer det likevel en tid der frihetsgraden til forandringer blir mindre og mindre avhengig bde av hvor mye man har kodet og hvilke konomiske rammer prosjektet har. Den enkelte utvikler vil arbeide bde selvstendig, med ansvar for en gitt oppgave og i gruppe. Hovedfokus i arbeidet er ofte p produktet selv om prosessen ogs vektlegges. Det synes likevel vre relativt vanlig at systemutviklingsprosessene ikke flger bestemte metodologier, men at utviklerne tar elementer fra forskjellige metodologier/metoder og bruker disse ettersom det passer i utviklingsarbeidet(Kiely og Fitzgerald 2003). I det flgende sette vi opp prosessene for bde skapende teater og systemutvikling i en tabell med stikkord relatert til de enkelte aktivitetene i prosessen. Vi har valgt bruke stikkord fra de to fagomrdene som de brukes innen omrdene I systemutviklingsframstillingen har vi valgt inkludere stikkord fra alle tre metodene, fossefall, iterativ og agile utvikling. Disse tre metodene inneholder elementer av de forskjellige stikkord, men i forskjellige grad og i forskjellige deler av prosessen. Innen
32
fossefall vil f. eks analysefasen vre lokalisert i en gitt del av prosessen selv om det vil foreg analyseaktiviteter i hele utviklingsprosessen. I iterativ utvikling vil det kunne vre analysefase innen hver iterasjon. Skapende teater Forunderskelse Research, ske kunnskap om ide, mlgruppe, Energi, entusiasme og nysgjerrighet er viktig Valg av utgangspunkt ide eller tekst, impuls Fra konkret eller abstrakt ide Kaos - pen prosess Mange muligheter Fantasi Idemyldring Impulsivt Spontant Tilfeldigheter oppstr Lek-lyst-flyt Intuisjon Uenighet Ikke gjr valg for tidlig Valg av form Dramaturgiske modeller lukkende prosess: Aristotelisk (liner utvikling) Episk (minst to fiksjonslag-fortellende) Simultan eller visuell (samtidig-flere fiksjonslag) Metafiksjonell (fiksjoner kommenterer hverandre) Ironisk-logisk motivert Blandingsformer av ovennevnte (vanligst) Dette valget kan taes i forbindelse med konseptvalg Improvisasjon Utprving - pen prosess Ide-raffinering Skaperglede Motstand-frustrasjon-vegring Fleksibilitet Systemutvikling Forunderskelse Utgangspunkt i problemer, ting fungerer ikke Foretas enten internt, via konsulent eller utviklere Analyse Gr dypere inn i situasjonen datainnsamling Intervjuer Studere skriftlig materiale N-situasjon, nsket situasjon Forhold til strategier og planer i bedriften Analysen forsker beskrive virkeligheten slik fremtidige brukere vil se den innen sitt arbeide Prototyping Brukerkontakt/involvering
Design Utgangspunkt i resultatet av analysen Tegner og foreslr hvordan det nye systemet skal se ut Dette inneholder bde grensesnittet, funksjonene og modellen (databasen) Designet m forholde seg til den virkelighet brukerne er i M lse brukernes oppgaver Brukerkontakt/involvering
Improvisasjon Flere lsningsforslag Tilbakemelding Gjelder i alle aktivitetene i utviklingen men spesielt innen analyse og design aktiviteter
33
Lytte scenisk, vre oppmerksom, ikke blokkere andres ideer Kunne kill your darlings Samarbeide-samspill-involvering Lage et konsept Gjre valg lukkende prosess Lage helhet samlet uttrykk Lytte analysere se muligheter Flge konvensjoner bryte konvensjoner Notere Gjenta Raffinere Et skritt fram to tilbake Motstand Konsentrasjon Fokus Et dialektisk prosess mellom kaos og valg Teatervinger med regi Valg er tatt vinger med prving og feiling Tlmodighet Venting Utmattethet Konsentrasjon
Programmering Designet omsettes i praksis Valg av plattform Programmeringssprk Lite kontakt med brukerne Her foregr ogs testing Modultest System test Disse i forhold til brukerne
Prveinstallasjon Testing av kunden Tilbakemelding om feil nske om forandringer
Prveforestilling Vises for et utvalgt publikum Tro p produkt pen for innspill Teaterforestilling teater oppstr i mtet med publikum hver forestilling ny opplevelse yeblikkets kunst innlevelse mestring suksess/fiasko kick
Implementering Systemet settes i bruk i bedriften Opplring av brukerne Konvertering fra gammelt system Gradvis, parallelt eller revolusjon Bruk Brukerne tar det i bruk i sitt daglige arbeid The proof of the pudding is in the eating Her mter man gjerne problemer som ikke var forutsett nr systemet ble laget.
34
Videreutvikling gjentagelse skaper noe nytt kommunikasjon med publikum gir ny inspirasjon for- og etterarbeid refleksjon Avslutning Teaterstykket tas av plakaten
Videreutvikling Systemet trenger forandringer Feil rettes opp Brukerne melder feil eller utviklerne finner dem Avhending Systemet avhendes Rutiner Destruksjon av sensitivt materiale (siker etc)
Tabell 1. Skjematisk framstilling av utviklingsprosessene i skapende teater og systemutvikling
Produktene
Utviklingsprosessene sikter mot et produkt. Vi vil beskrive dette i det flgende:
TEATERPRODUKTET
Skuespillerne har valgt en mlgruppe for sin forestilling. Teater oppstr i mtet mellom scene og sal. Ethvert mte vil vre unikt. Ingen forestillinger er like. Hver gang en forestilling spilles, skapes den p nytt. Det hele avhenger av den uuttalte kommunikasjonen mellom scene og sal. Teater er yeblikkets kunst. Publikum opplever et produkt. For kunne vre i en prosess og produsere dette produktet m en ha kunnskap om teatrets virkemidler. Hver enkelt publikummer fr sin spesielle opplevelse. Som teaterarbeider kan en aldri vre sikker p at teatergruppens intensjon har kommet over scenekanten. Det er i mellomrommet medskapningen fra publikum sin side skjer. Forestillingen skapes i et samspill mellom skuespiller og publikum. En hovedintensjon kan vre gi publikum en interessant opplevelse. Det vil vre stor forskjell p responsen (den sagte og usagte) ut ifra i hvor stor grad publikum kan noe om teater. Et teatervant publikum vil kunne lese teaterkodene og konvensjonene lettere. Skuespillerne kan f en erkjennelse ved utfre en form. Det vil ogs som oftest gi et flelsesmessig kick som ogs innbefatter en mestringsflelse, og som brukes som energi ved neste forestilling. En forestilling m spilles flere ganger for kunne utvikle seg. Refleksjonen og evalueringen av hver oppsetning kan fre til videreutvikling av produktet. Utviklingspotensialet frer som oftest til bedre kvalitet.
INFORMASJONSSYSTEMET
Et informasjonssystem bestr av mennesker, rutiner, applikasjon og maskiner. Disse fire inngr i et samspill for lse de oppgaver som bedriften trenger utfre. Utfordringene
35
er at et slikt system med disse 4 forskjellige bestanddelene har forskjellig oppfrsel og forskjellige krav til hvordan de kan oppfre seg. Det maskin- og applikasjonsrelaterte krever at all input m vre p en gitt form. Mennesker kan utfre sine rutiner med mer eller mindre nyaktighet. Applikasjonen som kommer ut av en systemutviklingsprosess, spesielt hvis det er skreddersydd for bedriften, skal vare en del r. Brukerne vil derfor kunne slite med mangler i produktet over tid eller m gjre forandringer p et relativt nytt system. Det skjer ofte at produktet oppgraderes og videreutvikles. Feil rettes mer eller mindre fortlpende med utgivelse av enten korrigerte programdeler eller nye versjoner. I dagens situasjon vil sikkerhet vre i hysetet og eventuelle sikkerhetshull tettes fortlpende i mange systemer. Informasjonssystemet er bde fysisk, synlig og varig samtidig som det er usynlig. Det formelle, automatiserte systemet er implementert i rutiner, programvare og maskinvare. Informasjonssystemet m ogs skapes i kontakt med sine brukere for kunne vre av god kvalitet. Samtidig kan slik kontakt vre problematisk siden bde utviklere og brukere utvikler seg og lrer gjennom kontakten. Det kan fre til at systemspesifikasjonene forandres over tid. Det kan vre forskjellige personer som deltar i forskjellige deler av utviklingsprosessen.
Diskusjon
I diskusjonen diskuterer vi frst produktet deretter prosessen. Produktet diskuteres under to overskrifter, hensikt og varighet, mens prosessen diskuteres i relasjonen mellom brukerne, utviklerne og produktet. Innen skapende teater brukes ofte betegnelsene skuespillere og publikum. Tilsvarende brukes ofte betegnelsene utviklere og brukere innen systemutvikling. I denne sammenhengen velger vi bruke begrepene utvikler og bruker for begge fagomrdene.
PRODUKT
Vi kan se produktet fra forskjellige perspektiv. Vi begynner med hensikten med produktet og skriver deretter om varigheten. Hensikten Teaterforestillingen har til hensikt gi brukerne en kunstopplevelse. En kunstnerisk form formidles. Denne formen kan framkalle sanseopplevelser, tanker og flelser. Disse erfaringene kan skape nye erkjennelser. Brukeren kan vre meddiktende i forestillingen ved at forestillingen danner indre bilder hos bruker. Teater kan gi virkeligheten en fortolkning, vre en stilisert utgave av verden som gjr at en ser nye aspekter og finner nye sammenhenger i tilvrelsen. En teaterforestilling kan gi en god og viktig opplevelse selv om brukeren ikke har forsttt forestillingen intellektuelt. Brukeren kan ha skapt nye symboler og sett andre sammenhenger enn utvikleren har tenkt. Det viktigste er at produktet kommuniserer. En teaterforestilling er et mte mellom bruker og utvikler. Begge parter har en innvirkning p, og ansvar for kommunikasjonen.
36
Systemutvikling handler prinsipielt om forandring. Et nytt system utvikles gjerne for forbedre lsningen av eksisterende problem eller lse enten nye problem. Et informasjonssysstem har derfor ingen hensikt i seg selv. Brukerne av systemet vil lse sine arbeidsmessige problemer ved hjelp av systemet. Det betyr at utviklere og brukere trenger ha tett kontakt siden brukerne vil bli vrende med systemet i lang tid. Varighet Teater er yeblikkets kunstart. Produktet er tilstede i her og n dimensjonen. Men forestillingen kan gjre noe med brukeren over tid. Produktet kan ha kommunisert p en slik mte at opplevelsen lever videre og skaper noe nytt i ettertid. Bruker kan assosiere og fabulere videre p forestillingen. Utviklerne spiller oppsetningen en viss tid. Etter refleksjon og rapport er forestillingen ferdig og taes kun i sjeldne tilfeller opp igjen p et seinere tidspunkt Brukerne av systemet nsker en enklere hverdag der systemet hjelper dem i utfre det arbeid de er ansatt for gjre. Det er derfor viktig at systemet er tilpasset den arbeidssituasjon som brukerne er i til enhver tid. Det leder til mulig videreutvikling av systemet. Et informasjonssystem er vellykket i den grad det over tid lser de problem brukerne er interessert i lse. Et system kan vre i bruk i en rrekke i samme organisasjon og vil derfor vre med pvirke hvordan brukerne arbeider. Brukerne av et informasjonssystem er ofte avhengige av systemet og vil derfor vre opptatt av at det er i drift og fungerer etter det de skal bruke det til. Det kan vre alt fra taste inn data, gjre regnskap eller ta ut rapporter til bruk i beslutningssituasjoner.
PROSESS
En teaterprosess veksler mellom vre en lukket og pen prosess (Szatkowski 1991). Prosessen er en kunstnerisk prosess hvor mange muligheter og forslag prves ut gjennom improvisasjon og vil kunne betraktes som pen og uforutsigbar. Denne delen vil ogs ha elementer av kaos i seg. Men i prosessen m det gjres valg, og dermed lukker man eller strukturer deler i prosessen skaper orden i kaos. En skapende teaterprosess er lsningsfokusert. Prosessen kan starte med en pen problemstilling. Viktige ressurser for utvikle en forestilling er kunnskap og skaperglede. Teaterprosesser er aldri like. Men felles er at de frer til et produkt. Fra den pne problemstillingen kommer man gjennom prosessen fram til et konsept for s skape et produkt. Prosesser som beskrives til bruk i utvikling av informasjonssystemer brukes ofte ikke slik de er beskrevet (Mathiassen 2000). Utviklerne gjr utvalg ettersom de mener det tjener utviklingsarbeidet. Dette kan avhenge av blant annet strrelsen p systemet, utviklernes kompetanse og erfaring, tilgang p programvare, krav fra brukerne/oppdragsgiver etc. det som faktisk skjer i utviklingen vil derfor avhenge av disse faktorene og man kan f forskjellig prosess for hvert enkelt utviklingsprosjekt.
37
Mye av utviklingen skjer i prosessen som finner sted mellom utviklere og brukere, brukere og produkt samt mellom utviklere og produkt. Vi vil derfor diskutere disse forskjellige relasjonene i det flgende. Utviklere - brukere Det er viktig ha kunnskap om mlgruppen generelt. Ved henvende seg til brukerne og underske hva som opptar dem kan en f materiale til starte en teaterprosess. Dette kan en gjre ved intervjue, samtale, observere, men ogs ved ta tidsnden som kommer fram i forskjellige medier. En teaterforestilling trenger ikke direkte hente impulser fra mlgruppen. En teaterid som utviklerne har kan prves ut hos mlgruppen, og p den mten bli testet og f respons. S snart en prveforestilling er klar kan den vises for brukerne. De gir s innspill til utviklerne om endringer som med fordel kan gjres for at brukerne skal kunne f mer igjen for forestillingen. Dette skjer i samtale med de som har utviklet forestillingen. Det er imidlertid ingen selvflge at alle forslag bygges inn i forestillingen. Det er flere hensyn ta. Men responsen taes vare p, og forslag videreutvikles. De samme brukerne/prvepublikum/ ressursgruppe kan komme flere ganger, og p den mten utvikles et samarbeid. Enhver utvikling av et IS m ta hensyn til brukerne og deres behov. Det kan fre til en nr kontakt mellom de to partene. Ofte vil det vre motsettende interesser siden en utvikling ofte vil involvere ressursbruk og krav til lsninger. I tillegg vil brukerne ofte lre under utviklingsprosessen. De ser hva som er mulig f til. Dette kan fre til strre krav i forhold til lsningen som igjen kan lede til strre press p utviklerne. Utviklerne kan st i fare for tenke p hva de er i stand til gjre rent teknisk og bli opptatt av egne lsninger uten at dette ndvendigvis gir seg utslag i annet enn en demonstrasjon av egen fortreffelighet. Brukerne kan gi respons p utviklingen og produktet til utviklerne. En utfordring her er om utviklerne forstr responsen. En annen er om de er villige eller i stand til forholde seg til den. En tredje er om de er i stand til gjre noe med den som blir til fordel for brukerne. Dette vil nok variere med utviklernes involvering og interesse i prosjektet. Brukere - produkt En teaterforestilling er som en levende organisme. Selv etter premiere vil mlgruppens responser kunne pvirke forestillingen, og endringer gjres. I en forestilling som er improvisasjonssbasert vil det ogs under selve forestillingen kunne gjres sm endringer ut i fra publikums respons. En teaterforestilling fr ofte en umiddelbar og direkte respons - applausen. Hvordan applausen er er ogs et signal en m lre seg tolke. Det kreves at utviklerne utarbeider et kunstpedagogisk for- og etterarbeid til forestillingen. Dette for at brukerne skal kunne bli kjent med elementer av forestillingen i forkant, og bearbeide
38
inntrykk i etterkant. Noe som igjen kan sette i gang prosesser hos mlgruppen som skaper nye opplevelser og erkjennelser. Vanligvis blir brukerne vrende med informasjonssystemet som er utviklet. Det kan vre i bruk i organisasjonen i flere r. Dette kan fre til at brukerne stiller krav til utviklerne om bruk hvordan systemet skal vre. Brukerne fr ogs, i alle fall om det er snakk om skreddersm, ofte se prototyper av systemet og gi tilbakemeldinger p disse. Utviklere - produkt For komme fram til et teaterprodukt m utviklerne i fellesskap ha skapt dette. Men produktet kan ha blitt til med mye motstand i enkelte faser. Motsand er imidlertid ofte vesentlig for komme fram til et kvalitativt godt produkt. Det vre seg motstand fra hverandre eller motstand fra materialet de har arbeidet med. Det ha skapt et nytt produkt gir mestringsflelse og en form for kick som gir energi og glede, og som m sees p som viktig lring. Informasjonssystemet utvikles ved hjelp av verkty som utviklerne bruker. Utviklerne er avhengige av arbeid sammen for n et ml, nemlig lage et informasjonssystem. Det hender likevel at de enkelte utviklerer har egne agendaer og disse kan i noen tilfeller vre sterkere enn de felles agendaene som prosjektet som helhet har.
Konsekvenser for lringsmiljet

Gjennom studiet skal studentene forberedes for bde beherske prosessen og gjennom den framstille et produkt. Vi vil i det flgende kort vise hva vi mener de to fagmiljene kan lre av hverandre i forhold til lringsmilj.
PRODUKT
Hensikt Bde en teaterforestilling og et IS-system gir en opplevelse. Det er opplevelser med forskjellig kvaliteter. Teater er som nevnt yeblikkets kunstart og opplevelsen er tilstede i mtet mellom bruker og utvikler. Et IS-system brukes muligens hver dag og opplevelsen ligger p et annet plan, men en kan i begge tilfeller snakke om god eller drlig opplevelse. I begge tilfeller kan kvalitetsbegrepet anvendes, analyseres og kritiseres. Det er derfor viktig gjre studentene oppmerksomme p konsekvensene av opplevelsene bde intellektuelt, estetisk arbeidsmessig. Dette vil kunne gjre dem mer bevisst p egen rolle i utviklingen og p de konsekvenser brukerne sitter med ved god eller drlig kvalitet p opplevelsen. Varighet
39
Teaterforestillingen er fysisk tilstede over kort tid (1-2 timer). Opplevelsen kan leve videre i brukeren bde rasjonelt og emosjonelt. Hvis forestillingen ikke har hatt kvaliteter for brukeren, er det ute av ye ute av sinn effekten. Teateret kan kanskje lre hvordan produktets varighet kan utvides for eksempel ved at det kan tas opp igjen senere og spilles p nytt. Et IS-system kan ikke velges bort. Det er fysisk tilstede hver dag. Fokus p varigheten av produktet frer til at studentene trenger oppmerksomhet p sin egen rolle og sitt eget arbeid og hvordan sette brukerne i sentrum for utviklernes aktiviteter. Generelt kan vi si at et informasjonssystem kan involvere strre risiko i alle fall dersom det er snakk om et strre system og en mislykket utvikling vil kunne gi store tap bde for utvikler og bruker.
PROSESS
I begge prosessene kan vi anvende begrepet lsningsfokusert. For komme fram til en best mulig lsning m muligheter og ideer arbeides med. Det m gjres valg for s utvikle formlstjenelige muligheter videre. Det stilles i begge prosesser krav til hndverkskompetanse for f best mulig kvalitet. I en slik prosess hvor produktet er et ml kreves det bl. a. nysgjerrighet, engasjement, energi, en pen innstilling og samarbeidsevne. Dette gjelder bde utviklere og brukere. For kunne delta i slike prosesser er det viktig at studentene i begge fagomrdene lrer et godt hndtverk, kommunisere, vre bevist p den estetiske dimensjon av arbeidet samt lre seg selv bedre kjenne samt hvordan ens vremte kan pvirke brukerne av produktet bde i utviklingsprosessen og i bruksfasene. I tillegg er det viktig vge bruke egne kreative ressurser. Her kan skapende teater og systemutvikling lre av hverandre. En viktig forskjell mellom prosessene i skapende teater og systemutvikling er at skapende teater har stort sett de samme aktrene under hele utviklingen. I systemutvikling kan det vre forskjellige aktrer som analytikere, designere, programmerere osv. Det kan vre en stor utfordring for kommunikasjonen i prosjektet siden de enkelte aktrer er avhengig av informasjons fra aktrer de ikke har sett. En interessant aspekt i begge prosessene er det som kalles pne og lukke i teaterprosessen. Disse prosessene forekommer i begge fagfeltene, men har muligens forskjellig betydning og premisser/konsekvenser i utviklingen.
Konklusjon
Vender vi tilbake til forskningssprsmlet kan vi kort konkludere med at de to fagomrdene kan lre noe av hverandre.
40
Utviklere av IS-systemer kan lre bruke mer av egne menneskelige ressurser, og f en strre bevissthet om at mer enn rasjonelle egenskaper kan anvendes i en prosess fram mot et produkt. Utviklere av teaterforestillinger kan opparbeide en kompetanse for kunne forlenge varigheten av et produkt. En kan med fordel utvikle en plan for raskt kunne ta opp igjen et produkt, og p den mten forlenge en forestillings levetid. En kan lre av hverandre ved formidling og utprving av metoder, for utvide og erfare flere muligheter og lsninger i prosessarbeid. Vi ser det som faglig interessant starte opp et samarbeidsprosjekt mellom studenter som utvikler IS-systemer og studenter som utvikler teaterforestillinger. Et av mlene for et slikt prosjekt vil vre ke forstelsen for hverandres prosesser og produkter. En vil ogs kunne utvide kunnskapen om prosess og produkt generelt. Det kan ke bevisstheten om fagfeltet og studentene vil lre hverandres begrep og koder. Ved analysere hverandres kreative prosesser vil en kunne se flere sider ved arbeidsmetoden, og p den mten vil begge grupperingene av studenter kunne profitere p hverandres erfaringer og kunnskaper. For f best mulig kvalitet p dette prosjektet br studentene kunne observere deler av hverandres prosesser og reflektere over observasjonene. Vi ser at dette kan fre til et fruktbart samarbeide og verd prve ut i praksis. Det kan ogs vre inspirerende samarbeide i forhold til bevisstgjring av de ovennevnte forhold. pne og lukke aktivitetene i utviklingsprosessene br ogs vre interessante omrder for videre studier innen de to fagfeltene.
41
Kildehenvisninger
Avison, D. E., Fitzgerald, G. (1988) Information Systems Development. Methodologies, Techniques and Tools, Blackwell Scientific Publications. Avison, D. E., Fitzgerald, G. (1995) Information Systems Development: Methodologies, Techniques and Tools. London, McGraw-Hill Book Company. Braanaas, N. (1999) Dramapedagogisk historie og teori, Tapir Trondheim Ciborra, C. (1999) Notes on improvisation and time in organizations. Accounting, Management and information Technologies 9 pp. 77-94. Cockburn, A. (2003) People and Methodologies in Software Development, Ph. D, University of Oslo Guss, F.G. (2001) Drama Performance in Childrens Play-culture: The Possibilities and Significance of form. HiO-report 2001 nr.6 Heathcote & Bolton (1994) Drama for learning, Heineman, Portsmouth Huisman, M., and Iivari, J. (2002) The individual deployment of systems development methodologies, CAiSE 2002, Springer-Verlag Berlin. Iivari, J., and Hirschheim, R. (1996) Analyzing information systems development: A comparison and analysis of eight development approaches, Information Systems Journal 21(7) pp. 551-575. Kendall, K.E. & Kendall, J.E. (2002) Systems Analysis and Design, Fifth Edition. Prentice-Hall International, Inc Kiely, G., and Fitzgerald, B. (2003) An investigation of the use of methods within information systems development projects, In Proceedings of IFIP WG 8.2 Conference, Athens Greece Lecoq, L. (2002) The Moving Body. Methuen, London Mathiassen, L., Munk-Madsen, A., Nielsen, P.A., Stage, J. (2000) Object-Oriented Analysis and Design. Forlaget Marko, Aalborg Murray, S. (2003) Jacques Lecoq. Routledge, London Oddey, A. (1994) Devising Theatre. Routledge, London Orlikoswski, W.J. (1996) Improvising Organizational Transformation Over Time: A Situated Change Perspective. Information Systems Research. Vol. 7, No. 1, pp. 63-92. Omland, H.O. (2004) Relationships between developers competence, methods and practice in information systems development: A case study. Proceedings of the Thirteenth International Conference on Information Systems Development, Vilnius, Lithuania Septhember, pp. 305-316 Robey, D., and Markus, M. L. (1984) Rituals in information system design, MIS Quarterly (MARCH 1984) pp 5-15. Ross, M. (1978) The Creative Arts. London Seljeseth, J & Elnan, M. (2003) Hva er aller viktigst? Lekens rolle i teatraliteten. Nordisk Dramapedagogisk tidsskrift nr.4 Seljeseth, J. & Elnan, M. red. (2003) Det lekende barnet og den skapende skuespiller. Hia kompendium Slade, P. (1954) Child drama. London Szatkowski, J. (1991) Det bne teater, Drama 2-91 42
Way, B. (1967) Development through drama. London Witten, J.L., Bently, L.D., and Barlow, V.M. (1994) Systems Analysis and Design Methods, 3rd edition. Irwin rvik, T. U., Olsen, Dag H., and Sein, M. K. (1999) Deployment of System Development Methods. Exploring Paradigmatic Mismatches. In Evolution and Challenges in System Development, Zupancic e. al., Kluwer Academic/Plenum Publishers pp. 19-31, New York.
43
Developing e-Government Portals: From Life-Events through Genres to Requirements

Merethe Haraldsen, Thilde Dalsmo Stray, Tero Pivrinta, Maung K. Sein Agder University College, Department of Information Systems PO Box 422, 4064 Kristiansand, Norway Merethe.Haraldsen@adm.hio.no, Thilde.Dalsmo.Stray@visma.no, Tero.Paivarinta@hia.no, Maung.K.Sein@hia.no
Abstract
While there is a small but growing body of research on e-Government portals, a methodology especially designed for developing such portals is yet to be described. This is of growing importance as the possible benefits from e-Government solutions is receiving attention from governments around the world, and the fact that the potential audience for an e-Government portal reaches throughout the entire society. In this paper, we combined the concepts of life-events and genres to develop the information requirements analysis phase of a methodology to develop e-Government portals. To illustrate our method we chose the life-event building a house and identified genre systems based on a case study of the city of Kristiansand, a government municipality in Southern Norway. We then translated the genre systems to use cases. From the use cases we developed class, sequence and activity diagrams to demonstrate how genres can be used to develop such diagrams. We evaluated our proposed method by first presenting it to two experts, one a practitioner and the other an academic who stated that our method can be applied in a development project. We also used guidelines for design research to further evaluate our method. The results indicate that our method can be applied in conjunction with, or as a supplement to established system development methods.
Introduction
E-Government is an emerging topic, with an enormous potential to transform and rationalize public sector work through the use of Information and Communication Technologies ICT (Schware and Deane, 2003). An increasing number of governments realize the potential in e-Government solutions, and aims to utilize this and provide better services to the citizens (ibid.). Currently, this is being done mainly through the use of eGovernment portals.
44
Recently, considerable attention has been paid in both academia and practice to developing e-Government portals. The need for robust development process, including citizen representatives as active participants, has been addressed (Detlor and Finn, 2002). However, beyond general-level application of principles of participative design (Detlor and Finn, 2002), the literature gives little methodological guidance for the development of e-Government portals. A variety of systems development methodologies (SDMs) exist today (described as the methodology jungle by Avison and Fitzgerald, 1995). Tuunanen and Rossi (2003) assert that requirements elicitation for non-organizational users create additional issues than those addressed by existing SDMs, especially because of the large and heterogeneous user group. Precisely the same issues arise in an eGovernment context. One approach that has been suggested in the literature as a basis for e-Government portals is the life events method (Vintar and Leben, 2002; Vintar et al., 2002). Instead of focusing on the governments needs, these portals consider governmental operations from the citizens point of view. A life-event, e.g., marriage, usually involves more than one governmental agency, and the citizen has to contact more than one of these. The idea is that the e-Government portal should be cross-functional and reach through governmental bureaucracy. While this is a promising approach, life events do not provide the level of granularity required to precisely define systems requirements. Our suggested solution to this problem is to combine life-events and genre-based methods. Genre-based methods focus on organizational communication, and can be used as a method for information systems planning (Pivrinta et al. 2001). We took a design research approach and used the life-event building a house as an example. This lifeevent involves extensive communication with different governmental agencies and is considered a tedious process by the general public. We explored this application process in the city of Kristiansand. To determine the information flow between the citizen and the other involved parties we used a diagonal matrix session. The session identifies the different genres of information involved in the process. From the genre list generated in the diagonal matrix session, we elicited system requirements through an established systems development approach, UML. We converted the genres into UML-diagrams which can be converted into an actual physical design of a system. To evaluate our method, we used two approaches. First, we presented the logical design to two experts, one a practitioner from a company that develops e-Government solutions, and the other an academic whose research area is ISD. Second, we used the criteria for design research proposed by Hevner, March, Park and Ram (2003). The evaluations demonstrate that our method could be a usable and practicable method to develop eGovernment portals. The rest of the paper is organized as follows. Section 2 discusses the problem of developing e-Government portals. In Section 3, we describe the two cornerstones of our method, life events and genres. In Section 4, we describe the context we use to illustrate
45
our method, namely, the life event of building a house in the city of Kristiansand. Section 5 describes the application of the method in detail. We present the evaluation of our method in Section 6. We conclude the paper in Section 7 with a discussion on the research and practice implications of our paper and suggest future research directions.
E-Government Portals
E-GOVERNMENT
The term e-Government can be defined as making public services accessible through a shared electronic interface, e.g. the Internet. The following quote from Reffat (2003, p. 2) illustrates e-Government quite well; E-Government is not about putting in computers or building a Web site for information access; it is about transforming the fundamental relationship between government and the public. It is about transforming government service delivery through the use of technology. The ability to connect all transaction stakeholders across time and space barriers thorough the Internet and through a standard interface is fundamental to e-Government (Schware and Deane, 2003). Schware and Deane also state that the object of e-Government is to bring government closer to citizens by providing easier access to information, modernize public services by connecting different services and making them work more effectively. Achieving this objective requires the integration of various governmental agencies and connecting them with the public. The most popular solution for this is e-Government portals.
PORTALS
E-Government portals connect the government with the citizens, and vice versa. These are designed to broadcast content to the community at large, and cater for all citizens needs for information online. E-Government portals have the potential of providing benefits to both the citizen and the governmental agencies. The nonhierarchical character of the Internet delivery frees citizens to seek information and services at their own convenience, not just when a government office is open. The portals also provide access to accurate and up-to-date information at all times. In addition the portals might save the citizens numerous trips to the local government office, as everything is available online (Hutton, 2003). For the governmental agencies e-Government portals have the potential of providing a wide range of benefits. The self-service nature of the portals reduces the service administrative costs to a minimum and frees technically skilled staff to perform mission critical tasks. Portals also provide faster communication across all levels of government
46
and transactions may be performed rapidly and with less error. In addition the ability to share data and applications across agencies leads to less redundant data and less duplication of effort. All these lead to higher quality content and thereby improved trust and reputation of governmental services among the citizens (ibid.). The complexity in governmental services is reflected not only in systems design, and rather complicated requirements but also raises additional concerns. Today, people can get information on governmental websites, but to solve a problem they have to make a request to the right office. Most importantly, a citizens interaction with government is almost always centered on a specific event in his/her life. One approach that is used to deal with these complicating e-Government issues is the life-event approach (Vintar and Leben, 2002) and the concept of life-events has emerged as a starting point for developing e-Government portals (Bleek, 2001; Vintar and Leben, 2002; Vintar et al., 2003).
Proposed Solution: Live-Events and Genres

LIFE-EVENT PORTALS
Life-events are situations in life, which in this context are likely to involve some interaction with the government. Examples of life events are birth, marriage, death, graduation etc. According to Kavaidas and Tambouris (2002, p. 3), Life-events describe situations of human beings that trigger public services. For example, the life-event moving, triggers several interactions with the government. The problem is that the government is organized along different departments and divisions. When the user contacts the government, he or she only knows what he or she wants to achieve, and not how to achieve it. The citizens problems usually do not concern only one single public institution. The citizen does not know what administrative procedures are necessary, or in what order to complete these. For example if the citizen moves from one place to another, he or she has to change address, move the house phone number, inform the bank, employer, school, mail service and so on. To perform these changes the citizen has to contact several different departments or institutions. To change address one has to contact the national register. To change school one has to fill out a form and send it to the new school. If the citizen does not know which school the child belongs to, he/she has to contact the central school office in the municipality. In addition, the citizen has to contact relevant organizations and institutions like insurance company, Norway Post, phone company etc. For each procedure the citizen has to fill in an application, and data is often duplicated. Inconvenient office hours also complicate the process, and cause delays. The citizen her/himself has to find out which procedures are involved and in what order to apply and how. This process is complicated and time consuming, and often leads to frustration.
47
Life-event portals are one way of solving this problem. Instead of focusing on the internal, departmentally focused, needs of the government, these portals consider government operation from the citizens point of view and from the perspective of everyday life (Leben & Bohanec, 2003). The life-event has to include all services and the corresponding processes needed to solve the citizens problem from the beginning to the end. In this way all services needed are either linked or integrated into one single service (Vintar et al., 2002). Figure 1 shows a life-event portal.
Registry of normative regulations Registry of procedures and forms
Registry of life-events Electronic guide through life-events
Life-event Public portal
Communication interface Public servants Citizens and business
Figure 1: Active life-event portal (Vintar and Leben, 2002)
Some countries have started to develop life event portals, which are some of the most important initiatives in e-Government programs in Europe and worldwide today. These portals are still at early stages, and are continuously evolving. According to Vintar and Leben (2002, p. 1), the main objectives of these projects are to bring governments and citizens closer and to improve efficiency, effectiveness and transparency of its operation and increase the quality of its services. While the life-event approach provides a good starting point for e-Government portal design, the lack of clear development guidelines might be a weakness of this approach. The life-event approach does not provide the level of granularity required to develop systems requirements, and must therefore be used in conjunction with established methods and techniques. Kavadias and Tambouris (2003) proposed the use of GovML, a
48
specialized type of XML, in public system development projects. However, GovML is a development platform and does not address requirement definition. Since the life-event approach requires some level of interaction with the citizens, the intended users of the proposed system, the chosen systems development methodology must somehow include the user/citizen in the development process. Consequently, given the heterogeneity of the citizenry, eliciting user requirements becomes complex and difficult. As Tuunanen and Rossi (2003) argue, developing information systems for nonorganizational users create the problem of requirements elicitation because users do not know explicitly what they want from a system and it is difficult for them to express their ideas. In addition the user groups can be heterogeneous and be at different levels of computer literacy. Precisely the same issues arise in an e-Government context. In addition governmental agencies must take extra care to not exclude any group of citizens. More importantly, in terms of systems development, life-events do not provide the level of detail or granularity that is needed to generate systems requirements. As a solution, we propose adding genres to the method. Genres focus on the organizational communication, and requirements are identified based on the identified communication.
GENRES AND THE GENRE-BASED METHOD

The word genre originates from the Greek word genos, meaning race, kind, sort or class (Zimmerman,1994). Yates and Orlikowski (1992) define a genre of organizational communication as a typified communicative action which is invoked in response to a reoccurring situation. This means that a situation that occurs regularly in an organization, results in an organizations own standards (explicit or implicit genre rules) for communication in each particular situation. Every genre typifies communicative substance, being mediated through one or many forms identified typical with it. In this context, substance refers to the social motives, themes and topics referred in the communication, whilst form refers to the physical and linguistic features of the communication (Yates and Orlikowski, 1992). Genres form a part of a hierarchy that includes genre systems and genre repertoires. The relationship between these concepts is illustrated in Figure 2. Yates and Orlikowski (1998) describe genre systems as consisting of independent genres; these are typically enacted in a given sequence in relation to each other. Yates and Orlikowski (ibid. p.3) argue that genre systems are useful when studying interaction, because it focuses on how people use sequences of communicative actions to coordinate their activity over time and space. Kavaidas and Tambouris, (2002, p. 3) define life-events as situations of human beings that trigger public services. We thus argue that life-events trigger a genre system. While the life-event happens at a given point in time, the genre system spans a period of time. As individual genres, genre systems are organizing structures within a community; the genre systems give expectations about the purpose, content, form, participants, time, and place of communicative interaction. Examples of genre systems 49
are hiring process, consisting of genres such as job advertisement, application letter, and resume, and meeting consisting of genres such as agenda, the meeting it self, and summary of the meeting. Yates and Orlikowski (1994) define a genre repertoire as a shared set of genres within a community or an organization; this means a set of genres that are routinely enacted by members of the community. To understand the communitys communicative practices one must study the communitys genre repertoire.
Genre repertoire
A set of genres that are routinely enacted in a community.
Genre system Independent genres that are enacted in a given sequence.
Genres
Concept for describing organizational communication.
Figure 2: Hierarchy of genres
Genre theory has been suggested as a basis for IS development (Pivrinta, 2001). A genre-based method typically focuses on genres of communication occurring between different stakeholders communicating with each other in the domain of interest, resulting in an IS plan (Pivrinta et al., 2001) or requirements specification for a particular kind of information system, such as document/content management systems (Karjalainen et al., 2000; Honkaranta, 2003). First, stakeholders in the potential IS project are identified; this is done to select those who should participate in the planning effort. The next step is to identify producers and users of information (PUI-entities), in relation to the proposed IS project. The third step involves identifying and naming genres. This step seeks to identify the involved parties genre repertoire. The repertoire is established through a collaborative, facilitated session, using a technique called the diagonal matrix (Benson et al., 1988). We applied this technique, in line with Karjalainen et al., 2000), to identify genres that are shared among the PUI-entities as a basis for further analysis. After the 50
collaborative session, the facilitators list the genres, and the producers and users of genres. This list is the basis for the next step which is identifying properties of genres relevant to the information system planning. The following step is to gather metadata about the genres from the participants. The final step is analysis of the genre-based data to constitute an IS plan or requirements definition. Tyrvinen and Pivrinta (1999) report that genre-based thinking helps systems developers understand their development project from different viewpoints. They were able to expose several hidden assumptions on the information that were to be managed by the new system. This led to discussion and an in-depth insight into the problem area. In relation to systems development, Pivrinta (2001, p. 220) states that: genre concepts have the potential of bridging the gap between the social and technological aspects of information systems development. Identifying genres makes the users of the potential system able to participate in the communication on the technological implementation of information systems, and it makes it easier for the users to understand what the new system is about, and in what way it will interfere with their daily work practices. Based on the discussion above, we propose that using genres in conjunction with the lifeevent approach has the potential of providing better systems requirements in an eGovernment context. We adopted a design research approach to address this question. We chose the life-event building a house in the city of Kristiansand, Norway to illustrate a suggested use of our method.
Developing the Solution

THE CONTEXT: BUILDING A HOUSE IN KRISTIANSAND
Kristiansand lies in the county of Vest Agder and has a population of around 80000. To day most cases in the city are either handled by, or redirected by Servicebutikken. The Servicebutikk, which means Service Shop, is the engineering offices face towards the citizens of the city. It assists the public with issues such as rules and regulations concerning building and modifying houses, maps, building permits, waste management, snow clearance etc. Some of the information is available online, but most people stop by the shop with their questions. They also give out building permits over the counter. Completed applications, where the neighbors have approved, are considered within a day or two. The formal application process for a building permit is quite extensive. To be able to apply, one has to be an approved applicant. This means that most private persons have to go through a contractor or an architect who have been already formally approved. In some cases where it is obvious that the applicant has the formal qualifications, e.g., if the
51
applicant is a retired architect, a temporary approval can be issued. To study the process, we used a variety of methods: interview key personnel, studied documents (on Internet as well) and direct observation. Two of the authors spent two days at the Servicebutikken. The process is illustrated in Figures 3 and 4. Today this process in Kristiansand is performed with paper documents only, although all documentation is scanned into Doculive, which is a case handling system. Ideally this communication could be performed electronically.
52
Figure 3: Process overview
Figure 4: Appeal proces
53
Applying the method

GENERATING GENRE-BASED SOLUTION
To get an overview of the life-event building a house we used the method steps sketched by Karjalainen et al (2000) as the starting point. This is shown in Figure 5. As approval of a building permit involves much document handling and management, this model, originally targeted to analysing requirements for document management systems, is a good starting point. As shown in Figure 5, step one identifies units or stakeholders which are producing and using information in the domain (PUI entities). In addition we identify with whom the PUI entities exchange information, the external stakeholders. In step two we identify and name genres that are used to communicate between the different PUI entities. In the third step we define and gather developmental metadata of the genres. This can be information about existing storage format, desired storage format, archiving time, free comments on future development needs, or whatever issue of interest related to the genres . The fourth and last step is to streamline the metadata values and construct a requirements definition in light of the projects goals. In this stage general requirements for an e-Government portal are preliminarily discussed in order to configure a plan for future IS applications.
Information about the organization documented discussions Knowledge from the employees producing or using information in their work Information about the capabilities of existing e-gov portals and software packages LEGEND:
Information storage Step of the method Information flow
1. Identifying the units producing and using information
List of the units, persons and external stakeholders Diagonal matrices & list of genres including their producers and users List of the genres and additional properties of metadata about the genres A requirements definition for a future e-Gov portal A comprehensive map of the organizations information resources
2. Identifying the genres
3. Defining and gathering the desired metadata values of genres
4. Analyzing the metadata values and constructing a requirements definition
Figure 5: Genre-based method (adapted from Karjalainen et al., 2000)
Ideally the diagonal matrix session is performed in a collaborative context. Since we did not have access to all the involved PUI entities, we interviewed the head of Servicebutikken PD. Through the interview we got an overview of the entire application process as well as the governments point of view. Two of the authors represented the citizens point of view. The application process is standardized according to governmental regulations, and is therefore well documented. In another context where the problem in question is less formal or more complex, a collaborative session might be essential to achieve the desired results. The output of the Diagonal Matrix session was a matrix of raw data. Figure 6 shows an extract from this matrix.1. This matrix forms the basis for further analysis. Based on information about the organization and the interview with PD, we identified entities that use and produce information (PUI entities). This list of entities and the knowledge from employees producing and using information in their work is used to identify and describe the different genres.. The Diagonal Matrix session is used to identify the exchange of information between the different departments and stakeholders.
Figure 6: Extract of the Diagonal Matrix
The entire matrix and detailed tables and outputs mentioned later are available from the authors
55
While the matrix illustrates the the genres used to communicate between the different actors and departments that are involved in the application process, it does not say anything about metadata or functional requirements. We analyzed the matrix to extract the information that was relevant to us. The first step of the analysis was to create a list of genres, i.e., we listed all the genres found in a table. After organizing the genres in a table we started to identify metadata for each genre. We listed metadata collected through the Diagonal Matrix, such as user and producer of information, the owner of the genre, other stakeholders, desired medium, existing medium and existing file format. We also tried to give an estimate for the desired archiving time. We used metadata to describe the genre more thoroughly. By describing metadata such as producer, user, owner, stakeholder, medium and format, we can sort the genres according to these. In this way we can see what genres would be affected if any data system is replaced, or we can see who is producing or receiving what information. After collecting the metadata, we ended up with a list of genres where we also had described the values of the new metadata. This extended genre list formed the basis for the last step in the process. We then analyzed the values of each genres metadata, and from this constructed a requirement specification for a future e-Government system. In addition we get a comprehensive map of the organizations information resources. When constructing a requirements definition we have to standardize the values of the properties (Pivrinta et al., 2001). Doing this makes it possible to make a plan for future IS applications. First we collect answers from different participants, and then we synthesize this information and analyze it in the context of the projects goals. The result is presented in a table organized and clustered according to for example storage format (see Table 1). The genres can be sorted according to producer or storage format. In this way we can see which entity/department produces which genre, or to see which genres that are affected if one application is replaced. Based on the information in these tables we can create a requirements definition for the future e-Government portal. Table 1 List of Genres
Producer User medium Desired format storage time archiving
Genre Building order Signed contract Response to start-up Citizen Citizen Contractor Contractor Contractor Citizen E-gov system N/A Word Word
56
application Response to basic application Contract proposal Invitation to tender Contract Application for basic construction permit Application for startup permit Contractor Fellesarkivet E-gov system Contractor Contractor Contractor Contractor Citizen Sub-contractor Sub-contractor Fellesarkivet E-gov system E-gov system Word Word Word Contractor Citizen E-gov system Word
In this stage we also try to identify the genres that are most frequently used. This way we can find what functionality that is especially important for these genres. We could also identify needs to manage genres that are critical for the organization (Karjalainen et al., 2000). In our case this could be the applications for building permits. Karjalainen et al. also emphasize that requirements definition for the future system should be balanced according to the metadata values obtained for the document genres. To illustrate our method we developed UML-diagrams from the collected genres. UML does not specifically describe an analysis phase that proceeds developing of diagrams in the design phase. However, it is necessary to have a thorough investigation of the problem area before starting the design phase. The first goal with this phase is to identify use cases that correspond with business events, and the actors that cause these events (Larman, 1998). In our case we found that the use cases corresponded to the genres we identified as genre systems, and that the actors corresponded with the main producers and users of information. The UML diagrams, with examples of heuristics used to build them, are shown in the Appendix. Summary of the development process Our process for defining information requirements of e-Government portals can be summarized as follows: Choose target life-event(s) to be covered by the system. Identify the entities / stakeholders producing and using information. Identify the genres of communication taking place between the producers and users of information in the life-event(s) in question. Define and map metadata to the resulting genres. Analyze the metadata values. Use the genres to develop UML-diagrams Identify actors that produce input to the actual system, and actors that use information, by sorting genres on producers and users of information
57
Identify genre systems by sorting the genres according to granularity, the genre systems are identified as use cases, responses triggered by the genre systems are also identified as use cases Develop class diagrams from the use case diagram. The actors are identified as classes and identifiers are stored as attributes. The genre systems are also identified as classes and the attributes belonging in each class is extracted from the genres identified in the diagonal matrix session. Develop sequence diagrams from the use cases. The classes in the sequence diagram correspond with the classes identified in the class diagram. Develop activity diagram(s) from the use case(s). Our method for elicitation of requirements is illustrated in figure 7. The changes we made to Karjalainen et al. (2000) are marked in grey.
Evaluation
According to Rossi and Sein (2003) it is important to validate the results from a design research according to specific criteria. Admittedly, the best way to do so is to build a prototype and test and evaluate it. Since it was not possible to do so in our case, we performed validation of the concept and used two approaches. First, to assess whether the method can be useful for practice, we presented our method to two experts, one a practitioner and the other one an academic. The practitioner was a consultant at a Norwegian software vendor and consultant company, who develops systems for governmental agencies. The academic was an associate professor from Agder University College whose research area is system development methods and object-oriented system analysis. Next, we evaluated our method on the basis of normative criteria from Hevner et al. (2003).
58
Information about lifeevents Todays practice
1. Identifying relevant lifeevents
List of life-events Ideal practice
Information about lifeevents Todays practice
2. Identifying the units producing and using information
List of the units, persons and external stakeholders Diagonal matrices & list of genres including their producers and users List of the genres and additional properties of metadata about the genres A future plan for the related information resources, genre by genre
Knowledge from the employees producing or using information in their work Information about the capabilities of existing e-gov portals and software packages
3. Identifying the genres
4. Defining and gathering the desired metadata values of genres
5. Analyzing the metadata values
LEGEND: Information storage Step of the method Information flow
6. Requirements using UML
A requirements definition for a future e-Gov portal
Figure 7: Genre-based method (based on Karjalainen et al., 2000)
We first presented our method to the practitioner. We described our method explaining how we went from life-events and genres to UML. The expert stressed that our method might be implemented in practice. However, the consultant had no technical background and was not familiar with the concepts of UML. The consultant was of the opinion that existing systems development methods do not cater for the retrieval of the users viewpoints in a sufficient manner, and that there is no way to be sure that all user needs
59
are covered. After being introduced to the diagonal matrix he was of the impression that this method would be able to solve some of these problems. The diagonal matrix identifies the information flow between the different users of the system, and the fact that it includes the users in this process makes it likely that all users needs are catered for in the requirements definition. Drawing from his experience, the consultant stressed that the path from the information gathered from interviews with clients to design requirements seems a bit unclear. This is where the consultant thought using a more structured method for information gathering, such as the diagonal matrix, would be useful. To validate our methods transition from genres to UML we presented our work to an associate professor at Agder University College, an expert in UML. We described our method in the same way we did to the consultant, explaining all the steps from life-events and genres to UML diagrams. The professors reaction was positive. He was of the impression that our method is easy to use, assuming that the user has some knowledge of system development methods. He was also of the opinion that this method is possible to implement, and that it could be used in conjunction with other system development methods. Next, we performed a conceptual evaluation using Hevner et al.s (2004) seven guidelines for design research. A practical implementation is the only way to truly validate our method. However, using the guidelines gives an indication on whether an implementation could be successful or not. Hevner et al. (ibid.) argue that each of the guidelines should be addressed to complete the research. Table 2 shows how we addressed Hevner et al.s criteria. Table 2 Summary of evaluation (based on Hevner et al. (2004)
Guideline 1: Design as an Artifact Guideline 2: Problem Relevance Guideline 3: Design Evaluation Our artifact is the resulting method for requirements elicitation. e-Government portals are widely being developed as governments see the potential in e-solutions. Presentation of method to one academic and one practitioner. Testing using data from Servicebutikken. Descriptive method, using existing theory in arguing our methods feasibility. Our contribution is adding to the knowledge about the design of e-Government portals, and proposing a new method for elicitation of systems requirements for e-Government portals. Rigor is achieved using the existing knowledgebase and validation. Our research is iterated in several steps. Making the thesis available to a large audience.
Guideline 4: Research Contributions
Guideline 5: Research Rigor Guideline 6: Design as a Search Process Guideline 7. Communication of research
60
Discussion
In this study, we demonstrated that using life events in conjunction with genres is a practicable and method that can be used to elicit requirements to develop e-Government portals. In the following, we discuss the specific contribution s made by our study.
E-GOVERNMENT PORTALS AND LIFE EVENTS

A greater focus on e-Government solutions and how these improve performance puts pressure on governments to develop and implement initiatives such as e-Government portals. The e-Government maturity model (Layne and Lee, 2001) describes four different levels of e-Government maturity. The ideal situation is horizontal integration where systems are integrated across different functions and provide real one-stop shopping for citizens. Life-event is an established concept for structuring e-Government portals. Using lifeevents and employing the active matrix of life-events provides a way of reaching the highest level of e-Government maturity. The active matrix of life-events provides twoway interaction and contributes to a greater responsiveness and improved accessibility. Even though the concept of life-events is quite thoroughly explained in the literature; using the term event might lead to some confusion among those that are not familiar with the concept. Event as a term was first introduced by Yourdon in the Yourdon Systems Method (YSM), and has since been used as a term for something that triggers an operation or a series of operation. In YSM events are described as stimuli that occur in the environment of the system, to which the system must respond (Avison and Fitzgerald, 1995, p 276). However, we do not find that there is a contrast between the use of events in systems development and life-events. In Kavaidas and Tambouris (2002, p. 3) life-events are defined as follows: Life-events describe situations of human beings that trigger public services. This indicates that life-events are at different abstraction level than events in systems development. Using Kavaidas and Tambouriss (ibid.) definition, life-events corresponds to Yourdons events as something that happens in a persons life, and the system that responds to the event is the public services that are triggered by the life-event. Nevertheless, in systems development events are described as something that happens at a given point in time. Our life-event building a house can also be described as a continuous process that lasts until the house is finished. We argue that building a house also is a life-event because the interaction with the governmental agencies happens at a given point in time. Alternatively, building a house could be described as a superior lifeevent to events such as applying for basic construction permit, or applying for start-up permit. In addition building a house is listed as a life-event in Bleek (2001). It is also possible to view the life-event as an event that triggers a genre system. In this case genres are used to structure the resulting communication. Evidently, it is important to be clear on how the concept of life-events is used to avoid confusion.
61
INFORMATION SYSTEMS DEVELOPMENT

In the suggested method, all organizational communication and information flow is catered for using the diagonal matrix to identify genres. The diagonal matrix session is easy to comprehend and everyone in an organization can participate with out any previous knowledge of the technique. This gives a valuable contribution to the analysis phase as the entire organization can be involved in the diagonal matrix session if desired. In addition, the illustrated transition from genres to UML provides a correspondence between the UML-diagrams and the results from the analysis ensure a consistent design. IEEE (1998) states that the software requirements specification (SRS) should address the following issues; functionality, external interfaces, performance, attributes, and design constraints imposed on an implementation. We argue that our method addresses functionality and external interfaces, by developing UML-diagrams and catering for external interfaces in the metadata functionality. Performance issues, attributes and design constraints must be catered for using other methods available. This indicates that our method must be used in conjunction with larger methodologies or current practices, as part of the analysis and design phase. We have used life-events and genres to develop UML-diagrams. Any systems developer that wishes to develop UML-diagrams needs to perform some kind of analysis in advance. To use a specific example, Larman (1998) proposes a four stage process of object oriented design. The stages are: plan and elaborate (including the key requirement specification phase), analyze (including building a conceptual model), logical design and physical design. We argue that using our method in the plan and elaborate phase, the analyze phase and the logical design phase provides a more clear-cut procedure for the analysis that leads to UML-diagrams. To give a specific example, Larman (ibid.) emphasize that the primary goal of the requirements phase is to identify and document what is really needed in a software product and communicating these requirements to the client and the developers. Use cases serve this purpose although Larman (ibid.) does not describe a formal method for identifying the use cases or actors. We argue that our method can be applied to identify use cases or actors. Some may argue that using the diagonal matrix and genres to define UML-diagrams is not something new, as some kind of analysis must be performed to develop UMLdiagrams. However, we think that using our method is fairly straightforward and is a more structured way of performing this analysis. In addition, our method can be used in conjunction with, or as a supplement to, traditional analysis. We have used genres and the diagonal matrix to identify organizational communication in one specific context. When designing an e-Government portal other requirements related to portal functionality must be identified as well. We argue that this does not have consequences for our method, as our method is not intended to address all issues concerning e-Government portal design; it is intended as a technique that can be used in conjunction with or as a supplement to other methods. In addition, other portal 62
requirements, such as for example system performance measured in e.g. response time or navigation features, are common for all portals and should be catered for in the overall portal design.
LIMITATIONS
Our study is largely conceptual and our method has not been implemented in a real life context by building an e-Government portal. Being able to test the method and evaluate the resulting system, would be a far better and more precise evaluation than what we have managed to perform using various techniques. In the data collection phase we have used a technique called the diagonal matrix. Ideally, this should be performed in a collaborative session, involving representatives from all stakeholders. As this was not possible we performed this session ourselves, using data collected from interviews and laws and regulations. Although we think we have represented all involved stakeholders in the application process (we are citizens too!), we do not know this for sure. However, as the application process we have used as an illustration of our method is regulated by national laws, it is well documented and there is little room for personal interpretations.
IMPLICATIONS FOR PRACTICE

The life-event approach is a way to design e-Government portals. Combining this with the genre-based method provides a new approach to elicit requirements. The diagonal matrix identifies all involved stakeholders and thereby makes sure that all views are represented. The practitioner who evaluated our design immediately recognized the concept of life-events as similar to LivsIT, which has been applied in design of eGovernment portals in Norway (Aaberge & lnes, 2003). The LivsIt concept is described as a standard for connecting all information from the public sector related to a citizens situation in life in one portal solution. This standard is developed by Statskonsult. Examples of situations in life given are parent, job applicant or residential constructor (LivsIT prosjektweb, 2004). Both life-events and LivsIT are concepts for structuring information in an e-Government portal. Introducing the active matrix of life-events opens for two-way communication through an e-Government portal. However, focus is still on providing information through a portal. The consultant stated that offering interactive services, as processing of building applications online, is what he expects the municipalities will ask for next. In this context, he thought our method for gathering systems requirements might be applicable. The specific context of issuing building permits is a key area of attention in Norway. Currently there is a nationwide project aiming to make the consideration of building projects more efficient both for the citizen and the public authorities involved. This project is called Byggsok (Statens bygningstekniske etat, 2004). This project has already been piloted in Bergen where electronic applications for building permits are now being accepted. Currently this is only available for uncomplicated projects where no special
63
considerations are necessary. We suggest that out method can be helpful in extending these efforts to more complex situations.
IMPLICATIONS FOR FURTHER RESEARCH

As mentioned before, the evaluation of our method will not be completed until a practical implementation is performed and evaluated. A suggestion for further research is to complete the evaluation by developing an actual system. This can be done in an action research project, for example where the researcher participates in designing and implementing an e-Government portal. We believe that our proposed method might serve well as a basis for such a development. We have briefly looked into how our method can develop requirements for off-the-shelf solutions. As there seems to be an ongoing shift from in-house developing in organizations to a market where vendors provide ready-to-install products, a suggestion for further research is to explore this more in depth. Comparing our method in terms of the principles of systems development methodologies is another avenue of research. The concepts of life-events and LivsIT encompass the same principles, and put the users needs in focus. As we did not find any literature that describes these concepts in relation to each other, a suggestion for further research is to investigate the differences and similarities between these approaches. This could provide valuable input to each approach regarding techniques and procedures based on experiences with both approaches.
Biography
Merethe Haraldsen is Frstekonsulent in Oslo University College. She gained her Master in Information Systems from Agder University College, Kristiansand. Thilde Dahlsmo Stray is IT konsulent in Visma IT. She gained her Master in Information Systems from Agder University College, Kristiansand. Tero Pivrinta is Associate Professor of IS at Agder University College, Kristiansand. He gained his PhD (econ.) from University of Jyvskyl, Finland. His research has focused on the use of genre theory in ISD, enterprise content management, information systems architecture, and software architecture. He has published in Information & Organization, ACIT, and ca. 20 international books and conferences. He is a co-guest editor of an EJIS special issue on enterprise content management in 2005, a co-chair on the enterprise content management minitrack in HICSS, and the conference chair of IRIS 2005, to be hosted in Kristiansand, Norway.
64
Maung Sein is Professor of IS at Agder University College, Kristiansand. A PhD graduade from Indiana University, USA. He has served at Georgia State, Florida International, and Indiana Universities. He conducts research in end-user training, human-computer interaction, conceptual database design, systems development methods, data warehousing, computer personnel, IS management and currently in IT and national development. He has published in ISR, MISQ, CACM, and HCI and presented his research around the world. He has or is serving on editorial boards on MISQ, MISQ Executive, CAIS, and E-Service Journal and guest-edited CACM. He has chaired conferences and served on program committee of ICIS, ECIS and AMCIS.
Acknowledgements
We thank the two anonymous reviewers for their excellent comments that helped improve our paper. We would also like to thank Peder Dalen, Dagfinn Lauvsland, and Tore rvik for their time and advice.
References
Aaberge, T., lnes S. (2003). Skemotorassisterte informasjonsportalar. In (eds.) T.J.Larsen & B. Bygstad, NOKOBIT 2003, pp. 60-73. Avison, D.E. and Fitzgerald, G. (1995). Information systems development; Methodologies, Techniques and Tools. 2nd edition. McGraw-Hill, London. Benson, Y., Hyrylinen, M., Saaren-Seppl, K. (1988). Wall Chart Technique. Model Charts for Designing EDP-Systems. Wimage Oy, Helsinki. Boyne, G.A. (2002). Public and private management: Whats the difference? Journal of Management Studies, 39:1 Bleek, W.-G. (2001). Situations in Life to Support the Use and Modelling of Municipal Information Systems. The European Conference on e-Government 2001. [Online] Accessible: http://swt-www.informatik.uni-hamburg.de/publications/files/ECEG2001-bleek.pdf [2004, 03.02] Detlor, B., Finn, K. (2002). Towards a framework for government portal design: The government, citizen, and portal perspectives. In . Gronlund (Ed.), Electronic government: Design, applications, and management. Hershey, Pennsylvania: Idea Group, pp. 99-119 Hevner, A.R., March, S.T., Park, J. and Ram, S. (2004). Design Science in Information Systems Research. MIS Quarterly, Vol. 28, No. 1, 75-105 Honkaranta, A. (2003). From genres to content analysis. Experiences from four case organizations. PhD Diss. University of Jyvskyl, Jyvskyl. Hutton, G. (2003). Building a Business Case for E-Government Portals. Vignette, [Online]. Accessible: http://www.vignette.com/Downloads/fed/federal_value.pdf [09.02.04]
65
IEEE (1998). IEEE Std. 830-1998: IEEE Recommended Practice for Software Requirements Specifications. [Online]. http://standards.ieee.org [09.03.04] Karjalainen, A., Pivrinta, T., Tyrvinen, P., Rajala, J. (2000). Genre-Based Metadata for Enterprise Document Management. Proceedings of the 33 rd Hawaii International Conference of System Sciences 2000. Kavaidas, G. and Tambouris, E. (2003). GovML: A Markup Language for Describing Public Services and Life Events. Knowledge Management in Electronic Government: 4th IFIP International Working Conference, KMGov 2003, 106-115. Larman, C. (1998). Applying UML and patterns An Introduction to Object Oriented Analysis and Design. Prentice Hall PTR, New Jersey. Layne, K. and Lee, J. (2001). Developing Fully Functional e-Government: A Four Stage Model. Government Information Quarterly 18, 122-136. Leben, A. and Bohanec, M. (2003): Evaluation of Life-Event Portals: Multi-attribute Model and Case Study. KMGov 2003, 25-36 Leben, A. and Vintar, M. (2003). Life-Event Approach: Comparison between Countries. EGOV 2003, 434-437. LivsIT prosjektweb (2004). [Online]. Accessible: http://www.livsit.info [2004, 12.05] Reffat, R.M. (2003). Developing a successful e-Government. Proceedings of the Symposium on e-Government: Opportunities and Challenge. Muscat Municipality, Oman, [Online] Accessible: http://www.arch.usyd.edu.au/~rabee/publications_files/03Reffat_eGov.pdf [2004, 03.04] Rossi, M. and Sein, M.K. (2003). Design Research Workshop: A Proactive Research Approach [Online]. Accessible: http://tiesrv.hkkk.fi/iris26/presentation/workshop_designRes.pdf [2004, 20.04] Pivrinta, T. (2001). The concept of genre within the critical approach to information systems development. Information and Organization, 11, 207-234. Pivrinta, T. Tyrvinen, P. Ylimki, T. (2002). Defining organizational documents metadata: a case beyond standards. Proceedings of the Xth European Conference on Information Systems (ECIS), Gdansk, Poland Schware R. and Deane A. (2003). Deploying e-Government programs: the strategic importance of I before E, The journal of policy, regulation and strategy for telecommunications, Volume 5 No. 4, 10-19. Statens bygningstekniske etat, (2004) [Online]. Accessible: http://www.byggsok.no [2004, 24.03]
66
Tuunanen, T. and Rossi, M. (2003). Market Driven Requirements Elicitation via Critical Success Chains. Proceedings of the 11th IEEE International Requirements Engineering Conference. Tyrvinen, P. and Pivrinta, T. (1999). On Rethinking Organizational Document Genres for Electronic Document Management. Proceedings of the 32nd Hawaii International Conference on Systems Sciences. Vintar, M., Kunstelj, M. and Leben, A. (2002). Delivering better quality public services through life-event portals. [online] Accessible: http://www.unpan.org/conf_10NISPAceemeeting02.asp [2004, 30.01] Vintar, M. and Leben, A. (2002): The concepts of an Active Life-Event Public Portal. EGOV 2002, 383-390. Wimmer, M., Traunmller, R. and Lenk, K. (2001). Electronic Business Invading the Public Sector: Considerations on Change and Design. Proceedings of the 34th Hawaii International Conference on Systems Sciences. Yates, J. and Orlikowski, W. J. (1992). Genres of organizational communication: a structural approach to studying communication and media, Academic of Management review, 17 (2), 299-326. Yates, J. and Orlikowski, W. J. (1994). Genre Repertoire: The Structuring of CommunicativePractices in Organizations, Administrative Science Quarterly, 13, 541574. Yates, J. and Orlikowski, W. J. (1998). Genre Systems: Structuring Interaction through Communicative Norms. [Online]. Accessible: http://ccs.mit.edu/papers/CCSWP205/ [2004, 28.04]. Zimmerman, E.N. (1994). On definition and rhetorical genre. In A. Freedman & P. Medway (eds.) Genre and the new Rhetoric. Taylor & Francis, London, pp. 125-132.
Appendix: From Genres to UML diagrams

We suggest that when designing e-Government systems, the life-event approach is suitable, because it focuses on the intended users needs. As the target for our approach is to elicit requirements for an e-Government portal, we make the presumption that the scope of a development project that wishes to use this approach is clear. After deciding on life-events, we used a genre-based method to identify the information flow between the parties involved in the suggested life-events. The information flow was identified in a diagonal matrix session and the resulting genres were analyzed by defining metadata values; as producer, user, owner, current medium of communication and desired medium. The analysis defined the starting point for developing UML diagrams (See Figures a, b). First we developed use case diagrams (Figure a). To identify the actors we used the genre list sorted by producers and users of information and recognized the main producers. 67
These were identified as actors. The use cases were identified using a genre list where the genres were sorted according to granularity. The genres that were identified as genre systems are the genres that lead to interaction with the system; also none of the other genres would exist on their own, and could be defined as sub-genres to the genre systems. The genre systems correspond to the notion of events, and were therefore identified as use cases, as were the responses triggered by the genre systems. Two additional use cases were added resulting from the inherit user-focus in the life-event approach. From the use case diagrams we developed class diagrams (Figure b). The actors were identified as classes and the relevant identifiers were stored as attributes. The genre systems were also identified as classes and the attributes in each class correspond with the genres identified in the diagonal matrix session. If some of the classes contained the same attributes, a super class was defined to simplify the data storage. From the use case diagrams we also developed sequence diagrams for each use case (due to the limited space, we do not include them here, as there are rather straightforward rules how to develop the other diagrams together with use case and class diagrams in UML). The sequence diagrams show how the actors interact with the system; and the events that are triggered by the actors input. The classes in the sequence diagram correspond with the classes identified in the class diagram. Finally we developed an activity diagram, describing the three use cases register application for construction permit, register application for start-up permit and register request for certificate of practical completion.
68
5 HTXHVWQ P DW L IRU L RQ
5 HJL HU VW DSSO L L RQ FDW IRU EDVL FFRQ U L VW RQ XFW SHU L P W
5 HJL HU HVSRQ R VW U VHW EDVL FDSSO L L RQ FDW
5 HJL HU VW DSSO L L RQ FDW IRU DU VW W XSSHU L P W
&RQ U RU W DFW 5 HJL HU HVSRQ R VW U VHW VW W DUXSDSSO L L RQ FDW
&DVHKDQ O GHU
5 HJL HU HTXHVW VW U IRU FHUL FDW W HRISU L IL DFW FDO FRP SO L HW RQ
5 HJL HU W FDW VW FHUL IL H RISU L FRP SO L DFW FDO HW RQ
7U DFHDSSO L L RQ FDW
&LL HQ W ]
Figure a. Use case diagram
69
Figure b. Class diagram
70
Information Systems Innovation and Adaptation
SESSION 2 Information Systems Innovation and Adaptation
The broad theme for this track is processes of IS Innovation and Adaptation. The first paper by Bygstad et al. investigates the relationships between business innovation and the use of software development process. Next, Larsen and Sreb presents a study aiming at extending the ExpectationConfirmation Model (ECM) of IS continuance intention.
1. "What is the relationship between software development processes and IT based business innovation? An exploratory study in Norway" by Bygstad, Fagerstrm and stensen 2. "The Role of Actual IS Utilization Level: Extending the ExpecationConfirmation Model of IS Continuance Intention" by Larsen and Sreb
What is the relationship between software development processes and IT based business innovation? An exploratory study in Norway.
Bendik Bygstad, Asle Fagerstrm and Tore stensen The Norwegian School of Information Technology
Abstract
This paper investigates the relationship between software development methods and business innovation. Two hypotheses were tested; one that assumes that IT-based business innovation can be planned and managed as a part of a top-down strategy, and one that assumes that innovation is basically emergent and unmanageable. A pilot survey was done in the Norwegian software industry, using a web-based questionnaire. 123 companies responded, from a total population of 610. We find that neither the top-down business planning approach, nor the emergent "amethodical" view is supported. Instead we find that the innovative projects are associated with the use of adapted development methods or various techniques. Innovative projects are neither initiated from top management nor improvised; they are usually initiated at department level. Further, our findings suggest that they occur predominantly in young companies. We interpret these findings within a resource based innovation view: IT-based business innovation is best supported by local initiatives in a competent environment, using locally adapted techniques.
Introduction
Can technology based business innovation be planned and managed, or is it a creative and emergent phenomenon? In the past ten years there has been an increasing interest in technology based innovation, both in the research communities and in national policymaking. The main reason, on the national level, is the assumed impact of technology on economic growth rates. On a business level the ability to innovate is seen to be crucial for the company to survive in the global market place.
72
In the 1990s the tremendous potential of Internet technology significantly changed the ways we combined technology and business to innovate, and also our vision of what was possible. Subsequently, the history of the dot com crash in 2001 taught us some important lessons. First, technology in itself cannot provide competitive advantage. As IT has become a wide-spread commodity; its importance for business performance is increasing, but its importance to competitive advantage is decreasing (Carr 2003). Second, the nature of IT based innovations seems to be elusive. After the initial frenzy, both the BPR and the dot com booms faltered. Many companies tried to learn from the early innovators, like Amazon and eBay, without much success.
This implies that we need to understand better several aspects of IT-based innovations. A basic question is: Can IT-based business innovation be planned and managed? There are two opposing views on this question. The first view is prevalent at business schools, and assumes that IT-based business innovations are the results of careful analysis, strategic positioning and managed implementation (Porter 2001, Davenport 1993, Laudon and Laudon 2004). The second view assumes that innovation is basically creative, emergent and unmanageable in a business context. Echoing Schumpeter, the American economic historian Mokyr put it this way: "Technological creativity, like all creativity, is an act of rebellion" (Mokyr 1992). It is probably useless to try to answer this question on a general level, because there are too many contingency factors that influence innovation. But we think it makes sense to investigate the question on a more limited scale. We define IT-based business innovations to be a combination of software and business process change resulting in new business concepts or improved business performance. This change process may be managed by structured development methods, or it may be done on an ad hoc basis. Thus, it is interesting to investigate whether successful innovation is associated with structured development methods, or - oppositely - with ad hoc initiatives. Therefore, the research question we try to give some preliminary answers in this pilot study is: What is the relationship between software development processes and IT-based business innovation?
The rest of the paper is structured as follows. In the next two sections we present our hypotheses and discuss the key terms of business innovation and software development methods. Then we present our research approach, followed by a presentation of our findings and a discussion in the following section. Lastly, we conclude and suggest directions for further research. 73
Hypotheses
In this section we present our two competing hypotheses, one based on a top-down planning view, and one based on an emergent view.
Alternative 1: Innovation can be planned and managed

From a business perspective a top-down innovation approach is well documented in the Strategic Information Systems research stream (Laudon and Laudon 2004), emphasizing how planned IT-enabled business innovation can be used to achieve competitive advantage. This approach is reflected in most of the standard software development methods: Information systems projects should be anchored in a business and product strategy (Tabrizi and Walleigh 1997), and start with a top-down business analysis (Jacobson et al. 1999, Stapleton 2003). Failed projects are explained by being technology-driven and lacking a clear and structured process (Cadle and Yeates 2004). Building on this perspective, we suggest the following hypothesis: H1: A high degree of innovation is associated with a high degree of development process structure, given high formal product development process.
Alternative 2: Innovation is emergent

In a series of case studies of large corporate information infrastructures, Ciborra (2000) found that although the management literature recommends a top-down approach, the evidence for the viability of this approach is scarce. On the contrary, the authors document that business innovation is often the result of technological drift and unintended consequences. There are various reasons for this drift. One is that established business models and practices are badly suited to innovation (Christensen and Overdorf 2000). Another is that the dynamics of IT-based business development are not easily controlled by traditional management and development methods (Truex et al. 2000), and tend to develop more through "tinkering and bricolage" (Ciborra 2002). From this perspective we suggest the alternative hypothesis: H2: A high degree of innovation is associated with a low degree of formal development process structure.
Key terms: IT-based business innovation and development methods

In this section we define and discuss our key terms.
74
IT-based business innovation

IT-based innovation has been studied extensively for two decades, but there is hardly a consensus on how to define it. Some important contributions are: Diffusion of innovations theory (Rogers 1995), describing the diffusion pattern of an innovation in a user community. The mutual adaptation of technology and business (Leonard-Barton 1988), stating that the real innovation is the unique and mutually adaptive implementation in an organisation. Process Innovation (Davenport 1993), describing IT-enabled business process reengineering. The IS Innovation Framework (Larsen 1998), synthesizing innovation research in the three phases of ideas, creation and usage, focusing on the human activity system.
These contributions all emphasize that the organisation is as important, or more important, than the technology. Building on this insight we choose a relatively simple definition: At the core of IT-based business innovations is the combination of a new technology with a new or changed business process. To measure the degree of novelty of such innovations is obviously problematic, since a very small technical innovation (like SMS) may have large business consequences, while a very sophisticated concept (like the Semantic web) has had small impact so far. We still think it is reasonable to say that an IT-based business innovation usually includes both the use of new technologies and the design of a new business process. Following Evans (2003) we use the following criteria. We take the use of new, enabling technologies, like web, java, microsoft.net, web services, and peer-to-peer technologies to indicate technological innovation. Further, to measure business innovation we take novelties and changes in business processes supply chains knowledge management customer relations as indicators for the degree of business innovation. And lastly, we take combinations of technological innovation and business innovation to be a particularly strong indicator of IT-based business innovation. For details, see table 9.
Development methods
To investigate our hypotheses, we measure the co-variation between IT-based business innovation and development methods. Building on earlier research (Fitzgerald 1998, 75
Fagerstrom et al.2003) we assume that the majority of the companies use their own development method, that a minority of the companies use standard development method like RUP (jacobson et al. 1999) and XP (Stapleton 2003), and that some companies do not use methods at all. Further, we assume that many companies use elements from formal methods, adapted to their own environment.
Research approach
This section begins with a detailed description of the sampling and sampling design that was used, followed by how the research design and analysis of survey responses are determined. In the end, sample validity and non-response bias are discussed.
Sampling
Ideally, all companies involved in software development in Norway should be defined as the population for this research. This includes general private companies and public organisations as well as professional companies within the IT sector. However, the two last years research (Bygstad et. al. 2002 and Fagerstrm et. al. 2003) showed that it is very hard to get responses from the general private and public businesses. As a result, this evaluation includes only professional IT companies (IT consultants and professional software development companies) in the population. This population is limited by three different Norwegian industrial classification (IC) codes: 7220000 (System- and software consulting), 7260001 (IT consulting) and 7260003 (IT services). Based on experiences from last years research (Fagerstrm et. al. 2003), and in order to facilitate trend analysis and comparison with last years survey, only companies with 5 or more employees and a valid contact person were considered. There were in total 1007 such companies in May 2003. All the companies were telephoned and asked whether or not they were actually involved in system development, reducing the total population to 627. Of these, 287 responded in the survey. In May 2004 the same 287 companies were contacted and were asked if they wanted to participate in this years survey. 204 companies agreed to participate.
Research design
The survey was based on a questionnaire (see Appendix A) and implemented electronically by using the QuestBack system2. This system is based on e-mail
2
www.questback.com
76
distribution of a link to the actual survey and replies via a web browser on the internet. Email addresses were collected during the process of calling all companies in the survey sample. The QuestBack system has an automatic reminder, which was scheduled three or four times to those who had not responded after the request to participate in the survey was sent out. In this process, a number of respondents withdrew from the survey (by sending a separate mail), others just didnt reply. Also, there were some errors in the email addresses, which were returned by QuestBack or external e-mail servers. After about eight weeks the survey was closed with these results: Table 1. Sample segments when survey was closed Group Companies that agreed to participate 1. Actual number of replies 2. Not answered 3. Withdrawn Size 204 123 61 20 60 % 30 % 10 %
Sample validity and non-response bias

Non-responses can be attributed to two primary reasons: 1. No answer The respondent agreed to participate, but did not do so 2. Withdrawn The respondent withdrew after having read the questionnaire Our total population (companies which actually are involved in software development) can be estimated to around 627 companies. Thus, the response rate is ca. 20%. The results (estimated percentages for the total population) of the survey should be accurate to a level of plus or minus 8 %.
Findings
This section presents the results from the survey, and is divided into four parts: (1) type of IT project that the respondents selected to be the basis for answers in this survey, (2) development methods that had been used in the selected IT project, (3) degree of innovation and (4) degree to which product development is anchored in a product strategy.
Selected IT projects
Respondents were asked to choose one important IT project that has been accomplished the last year. Table 2. Chosen IT project that has been accomplished the last year. Selected IT project In-house SW development Frequency Percent 58 47,2
77
SW development for customer In-house adaption/implementation of standard SW Adaption/implementation of standard SW for customer Other Total
35 6 16 8 123
28,5 4,9 13,0 6,5 100,0
In table 2 we can see that most of the projects selected by the respondents are in-house software development IT project 47,2 percent chose this type of IT project. It is important to have this in mind when reading the rest of findings and discussion in this paper: This implies that in some cases the innovation is done within a technology firm (in-house), in other cases it is the result of a shared project between SW company and their customer.
Development methods
Responses to the question to describe which development methods were used in the selected IT project are given in Table 3.
Table 3. Development methods that were used. Development methods Frequency Percent Standard SDM (e.g. RUP, MSF) 9 7,3 Adapted SDM 13 10,6 Own (or consultant's) method 39 31,7 A combination of techniques 41 33,3 No method 21 17,1 Total 123 100,0 Table 3 shows that most of the companies in this survey use a combination of techniques (33,3 %) in the IT project that they selected to be basis for answers in this survey. Only 7,3 % use standard system development methods e.g. RUP, MSF. These results are consistent with our findings in a previous survey (Fagerstrm et.al. 2003), and other finding within this research field (Fitzgerald 1998).
Degree of innovation
To measure the degree of innovation in the IT project we asked the respondents to estimate the degree of innovation related to: (1) technology, (2) business processes and (3) outcome of the particular IT project that was selected. Table 4. Innovation related to technology. Technology Frequency Percent
78
Web Java or J2EE Microsoft.net Web services Peer-to-peer services Other
63 30 45 37 4 44
51,2 24,4 36,6 30,1 3,3 35,8
The percent of technology that was used are greater than 100 % because the respondent was able choose one or more of the alternatives. From table 4 we can see that Web and Microsoft.net are widely used by the companies in this survey. It may be somewhat surprising that only four companies say that they use peer-to-peer services technology. As table 5 shows 35,8 % of the respondents say that the selected IT project caused no change of business processes, while 51,2 % (adding the other categories) say that the IT project resulted in completely new business model or change in parts of the business model.
79
Table 5. Innovation related to changes in the business processes. Change in the business processes Frequency Percent Completely new business model developed 18 14,6 Supply chain are changed 9 7,3 New solution for knowledge management 12 9,8 New/changed solution for customer relations 24 19,5 No change of business processes 44 35,8 Other 11 8,9 Missing 5 4,1 Total 123 100,0 Table 6 shows that around half of the projects are focused on IT-based business innovation (combinations of new technology and business change), while the other half are purely SW projects Table 6. Outcome of the particular IT project. Outcome of the particular IT project Frequency Percent New version of existing software 45 36,6 Completely new software 19 15,4 New/changed SW and improved business process 30 24,4 New/changed SW and new business process 10 8,1 New/changed SW and new business concept 13 10,6 Other 3 2,4 Missing 3 2,4 Total 123 100,0
Anchoring in a product strategy

To be able to measure the degree of anchoring in a product strategy we asked the respondents to estimate: (1) if the project was a isolated project or part of product or business development, and (2) how formally the project was related to a product development process. Table 7. Relation to a product or business development process. Relation to a product strategy Frequency Percent Isolated project 36 29,3 Part of product- or business development 87 70,7 Total 123 100,0 We can see in table 7 that the majority (70,7 %) of the projects were related to a product development process. A follow-up question to the respondents who said that the project
80
was related to a product development process (n=87), yielded the results shown in table 8. Table 8. Degree of relationship to a product development process. Degree of relation to a product development process Frequency Percent Ad hoc 3 3,4 Part of commission 29 33,4 Initiated by product development department 41 47,1 Overall product development process 13 15,0 Dont know 1 1,1 Missing 0 0,0 Total 87 100,0 From table 8 we can see that 33,3 % of the IT projects were part of a commission and 47,1 % initiated by product development department. Few respondents (only 3,4 %) say that the project was a result of an ad hoc situation.
Discussion
This section presents our key findings related to the hypotheses, and discusses the implications.
Innovation index
To measure the association between IT-based business innovation and development method, an innovation index was constructed from the variables introduced earlier, consisting of the following weighted elements: Table 9. Innovation index Category Technology
Business innovation
Combinations of
Variable - Web technology - Java (or J2EE) - Microsoft.net - Web services - Peer-to-peer services - Other - New business model - Changed supply chain - Knowledge management solution - Customer Relation Management - No change of business process - New version of existing software
Weight 1 1 1 1 1 0 5 3 3 3 0 1
81
- New software 2 - New software and business process 3 improvement - New software and new business process 4 - New software and new business model 5 The elements were summarized for each respondent, giving each project an innovation index with a minimum score of 0 and a maximum score of 15. This variable is converted to an innovation level according to table below. technical and business change Table 10. The relation between Innovation index and Innovation level Innovation level low medium high Innovation index from to 0 3 4 7 8 15
The Innovation level variable was constructed with the aim of distributing the Innovation index with as close as possible to one third of the observations for each of the three levels (low, medium and high): thus creating a relative innovation level between the companies in the survey. This was cross-tabulated with three categories of development methods: Formal methods: Formal methods, such as RUP, XP and including methods from consultancies Adapted method: Adapted formal method + use of various of techniques No method: No use of method
Table 11: Innovation level and development method Develoment Method Formal Adapted No method Missing Total --------- Innovation level ---------Low Medium High 10 (21,3%) 9 (19,1%) 28 (59,6%) 11 (20,4%) 19 (35,2%) 24 (44,4%) 4 (21,1%) 5 (26,3%) 10 (52,6%) 31 (25,8%) 51 (42,5%) 38 (31,7%)
Total 47 (100%) 54 (100%) 19 (100%) 3 (100%) 123
As table 11 shows formal methods are associated with a medium degree of innovation, adapted methods with a high degree of innovation, and no use of methods is associated with a low degree of innovation.
82
At this stage it is reasonable to ask whether there is different innovation level related to in-house development and development for an external customer. Table 12 shows that there are no significant differences between in-house and external projects.
83
Table 12: Cross table of SW development category and innovation level Software development category In-house SW development or adaption/implementation of standard SW SW development or adaption/implementation of standard SW for customer Other Missing Total Our first hypothesis was: H1: A high degree of innovation is associated with a high degree of development structure, given high formal product development process. As seen in table 11 formal methods are associated with a medium degree of innovation. We also hypothesized a stronger association with innovation when the use of formal methods was part of an over all product development process. This was tested in table 13, and the result is negative. Thus, there is no support for hypothesis 1. Table 13. Formalisation of the product development process related to the innovation index. Formalisation of the product development process Ad hoc Part of commission Initiated by product development department Overall product development process Total Innovation index Mean 8,0 6,3 5,5 5,1 5,8 N 3 28 41 13 85 --------- Innovation level ---------Low Medium High 17 (26,6%) 27 (42,2%) 20 (31,2%)
Total 64 (100%)
10 (20,5%)
22 (44,7%)
17 (34,8%)
49 (100%)
4 (57,2%) 31 (25,8%)
2 (28,6%) 51 (42,5%)
1 (14,2%) 38 (31,7%)
7 (100%) 3 (100%) 123
As seen in table 13 the innovation index declines as formalisation of the product development process increases. Our alternative, second hypothesis was:
84
H2: A high degree of innovation is associated with a low degree of formal development process structure. As seen in table 11, most of the projects that use no method have a low degree of innovation. It is hard to see any traces of creative and innovative "amethodical" development. Thus, there is no support for this hypothesis.
Adapted methods, techniques and innovation

Although we found no support for either hypothesis, we have identified a relevant finding. As table 11 shows, there appears to be an association between the use of adapted methods and techniques, and project innovation. The innovative projects use a collection of techniques or customised development methods. How do we explain this finding? Unfortunately, our data do provide more information on this revelation. However, the literature offers at least two possible explanations. First, it may be explained from a resource based view (Andreu and Ciborra, 1996): An innovation is the result of tangible and intangible resources, performed - and often improvised - by clever people that possess a certain combination of skills and ideas. In an IT-based context this implies that innovators would use parts of development processes and techniques, and adapt them to the actual challenge. This is an important resource for innovations, because these skills are integrated as a part of the core competence in the firm (Barney 1991). In a way, this transfers the research question into knowledge management theory Second, it may be assumed that the opportunity to experiment and adapt methods is associated to smaller and younger firms, where formalism is less rigid. Table 15 gives some support to that assumption, in the sense that the younger the firm, the stronger is the innovation index of the project Table 15. Company age and innovation level Age of company Young < 5 years Medium 5-9 years Mature 10-14 years Old > 15 years Missing Total --------- Innovation level ---------Low Medium High 3 (14,3%) 8 (38,1%) 10 (47,6%) 8 (17,4%) 19 (41,3%) 19 (41,3%) 8 (32,0%) 12 (48,0%) 5 (20,0%) 12 (42,9%) 12 (42,9%) 4 (14,2%) 31 (25,2%) 51 (41,4%) 38 (30,8%)
Total 21 (100%) 46 (100%) 25 (100%) 28 (100%) 3 (100%) 123
Limitations
85
We acknowledge that there are limitations within this pilot study. Validity may be questioned because our sample - for practical reasons - includes only SW development companies. Since over half of the projects are in-house in these companies, it may be argued that our results are valid only for this population. There are, however, no significant differences in the degree of innovation between the internal and external projects, as shown in table 12. We also realize that our technology index may be questioned, both on its constructs and its weighting. The same applies to our grouping of systems development methods. However, the results are not significantly changed when these indexes are composed differently or given another weighting. Finally, the analysis of the innovative projects could have been improved by asking for more contextual information from the respondents.
Conclusion
This is a pilot study, investigating the relationship between IT-based business innovation and software development methods. In a quantitative survey 123 companies were asked to identify an important development project last year, and describe attributes of that project. We find that neither the top-down business planning approach, nor the emergent "amethodical" view is supported. Instead we find that the innovative projects are associated to the use of adapted development methods or various techniques. Innovative projects are neither initiated from top management nor improvised; they are usually initiated at department level. Further, they occur in young companies. We interpret these findings within a resource based innovation view: IT-based business innovation is best supported by local initiatives in a competent environment, using locally adapted techniques. Further research should collect more data on the nature of this local competence environment, and also of more contextual factors, like top management support and financing issues. Ideally, a quantitative study should have an international focus, with data from more nations. Another track could be a more qualitative study, concentrating on a few case projects to study IT-based innovation over time.
Acknowledgements
We thank the two anonymous Nokobit reviewers and Cecilia Haskins for providing useful comments to improve this article.
86
References
Andreu R. and Ciborra, C. (1996) Organisational learning and core capabilities development: The role of IT, Journal of Strategic Information Systems 5, pp.111-127. Barney, J (1991) Firm Resources and Sustained Competitive Advantage, Journal of Management, 17, 99-120 Bygstad, B., Fagerstrm, A., stensen, T. (2002). Vi bruker vr egen metode, og den fungerer utmerket. En underskelse av hvilke utviklingsmetoder som er i bruk i norske programvare utviklingsmiljer, Proceeding of Nokobit, Kongsberg. Cadle, J. and Yeates, D. (2004) Project Management for Information Systems, Prentice Hall. Carr, N. (2003) IT Doesn't Matter, Harvard Business Review, May. Ciborra, C. (2002). The Labyrinths of Information: Challenging the Wisdom of Systems. Oxford University Press Davenport, T. H. (1993). Process Innovation. Boston, Ernst & Young. Evans, N.D. (2003). Business Innovation and Disruptive Technology. Prentice-Hall. Fagerstrm, A., Bygstad, B., stensen, T. (2003) Adoption of Systems Development Methods and Market Orientation: An Empirical Investigation in Norway, Proceedings of NOKOBIT, Oslo. Fitzgerald, B. (1998 A). An empirical investigation into the adoption of systems development methodologies. Information & Management, Vol. 34, pp. 317-328. Jacobson, I., Booch, G. and Rumbaugh, R. (1999) The Unified Software Development Process, Addison Wesley, Reading. Larsen, T. J. (1998). "Information Systems Innovation: A Framework for Research and Practice". In Larsen and McGuire (eds.): Information Systems Innovations and Diffusion: Issues and Directions, Hershey, PA:,Idea Group Publishing, pp. 411-434. Laudon, C. K. and Laudon, J.P. (2004). Management Information Systems. Upper Saddle River, New Jersey, Pearson Education. Leonard-Barton, D. (1988). " Implementation as Mutual Adaptation of Technology and Organization." Research Policy Vol. 17(5), pp. 251-267.
87
Mokyr, J. (1992). The Lever of Riches: Technological Creativity and Economic Progress. Oxford University Press. Porter, M. E. 2001. "Strategy and the Internet". Harvard Business Review (March), pp.63-78. Rogers, E. M. (1995). Diffusion of Innovations, Fourth edition, Free Press. Stapleton, J. (2003) DSDM: Business Focused Development. Addison-Wesley. Tabrizi, B. and R. Walleigh (1997). "Defining Next-Generation Products: A Inside Look", Harvard Business Review, pp.117-124. Truex, D., Baskerville, R., and Travis, J. (2000). Amethodical Systems Development: The Deferred Meaning of Systems Development Methods, Accounting, Management and Information Technology, Vol.10, pp. 53 - 79.
88
Appendix - Questionnaire
89
90
91
92
The Role of Actual IS Utilization Level: Extending the Expectation-Confirmation Model of IS Continuance Intention
Tor J. Larsen, PhD Associate Professor, Department of Information Technology and Knowledge Management Norwegian School of Management, Norway, Tor.J.Larsen@bi.no
Anne Sreb, MSc in IS University College Teacher, Faculty of Education in Business Administration Buskerud University College, Norway, Anne.Sorebo@hibu.no
Abstract
The concept of actual utilization level is added to the original four constructs in the Expectation-confirmation model (ECM) of IS continuance intention. The rationale for the inclusion is that actual behavior may vary relative to feelings and perceptions. The analysis suggests that actual utilization level contributes to explaining IS continuance intention without reducing the impact of perceived usefulness and satisfaction.
ACM categories: J.1 Administrative Data Processing, J.4 Social and Behavioral Sciences (ECM), K.3.2 Computer Information Science Education, K.8.1 Application Packages.
93
Key words: Expectation-confirmation model of IS continuance, actual utilization level, e-learning
Introduction
Bhattacherjees (2001) key work on continuance of use makes two important contributions to our understanding of user acceptance of information systems (IS), when the use is voluntary. First, it presents a strong case for an in-depth examination of the separation between users initial acceptance and long-term acceptance of IS. The latter is a question of so-called IS-continuance; i.e. users demonstrate a willingness to take advantage of a new IS for the tasks it is designed to support beyond the period of firsttime use3. Second, it provides a theoretical framework, namely expectation-confirmation theory (ECT), as a basis for explaining IS continuance. A valid test of a theorys suitability requires proper replications, extensions, and generalizations (Rosenthal 1991; Tsang and Kwan 1999). Replications play an important role in the construction of IS knowledge (Berthon et al. 2002) and contribute to the building of a cumulative tradition in IS (Benbasat and Zmud 1999; Sambamurthy 2001). The Expectation-Confirmation Model of IS Continuance is highly recognized as a contribution toward theory building in IS but Bhattacherjee calls attention to the need for extending the current research (2001:366). The present study introduces user's actual usage patterns as an endogenous construct in addition to those originally proposed by Bhattacherjee (2001) because feelings and perceptions may not vary systematically with usage levels. The objective of the present research is to examine the influence of actual usage patterns on users IS continuance intention. We believe there are two significant reasons for investigating how actual usage patterns influence users willingness to continue to use a particular technology. First, continued use can only be studied when usage patterns already exists. We assume that these established usage patterns influence future IS use, at least in the domain of voluntary use. This is what Seddon (1997) label Meaning 2, i.e. use as a behavior or a pointer of future use. Also, usage patterns are related to the key concept of user satisfaction in the Expectation-Confirmation Model of IS Continuance (DeLone and McLean 1992, 2003; Gelderman 1998). For example, Delone and McLean (2003) describe usage as a positive experience with use, which in turn will lead to greater user satisfaction. It has also been argued that when use is mandatory, system use and usefulness are proxies (Seddon and Kiew 1996). In the Expectation-Confirmation Model of IS Continuance, usefulness is one of the two antecedents of satisfaction. Based on these arguments we conclude that there is adequate reason to investigate the impact of users actual use patterns on users IS continuance intentions.
3
The authors define first-time use as a context dependent construct. For example, in business-toconsumer electronic commerce first-time use refers to the first-time a user places an order and in elearning first-time use refers to the first-semester a teacher employs IS in a learning environment.
94
Research Model
Expectation-Confirmation Theory (ECT) was initially developed in the area of consumer behavior (Oliver 1980). Bhattacherjee (2001) extended ECT into the area of postacceptance of computer technology. The major transformation was changing the original mixed pre/post consumption aspect into a pure post-acceptance IS model (see Figure 1).
Figure 1. Bhattacherjees (2001) Expectation-Confirmation Model of IS Continuance The model builds on the assumption that users, after first-time acceptance and a period of initial use, form an opinion of the extent to which their pre-acceptance expectations are confirmed (e.g. Confirmation.) Simultaneous they form an opinion about the benefits (e.g. Perceived usefulness.) After a period of time the degree of confirmation and perceived usefulness are the basis for users perceived satisfaction with the IS (e.g. Satisfaction.) Finally, the perceived usefulness and satisfaction impact the willingness to continue to use the IS (e.g. IS continuance intention.) In this study, the extension made to the ECM model is actual (IS) utilization level, defined as an endogenous construct. It reflects IS functions that users take advantage of in the execution of specific job tasks and the frequency of use (of those functions.) We propose that Confirmation and Perceived usefulness impact Actual utilization level and that Actual utilization level impacts Satisfaction and IS continuance intention. The extended ECM model is shown in Figure 2.
95
Notes: Hypotheses 1-5 address relationships in the original ECM model, see Bhattacherjee (2001). Hypotheses 6-9 reflect new relationships because of the inclusion of Actual utilization level.
Figure 2. The extended Expectation-Confirmation Model of IS Continuance The arguments for the added hypotheses (numbers 6-9) are presented next. Users actual utilization level is hypothesized to be an antecedent of IS continuance intention. We believe that acceptance in the long run is not only a question of accepting an IS per se but first and foremost a question of accepting specific IS functions. Therefore, a high level of utilization of IS functions will strengthen users intention to continue their use. We propose the following hypothesis: H6: Actual utilization level is positively associated with IS continuance intention. An increased level of system use is predicted to lead to greater user satisfaction (DeLone and McLean 2003, Gelderman 1998). This assumption leads us to expect that when the actual utilization level increases satisfaction with the IS will also increase, hence: H7: Actual utilization level is positively associated with satisfaction. Seddon and Kiew (1996) argued that usefulness is an adequate proxy for tool-based use when use is mandatory. We investigate the opposite situation, i.e. use is voluntary. The proposition is that use is voluntary when use is not requested but a result of users own judgment of IS applicability in getting the job done. H8: Perceived usefulness is positively associated with actual utilization level. We infer that confirmation implies realization of the expected benefits of using an IS (Bhattacherjee 2001), which should result in a higher degree of use, hence:
96
H9: Confirmation is positively associated with actual utilization level.
Methods
Three university colleges, where an e-learning tool was made available to all faculty members, agreed to participate in the study. Since variables and items have been used previously and found reliable with acceptable validity, a questionnaire was developed as the vehicle for data collection (see Appendix A). The language of the questionnaire was Norwegian. English items were first translated into Norwegian and then back into English by a second person to ensure wording reliability. An early version of the instrument was presented to 10 prospective respondents. They where asked about their own and their coworkers present use of an e-learning application. Our test group next filled in a close-to-final version of the instrument without the researchers being present. Respondents were at this stage encouraged to write comments if items were found to be ambiguous or non-understandable. Improvements were made at each of these steps. The questionnaire was distributed, in the beginning of March 2003, through an internal mail system to all faculty members at the three colleges. Only faculty members who had taken advantage of an e-learning application for ordinary on-site courses were asked to respond. Respondents were instructed to respond only if they felt that they could at their own discretion decide to discontinue their use of the e-learning tool the subsequent semester. Returns were by ordinary mail. By the end of March 2003, 135 usable questionnaires were returned, for a response rate of 33 percent. The reasons why the response rate was somewhat low may be that not all the faculty members use e-learning in ordinary on-site courses and that some of the faculty members work within distance teaching only.
Thirty-two percent of the respondents were women and sixty-eight percent were men, reflecting the true gender distribution at the three university colleges. The average respondent was 45 years old (1% below thirty, 23% in the thirties, 51% between 40 and 54, and 25% above fifty-five), held a master degree, and had 15 years of experience in using computers. The recommended two step procedure for checking item data quality measurement before hypothesis and relationship testing was followed (Anderson and Gerbing, 1988). Items were checked for skewness and kurtosis and were found to have unproblematic univariate distributional characteristics (i.e. values below 2.0). In our sample 104 questionnaires contained no missing values, 29 questionnaires had between 1 and 9 missing values, and 2 questionnaires had more than 9 missing values.
Data Analysis
97
Because of its recent extensive use, we employed Partial Least Squares (PLS Graph, version 3.0) to analyze the data (Compeau and Higgins 1995, Plouffe et al. 2001, Wixom and Watson 2001). PLS has several strengths that makes it appropriate for our study, including its ability to handle complex predictive models and formative constructs. The analysis proceeds with measurement model results and structural model results.
MEASUREMENT MODEL RESULTS

The instrument for measuring actual utilization level was developed by the authors, and consists of formative or causal indicators. This means that the indicators determine the latent variable; i.e. utilization of each individual IS function determines the actual utilization level. Therefore, this instrument cannot be tested for internal consistency and reliability - since we cannot expect that indicators are correlated (Bollen and Lennox 1991). However, the literature suggests that formative constructs can be investigated for validity by evaluating two different conditions. First, psychometricans argue that formative constructs require a census of all the indicators that form the entire construct (Jarvis et al. 1993). Second, Bollen and Lennox (1991:312), state that: causal indicators are not invalidated by low internal consistency so to assess validity we need to examine other variables that are effects of the latent construct. Adhering to these conditions means paying attention to (1) census of the included items and (2) nomological validity (Jarvis et al. 2003). The criterion of census was addressed in questionnaire pre-test procedures. The participants were invited to evaluate the 8 items in the proposed instrument and propose other functions in the e-learning application that were not covered by these items. The eight items in our instrument covered the functionality that would be used in on-site courses. Nomological validity is addressed in the PLS analysis when testing the explanatory power of the construct of actual utilization level within the Expectation-Confirmation Model of IS Continuance. The other four ECM constructs (i.e. perceived usefulness, confirmation, satisfaction and IS continuance intentions) are reflexive. The adequacy of these reflexive constructs can be determined by looking at: (1) individual item reliabilities, (2) the convergent validities of measures associated with individual constructs, and (3) discriminant validity between constructs (Hulland 1999). All the items, except one item representing the IS continuance construct, have loadings close to 0.7 or above (see Table 1). Item two in the IS continuance intention scale was deleted due to item reliability below recommended threshold of 0.50 (i.e. o.582 = 0.34). Table 1 shows items, means, and standard deviations for the five constructs in the extended ECM model. For each construct the assessment of convergent validity or internal consistency (Fornell and Larcker 1981) is also included. The calculated value for internal consistency should exceed 0.70 (Nunnally 1978). The criterion is met for all five construct. 98
Perceived usefulness (Composite reliability = 0.92) Reflective Mean Using e-learning improves my teaching quality. 4,42 Using e-learning makes my students more 4,55 satisfied with my teaching. Using e-learning increases the control with my 4,23 teaching plans. 3,43 Using e-learning does it possible to do more
Std.Dev. 1,65 1,51 1,92 1,96 1,80 1,66 Std.Dev. 1,37 1,52 1,76
Loading 0.84 0.77 0.83 0.73 0.82 0.87 Weight 0.39 0.06 0.48
t-Stat 18.75 11.59 18.99 10.61 14.91 22.15 t-Stat 1.42 0.31 2.20
teaching than without. Using e-learning makes it easier to do my job as a teacher. e-learning is a useful tool in my teaching.
Actual utilization level Formative I make information about the courses available (e.g. plan of progress, literature references, etc.) On continuous basis I make necessary messages about e.g. changes of classroom, deadlines, etc. I make educational material available (e.g. documents, presentations, proposal for solutions, etc.) I make Multiple Choice or other tests for my students I utilize the electronic mail function for sending mail to my students (e.g. group mail or single mail) I receive reports from projects and so on I establish and utilizes discussion forums for my students I establish closed students groups (e.g. in connection with the students project work) Confirmation (Composite reliability = 0.94) Reflective My experience with using e-learning was better than what I expected. The service level provided by e-learning was better than what I expected. Overall, most of my expectations from using elearning were confirmed. Satisfaction (Composite reliability = 0.88) Reflective Based on my experience with e-learning, I am very contented with using the system. Based on my experience with e-learning, I am
4,16 5,01 Mean 6,17 5,96 5,76
2,03 4,96
1,76 2,09
0.24 0.29
1.49 1.47
4,36 2,97 3,38
2,29 1,97 2,35
0.14 0.07 0.09
0.79 0.27 0.37
Mean 4,30 4,06 4,49
Std.Dev. 1,51 1,44 1,41
Loading 0.91 0.93 0.90
t-Stat 31.81 44.91 22.11
Mean 4,71 5,31
Std.Dev. 1,35 1,51
Loading 0.93 0.90
t-Stat 57.81 34.20
99
very dissatisfied with using the system (reverse coded). Based on my experience with e-learning, I am 3,10 delighted with the system. IS continuance intentions (Composite reliability = 0.84) Reflective Mean I intend to continue using e-learning rather than 6,16 discontinue its use. My intentions are to continue using e-learning 6,22 rather than use any alternative means. If I could, I would like to discontinue my use of 5,92 e-learning (reverse coded). My intentions are to continue using e-learning, 5,96 at least as active as today, in the future
1,51
0.67
6.45
Std.Dev. 1,31 1,24 1,91 1,32
Loading 0.87 0.58 0.67 0.84
t-Stat 19.24 3.01 5.48 16.43
Table 1. Item Means, Standard Deviations, and Internal Consistencies Fornell and Larcker (1981) suggests that discriminant validity among constructs should be based on the average variance extracted (AVE), which is the average variance shared between a construct and its measures. As Table 2 shows, the square root of AVE values is consistently greater than the off-diagonal correlations, suggesting satisfactorily discriminant validity among constructs. Construct 1. Perceived usefullness 2. Confirmation 3. Satisfaction 4. IS continuance Intentions 1 0.82 0.60 0.59 0.51 2 0.92 0.79 0.48 3 5
0.85 0.64
0.81
Table 2. Squared Correlations Among Constructs & Average Variance Extracted
STRUCTURAL MODEL RESULTS

Figure 3 summarizes the structural model results. Standarized regression coefficients are shown above each path with appurtenant t-values in brackets. R2 is shown above each endogenous latent construct.
100
Notes: Stars indicate significance level: ***(p<.001). Hypothesis number is in bold adjacent to each path.
Figure 3. LISREL Analysis of the Extended ECM Model Five out of nine path coefficients have positive associations with dependent constructs. We conclude that hypotheses 1, 2, 5, 6 and 9 are supported. The structural model analysis documents acceptable levels of explained variance for IS continuance intentions (R2 = 62%), actual utilization level (R2 = 36%), satisfaction (R2 = 65%) and perceived usefulness (R 2 = 36%).
Discussion, implications and limitations

In our Extended Expectation-Confirmation Model of IS Continuance, we found support for five out of nine hypotheses. The results suggest that in addition to the already established role of satisfaction, utilization level contributes to explaining IS continuance intention. Our added construct demonstrates an adequate level of nomological validity. The concrete use of an IS and the perception of its benefit both play a role. The theoretical implication is that our extended ECM model has merit. In their interaction with users, IS practitioners should encourage a comprehensive level of utilization while simultaneously stimulate positive feelings toward the technology.
101
The lack of a significant relationship between actual utilization level and satisfaction is surprising. Prior research documents that the feeling should be related to behavior when use is voluntary (Winter et al., 1998). We interpret that there might exist two different and independent routes to IS continuance intention; one instrumental and one based on feelings. The instrumental route is the path confirmation -> perceived usefulness -> actual utilization level -> IS continuance intention. The feeling route is the path confirmation -> satisfaction -> IS continuance intention. Both routes have their origin in fulfilled expectations (i.e. confirmation). However, they represent different explanations for IS continuances intention. We suggest that users with a low actual utilization level can be highly satisfied (with the IS) and may therefore also exhibit a high level of IS continuance intention. Conversely, users with a high utilization level may be dissatisfied with the IS, but they may not perceive discontinuance as an alternative. If so, the score on IS continuance intention will be high. Further research is needed to confirm these assumptions. The present research has limitations. First, we measured IS continuance intention but not actual IS continuance level the following semester. The benefit of longitudinal data parallels the proverb: the proof is in the eating of the pudding. Besides, published research documents that an individual decision to utilize a technology is a conscious act, which can be sufficiently explained and predicted through behavioural intentions (Sheppard et al. 1988; Chau and Hu 2002). In conclusion, IS continuance intention is an adequate proxy for actual IS continuance behaviour. Second, the use of a cross-sectional survey data has limitations. This data collection approach does not allow explicit directional tests. This does not mean that the causal relationships we have specified in Figure 3 are void. The theoretical arguments for the ECM model provide support for the existence of causal relationships among constructs but also structural equation analysis. In spite of the support for causal relationships in survey research, conclusive statements about causality cannot be made since alternative explanations cannot be ruled out. Longitudinal design is the approach that should be employed to settle the issue. The response rate at 33% is not the best, even though it falls in the acceptable range. An earlier report from one of the participating colleges, suggested that 47 percent of the faculty were engaged in e-learning for on-site courses in 2001. Assuming an equal level of e-learning application use in the two other colleges our response rate is actually 70 percent of faculty that are active e-learning application users. Although more faculty members may have taken e-learning into use since 2001, we conclude that the response rate is acceptable. The major concern is that faculty members being active e-learning application users were more strongly inclined to respond to the questionnaire.
102
References
Anderson, J.C. & Gerbing D.W. (1988). "Structural equation modeling in practice: A review and recommended two-step approach," Psychological Bulletin, Vol. 103, No. 3, pp. 411-423. Benbasat, I. and Zmud, R.W. (1999). Empirical Research in Information Systems: The practice of relevance, MIS Quarterly, Vol. 23, No. 1, pp. 3-16. Berthon, P., Pitt, L., Ewing, M., and Carr, C.L. (2002). Potential research space in MIS: A framework for envision and evaluationg research replication, extension and generalization, Information Systems Research, Vol. 13, No. 4, pp. 416-427. Bhattacherjee, A. (2001). Understanding information systems continuance: An expectation-confirmation model, MIS Quarterly, Vol. 25, No. 3, pp. 351-370. Bollen, K.A., and Lennox, R. (1991). Conventional Wisdom on Measurement: A Structural Equation Perspective, Psychological Bulletin, Vol. 110, No. 2, pp. 305314. Chau, P.Y.K and Hu, P.J. (2002). Examining a Model of Information Technology Acceptance by Individual Professionals: An Exploratory Study, Journal of Management Information Systems, Vol. 18, No. 4, pp. 191-229. Compeau, D. R. and Higgins, C. A. (1995). Computer Self-Efficacy: Development of a Measure and Initial Test, MIS Quarterly, Vol. 19, No. 2, pp. 189-211. DeLone, W.H., and McLean, E.R. (1992). Information systems success: The quest for the dependent variable, Information Systems Research, Vol. 3, No. 1, pp. 6095. DeLone, W.H., and McLean, E.R. (2003). The Delone and McLean Model of information systems success: A ten year update, Journal of Management Information Systems, Vol. 19, No. 4, pp. 930. Fornell, C. and Larcker, D. F. (1981). Evaluating Structural Equation Models with Unobservable Variables and Measurement Error, Journal of Marketing Research, Vol. 18, February, pp. 39-50. Gelderman, M. (1998). The relation between user satisfaction, usage of information systems and performance, Information & Management, Vol. 34, No. 1, pp. 11-18.
103
Hulland, J. (1999). Use of Partial Least Squares, (PLS) in Strategic Management Research: A Review of Four Recent Studies, Strategic Management Journal, Vol. 20, No. 2, pp. 195-204. Jarvis, C.B., MacKenzie, S.B., and Podsakoff, P.M. (2003). A Critical Review of Construct Indicators and Measurement Model Misspecification in Marketing and Consumer Research, Journal of Consumer Research, Vol. 30, September, pp. 199218. Nunnally, J. C. (1978). Psychometric Theory, McGraw Hill, New York. Plouffe, C. R., Hulland J. S. and Vandenbosch M. (2001). Research Report: Richness Versus Parsimony in Modeling Technology Adoption Decisions Understanding Merchant Adoption of a Smart Card-Based Payment System, Information Systems Research, Vol. 12, No. 2, pp. 208-222. Rosenthal, R. (1991). Replication in behavioural Research, in Replication Research in the Social Sciences, J.W. Neuliep (ed.), Sage, Newbury Park, CA, pp. 1-30. Oliver, R.L. (1980). A cognitive model of the antecedents and consequences of satisfaction decisions, Journal of Marketing Research, Vol. 17, November, pp. 460469. Tsang, E.W.K., and Kwan, K.M. (1999). Replication and Theory Development in Organizational Science: A critical realist perspective, Academy of Management Review, Vol. 24, No. 4, pp. 759-780. Sambamurthy, V. Research in information systems: What we havent learned, MIS Quarterly, Vol. 25, No. 4, pp. 10-11. Seddon, P.B., and Kiew, M.Y. (1996). "A Partial Test and Development of DeLone and McLean's Model of IS Success, Australian Journal of Information Systems, September, pp. 90-109. Sheppard, B.H., Hartwick, J. and Warshaw, P.R. (1988). The Theory of Reasoned Action: A Meta Analysis of Past Research with Recommendations for Modifications and Future Research, Journal of Consumer Research, Vol. 15, No. 3, pp. 325-343. Winter, S. J., Chudoba, K. M. and Gutek, B. A. (1998). Attitude Toward Computers: When do They Predict Computer Use?, Information & Management, Vol. 34, pp. 275-284. Wixom, B. H. and Watson, H. J. (2001). An Empirical Investigation of the Factors Affecting Data Warehousing Success, MIS Quarterly, Vol. 25, No. 1, pp. 17-41.
104
105
Appendix A: Questionnaire Items

Perceived usefulness using a Likert type scale ranging from 1 (strongly disagree) to 7 (strongly agree) Using e-learning improves my teaching quality. Using e-learning makes my students more satisfied with my teaching. Using e-learning increases the control with my teaching plans.
Using e-learning does it possible to do more teaching than without. Using e-learning makes it easier to do my job as a teacher. E-learning is a useful tool in my teaching.
Actual utilization level using a Likert type scale ranging from 1 to 7 (1 = never, 4 = sporadic and 7 = actively) Through e-learning: I make information about the courses available (e.g. plan of progress, literature references, etc.) I make necessary messages about e.g. changes of classroom, deadlines, etc. available I make educational material available (e.g. documents, presentations, proposal for solutions, etc.) I make Multiple Choice or other tests for my students I utilize the electronic mail function for sending mail to my students (e.g. group mail or single mail) I receive reports from projects and so on I establish and utilize discussion forums for my students I establish closed students groups (e.g. in connection with the students project work) Confirmation using a Likert type scale ranging from 1 (strongly disagree) to 7 (strongly agree) My experience with using e-learning was better than what I expected. The service level provided by e-learning was better than what I expected. Overall, most of my expectations from using e-learning were confirmed. Satisfaction using a Likert type scale ranging from 1 (strongly disagree) to 7 (strongly agree) Based on my experience with e-learning, I am very contented with using the system. Based on my experience with e-learning, I am very dissatisfied with using the system (reverse coded). Based on my experience with e-learning, I am delighted with the system. IS continuance intentions using a Likert type scale ranging from 1 (strongly disagree) to 7 (strongly
agree)
I intend to continue using e-learning rather than discontinue its use. My intentions are to continue using e-learning rather than use any alternative means. If I could, I would like to discontinue my use of e-learning (reverse coded). My intentions are to continue using e-learning, at least as active as today, in the future.
106
IS Research and Issues of Collaboration
SESSION 3: IS Research and Issues of Collaboration

This track concerns both research on collaboration within the IS community and research on IT/IS for supporting collaboration (CSCW). The first paper, by Molka-Danielsen empirically explores the relation between the number of registered research works and the level of collaboration among researchers. In the second paper of this track, Skattr et al. develops a framework for understanding mobile computing work among non-office workers as for instance nurses. The third paper, looks at ICT and organizational change initiatives in the health care sector in Norway. The authors conclude that ICT-related change projects in this context are typically constrained by existing institutional practices, information infrastructures, and regulations in terms of legislation and locally defined rules and procedures
1. "Collaborative Research Network: Case Study at Molde University College" by Molka-Danielsen 2. "Mobile work - Mobile ICT Supporting Secondary Work" by Skattr, Hasvold, Berntzen, and Engvig 3. "ICT-related change in complex organisations: The role of infrastructural and regulatory constraints" by Aanestad, M & Boulus
Collaborative Research Network: Case Study at Molde University College

Judith Molka-Danielsen Molde University College P.O. 2110, N-6402 Molde, Norway Email: J.Molka-Danielsen@himolde.no Web: http://home.himolde.no/~molka
Abstract
The co-authorship network of faculty at Molde University College is an example of an evolving social network. By querying the research database FORSKDOK of the BIBSYS library resources system we have access to the history of registered works and the link structure of this network. While many academic institutions are experiencing lower student enrollment, we see it as increasingly important to be able to demonstrate and improve our productivity in the workplace. A question worth exploring is, does working together increase the quantity of publications produced and can this be demonstrated in our network of researchers? This case study examines network properties and the correlation between the number of registered research works and the level of collaboration among researchers at Molde University College.
Introduction
Clusters of associations and even friendships can foster a sharing of ideas. Sometimes these exchanges lead to the co-production of scholarly works. While some may study how associations are formed,4 the primary focus of this paper is to study whether these
Another study that measured collaboration made use ofur@q s numberQhy@q mathematician that published over 1500 papervu$&phu@q uhh@q
hhCthvh i
er of zero. A
108
associations are beneficial to production of the researchers. This study examines one social network of researchers at Molde University College. The questions examined are: how much collaboration takes place within small faculties and how much between departments of the school? Do faculty members that collaborate have higher quantities of publications?5 More specifically, is there is an association between the total number of publications of a researcher and the number of unique co-author links (that is unique person-to-person links). This is examined using a cluster coefficient. (Watts and Strogatz, 1998) We next compare the connectivity of the network by looking at who is part of the largest cluster. Finally, based on our network data, we discuss the meaning of collaborative efforts for researchers.
Certain factors of this study must be clarified. The data for this study was collected on August 27, 2004. This study does not look at the evolution of collaborations over time, because we would need to look at not only when works were published but who was on faculty in a given year.6 This goes beyond the scope of this report, but we hope to examine this in a later study. In this study, we look at the association between collaborations and publications in the static network.
Second, this study does not attempt, and should not be seen as a measure of individual faculty productivity. Individual faculty members have different roles and responsibilities, and budget different amounts of working time towards publishable research. Comparing absolute quantities between faculty members cannot be used to draw conclusions.
co-huuhh@q
irs
6p
-author to a co-huuhh@q
irshq
This study confirmed that co-authors were not chosen randomly.(Grossman, 2003) 5 We note and explain later that quantities of publications cannot be compared between individual faculty members. 6 A similar survey of the FORSKDOK database was made on May 14, 2003. Data found in (MolkaDanielsen, 2004) Since that time the number of researchers or nodes has changed. At the time of the 2003 query there were 73 researchers. Since then the list of faculty has changed. 6 were removed from the list and 8 had joined. The set of nodes for this study is based on the 75 researchers that were listed in August 27, 2004 and all registered works up that that date.
Hrhruhruuhhurhvpvhuhrhy@q irhq'
109
Third, there may be a difference in reported numbers of works and real quantities produced. Although faculty members are encouraged to register their works, there are many reasons why these works are not registered. Some faculty members register work themselves and others ask the library staff to do this, or let them determine when to register a work. But, for works that are registered in FORSKDOK, the process of registration is timely and the database has been in use for several years.
Research Objective and Method

Recent years have witnessed many studies of the small world phenomenon.(Watts and Strogatz, 1998)(Watts, 1999) This phenomenon states that average distances across networks can be very small, that one must traverse only a few links on average to go between any two nodes of the network.(Albert, Jeong and Barabsi 1999) That is all of the members of the network are on average, closely connected. Why do some networks exhibit this phenomenon? Small world networks have been characterized to have a few members (or nodes) with many connections to other members. The popular members are called hubs. We see hubs in technology networks also, such as routers on the Internet. The routers (nodes) are connected by transmission lines (links). But, hubs exist also in certain social networks, such as networks of researchers connected by collaborative works. Is the researcher network at Molde University College a small world network? The data here will indicate that it is. The research objective of this study is to outline the properties or characteristics of this social research network, and to understand the meaning of these characteristics for the members. The characteristics that we will study are: the number of nodes, the number of links, the cluster coefficient of links, the distance across the network, and what percentage of nodes are included in the cluster. Finally, we will look at what these characteristics mean to the researchers.
110
LINK DATA
On the date, 27th August 2004, the FORSKDOK database of BIBSYS was queried. The number of publications for each researcher at Molde University College was recorded, along with the number of author-to-author collaborations. Here we define publications as the term for a research work. This can include any written or verbal category of research that is included in the Molde University College research reports and any specific work that has been recorded in FORSKDOK of the BIBSYS system.
A count was made of the number of unique co-author collaborations. A unique collaboration is with a unique person pair, A-to-B. The unique co-author collaboration pair is denoted as a link in the network. One pair of researchers can have several works together, but the repeated collaborations are not a new link in the association network. The unique co-author collaborations are of interest to study the connected network.
MEMBER NODES DATA

Member nodes were selected from academic faculty members with a research record in FORSKDOK. There were 75 researchers or member nodes used in this study. The member node data are based on the data accessed on August 27, 2004. However, as stated above, faculty can have additional publications that are not listed in the FORSKDOK database. For example, here are listed several conditions where faculty work is not recorded in FORSKDOK:
Faculty with guest positions can have careers outside of Norway. They may have
registered publication that only relate to our in-house faculty. In-house faculty can have publication records before entering Molde University
College. If the faculty member worked previously in another country, it is likely that these former publications are often not recorded in the database. But, it is also possible
111
for new faculty coming from other institutions within Norway to not have work registered in the database.
In-house faculty may choose to not record publications or other research activity.
Even though there are categories for every public -clearing of the throat, some of these activities will still go unregistered. Work that was registered after 27th August 2004 is not shown in this study. New
collaborations and links are constantly being added.
Findings and Comparisons

The following table summarizes the data of the total number of publications per researcher, the number of unique co-authors, the total number of co-author collaborations and the average number of co-authors per publications. The average number of coauthors per publication is the author working by themselves plus this number. In the query, the average number of publications per researcher is 20.6. The average number of unique associations per researcher is 7.6. The average number of works done with a co-author is 21.83. And so, the average number of coauthors on each work is 1.08. The data for the query is shown in Table 1. Table 1 Publications and Collaborations
# Unique co-authors 17 0 1 5 7 4 22 6 2 1 4 0 # Total co-authors 55 0 2 5 14 6 57 13 3 1 10 0 Avg # coauthors per publication 1,77 0,00 1,00 2,50 1,27 1,50 1,90 1,18 0,06 1,00 1,43 0,00
Researcher Aarseth Aass Arntzen Bakken Bedrings Berg Berge (J) Berland Bjrkly B Borgen Botslangen
#Publications 31 2 2 2 11 4 30 11 52 1 7 1
112
Brthen (J) Braute Brekk Bremnes Buvik Crainic Dahl Danielsen Dauzre-Prs Dedekam jr. Foss Gammelster (J) Gjerde Gribkovskaia Guvg Halskau Hauge K-W Hauge Olav Haugen (J) Hervik (J) Hjelle Hvattum Hveberg Jger Jkupsstovu Jrgensen Jrnsten Kamsvg Larsen (J) Lium Lbersli Lkketangen Lvlien (J) Lykkeslet Michaelsen MolkaDanielsen Nordhaug Odeck degrd Atle Ohren Olsen Olstad (J)
88 4 5 12 69 2 2 15 2 6 13 80 9 9 5 30 22 9 39 154 26 6 1 26 16 3 98 2 20 3 2 77 4 5 4 25 1 19 2 15 47 7
29 1 3 8 12 2 0 3 2 1 6 32 10 4 1 9 8 7 20 52 8 3 1 7 4 1 30 0 12 4 0 32 0 1 4 5 0 10 2 3 18 9
115 2 3 18 42 4 0 14 2 1 10 82 26 10 4 19 16 9 69 205 27 12 1 28 9 1 85 0 14 6 0 90 0 1 7 28 0 29 2 16 53 16
1,31 0,50 0,60 1,50 0,61 2,00 0,00 0,93 1,00 0,17 0,77 1,03 2,89 1,11 0,80 0,63 0,73 1,00 1,77 1,33 1,04 2,00 1,00 1,08 0,56 0,33 0,87 0,00 0,70 2,00 0,00 1,17 0,00 0,20 1,75 1,12 0,00 1,53 1,00 1,07 1,13 2,29
113
Opdal Ory Rasmussen Rekdal Risnes Rnhovde Rudi Stre Sttem Skrondal Solibakke Spring Standal Stavik Sundal Tipper Tornes Vatne Wallace AK Wallace SW Woodruff sum 1-75 average standard deviation
3 2 2 2 14 11 3 17 4 2 19 1 1 4 3 5 28 20 10 237 17 1543 20,57 36,67
4 0 0 7 2 10 0 3 0 0 0 4 3 5 2 3 13 8 6 64 6 573 7,64 11,34
8 0 0 7 2 24 0 20 0 0 0 4 3 8 2 7 19 9 11 241 30 1637 21,83 41,07
2,67 0,00 0,00 3,50 0,14 2,18 0,00 1,18 0,00 0,00 0,00 4,00 3,00 2,00 0,67 1,40 0,68 0,45 1,10 1,02 1,76 80,85 1,08 0,88
The data in Table 17 would characterize the researcher network at MC as a small world network. This is demonstrated in Figure 1 and explained as follows. The number of the number of unique person-to-person links follows an Exponential distribution. In other words, there are a few authors with many different collaborators.
The (J) in Table 1 indicates the person has a joint appointment with another research institution.
114
Unique Co-authors Distribution 14 12 10 8 6 4 2 0 0 20 40 60 80 #Unique Co-author Links
Figure 1. Unique co-author link distribution researchers represented individually Also, as shown in Figure 1, the degree distribution follows the shape of a Scale Free Network rather than a Random network. The degree distribution of random networks follows a bell shaped curve. So, if one plotted the number of nodes with links versus the number of unique links there would be a bell shaped distribution. That would mean that most nodes have the same number of links and that nodes with very large number of links do not exist. An example of a random network is a national highway system where the nodes are the cities and the links are the highways connecting them. Most cities have about the same number of highways connecting them and the establishment of a new link
vryrprqhqyShqrxurhvqprqiQhy@q hq6yspq Spvv ($(@q , 1959)

Another kind of network is called the scale free network. This follows a power law distribution such that most nodes have only a few links, while a few nodes have many
links. These nodes with many links are called hubs. This is similar to an air traffic system where many small airports are connected by a few hubs. Social networks have also been 115
#Researchers - with exactly this count
found to follow a scale free distribution. For example, a study was made by Malcolm Gladwell to measure how social a person is.8(Gladwell, 2000) His conclusion was there are a few people that have a knack for making friends and acquaintances and these are connectors. Similarly, our network of researchers appears to contain a few connectors. As reported in Table 1, the standard deviation from the average number of unique coauthors is broad. Another characterization of networks is that they change or evolve over time.9 (Barabsi, Jeong, Nda, Ravasz, Schubert and Vissek, 2002) Presently, the MC researcher network has some demonstrable hubs of researchers with many connections. But, the number of hubs may increase over time. For example, in another study by Albert-Lszl Barabsi of the Internet and the www network of web pages, he noted that if the web followed the scale free model, that there would be only a few web pages with very many links in and out, only a few hubs. (Albert and Barabsi, 2002) His data showed however that the occurrence of hubs on the web is not so rare and that new hubs can come into existence. Nevertheless, while there are a growing number of hub web pages such as the Google web page, he found the distribution of number of links (in and out) of web pages still follows a power law. (Barabsi, 2002) One of his conclusions is that the number of connections or links in new networks often follow a power law distribution, and in more mature stable networks follow a normal distribution. In our network, we may also have researchers that will change into connectors or hubs over time. That is, they may produce more publications and create more connections over time. The creation of hubs usually arises out a differentiating factor. Such as, the Google web site has a good search engine.
He surveyed college students at City College of Manhattan. He asked how many surnames they knew from a list of 248 names from a Manhattan telephone book. He later tested other control groups of people of similar age, education and income. He found the range to be wide. The range was from 2 to 95 for the student group, and 16 to 108 for an older group. 9 In future studies, we hope to study how our network model changes over time.
116
COLLABORATIONS WITHIN
Next we examined the level of collaboration within departments and between departments at MC. This is measured using a clustering coefficient (C). In a fully connected network the Total Number of Possible Links = N(N-1)/2 where N is the number of nodes. In this case N=75 researchers. The clustering coefficient is the number of existing links divided by the number of possible links.10 (Watts, 1998) The closer that the value C is to 1.0, the more likely it is that there are ties between the nodes. The cluster coefficients for our network are reported in Tables 2 to 5. In all cases the collaboration is greater than for a random network with the equivalent number of nodes and links. Once again, a link is a unique person to person collaboration between members of the 75 node network. Not included in this calculation are co-author links where the coauthor is not a member of the 75 researcher set. The following table and figures summarizes the findings. The number of internal links is the number of co-author links within a department, and in the final row the MC total this is the number of unique coauthor links between departments. We can see in Table 2 that the Logistics group (although not their own department, this group works closely and it is important to examine this as a cluster) is most collaborative. Logistics is followed by Social Sciences, Economics, Informatics and Health departments respectively. As a whole, the entire college, across all institutions does not show much collaboration. The findings within departments may reflect some of the traditions in the fields of study. For example, it may be the tradition for researchers in Health Sciences to work individually on research topics. We also look at other groupings in Table 3, such as the Economics and Logistics group as a whole and the Logistics group with the Informatics group. We see that Informatics & Logistics grouping has a higher cluster coefficient than Informatics alone. This shows a larger number of associations between these groups. But, the number of links within a group should normally be higher than the number of links between groups. This is also a method used to identify communities on the web. One study identified that
10
The Cluster Coefficient can also be written as C=2E/N(N-1) where E is the number of edges present and N is the number of nodes present. The cluster coefficient for a random network Cr=<k>/N where <k> is the average degree of the network of size N nodes. <k> is therefore the average number of links per node.
117
documents belong to the same community if they have more links to each other than to documents outside the community. (Flake, Lawrence and Giles, 2000) Using the same definition would indicate that it is difficult to identify the boundaries of the groups into groupings of collaborative interests even if one begins with the facts that persons are a priori associated with different departments. Figures 5-10 demonstrate the network links of co-authors can be found in Appendix. Table 2 Cluster Coefficients all researchers
# nodes (researchers) 14 21 20 10 10 75 # internal existing links 13 15 9 14 11 98 # possible links 91 210 190 45 45 2775 Cluster Coefficient 0,1429 0,0714 0,0474 0,3111 0,2444 0,0353 Cluster Coefficient Random = <k>/N 0,0663 0,0340 0,0225 0,1400 0,1100 0,0174
Department Economics Informatics Health Logistics Social Sciences MC total
Table 3 Cluster Coefficients other groupings - all researchers

Department Eco&Logis Info&Logis # nodes (researchers) 24 31 # internal existing links 32 40 # possible links 276 465 Cluster Coefficient 0,1159 0,0860 Cluster Coef. Random 0,0556 0,0416
Table 4 Cluster Coefficients of the connected network

# internal existing links 13 15 9 14 11 98 # possible links 36 153 21 36 28 1275 Cluster Coefficient for the Connected Network 0,3611 0,0980 0,4286 0,3889 0,3929 0,0769 %Nodes researcher s that are connected 64,29 % 85,71 % 35,00 % 90,00 % 80,00 % 68,00 % Cluster Coefficient Random = <k>/N 0,1605 0,0463 0,1837 0,1728 0,1719 0,0377
Department Economics Informatics Health Logistics Social Sciences MC total
# nodes researchers 9 18 7 9 8 51
118
Table 5 Cluster Coefficients other groupings - of the connected network

Cluster Coefficient for the Connected Network 0,2092 0,1140 Cluster Coefficient Random = <k>/N 0,0988 0,0549
Department Eco&Logis Info&Logis
# nodes (researchers) 18 27
# internal existing links 32 40
# possible links 153 351
In Tables 4 and 5 we examine the clustering coefficient for the network where the node set is only the nodes included in the large cluster of the 75 nodes. In total 51 of the 75 researchers are part of the largest cluster, that is 68%. In summary, not many in the Health department work together. There only 35% are part of the large cluster. But, of those that are part of this cluster, they have the highest clustering coefficient of 43%. In comparison, the Informatics department has nearly 86% of the researchers connected to the large cluster, but has the lowest clustering coefficient of 10% for the connected network. Also, the Informatics department is the only group with 2 sub-clusters that are connected to the greater cluster but separate from each other. As a result, an example of the longest path on the connected network would be between Borgen and Arntzen with a path of 7 links.
COLLABORATIONS AND PUBLICATIONS

Is there an association between the amount of collaboration and the number of works produced by researchers of this network? Figure 2 shows that working with a high number of unique co-authors are associated with a high number of publications. Figure 3 shows that the average number of co-authors per article is not necessarily associated with a high total number of works produced. We may conclude from these data that it helps to have many associations, but it does not help to have many names on one article. The data in these figures are based on the data in Table 1.
119
Unique Co-Authors vs. Publications

70 60 50 40 30 20 10 0 0 50 100 150 200 250 Number of Publications # Unique Co-Authors
Figure 2. Unique Co-Authors versus Publications
Average #Co-Authors vs # Publication

4,50 4,00 3,50 3,00 2,50 2,00 1,50 1,00 0,50 0,00 0 50 100 150 200 250 Number of Publications Average # of Co-Authors
Figure 3. Average # of Co-Authors versus Publications
Next, we examined if there an association between the average number of publications per faculty member within the department and the clustering coefficient by department.
120
Clustering Coefficinet vs. Average Production

Average #Pubilications/#faculty 30 25 20 15 10 5 0 0 0,05 0,1 0,15 0,2 0,25 0,3 0,35 Clustering Coefficients H I S E L
Figure 4. Average # of Co-Authors versus Publications Figure 4 displays this relationship and uses the clustering coefficient in Table 2. However, we cannot conclude by this comparison that there is an association. And, we cannot exclude the association either, because as stated previously, the departments are composed of different groups of faculty positions where some positions are suppose to do more teaching, while others do more research. We said previously that we could not compare productivity between individuals, and it is also invalid to compare between departments.
Last we examine if a greater ratio of external associations contributes to a higher number of publications. A study by Newman found the probability of acquiring new collaborations increases with the number of past collaborations.11(Newman, 2001) We do not project forward, but look the association of the present count to past collaborations. Table 6 shows that all departments are collaborating more externally, with co-authors in
11
Newman studied the Medline database of the National Institutes of Health.
121
an outward direction rather than within the school, because the ratios are all greater than 1.
122
Table 6 Ratio of Unique Co-Author Links by Department

Department Health Informatics Social Sciences Economics Logistics #Publications 170 302 187 386 498 #External Links to # Internal Links 6,22 6,33 7,73 9,77 10,57 #Publications to #External Links 3,04 3,18 2,20 3,04 3,36
There appears a trend in Table 6 that a higher external to internal ratio may be associated with a higher number of publications, but there are not enough categories to verify this correlation. The relationship also seems to be stronger in Table 7 for researchers with high numbers of publications. Observe that these researchers are called connectors because they have the greatest number of unique co-author links for their sub-clusters and not because they have the greatest number of publications.
Table 7 Ratio of Unique Co-Author Links by Researcher

Researcher Connectors DM Berge Lkketangen Gammelster Hervik S Wallace #Publications 30 77 80 154 237 #Unique co-author Links 22 32 32 52 64 #Exteranl Links to # Internal Links 1,00 2,56 3,00 2,47 6,11 #Publications to #External Links 2,73 3,35 3,33 4,16 4,31
Conclusion and Implications
123
In the first part of this study we analyze the properties of the social network of researchers at Molde University College and recognize that the degree distribution of collaborations is similar to other Scale Free Networks. Understanding how the members of the network connect to each other and the importance of connectors or hubs within this network may help researchers to come in contact with other researchers and to form new associations or links within this network.
In this study we analyzed the connection between researcher productivity as measured in the total number of publications and the existence of co-author collaborations. We found that some research departments have a stronger tradition for doing collaborative work than others. It was asked if one can we boost the total number of publications by increasing the number of co-authors on each paper? The answer seems to be no. In addition to Figure 3, it was also shown in Table 1 that the average number of co-authors per publication across all researchers was about 1 co-author per work regardless of the number of publications. Taking only one example, S. Wallace has 237 publications while Molka-Danielsen has only 25. This is an order of magnitude difference. However, the average numbers of co-author links are almost alike. They are 1.02 and 1.12 respectively. We therefore conclude that the existence of co-author research is not the only factor in creating a high number of publications. Figure 2 however, indicates another factor that can contribute to a greater number of publications that is the number of unique associations.
124
The relationship between the unique associations and the number of published works should not be seen as a one way causal relationship. Barabsi (2002) found that two factors are present in networks that exhibit the scale free structure that is growth and preferential attachment. Preferential attachment explains the phenomenon that any researcher would prefer to work with a more famous established experienced researcher than with an unknown. This phenomenon, also known as the rich get richer, explains the occurrences of hubs in these networks. This was also observed in the Newman (2001) study of the citation database. So, while more associations can lead to more production, it is also true that more production can lead to more associations. For a young researcher, it is only important how they can join this virtuous cycle.
Therefore, thirdly, we looked at the role of the connectors or hubs to see if the external connections of these researchers contribute to a greater number of publications. In addition to acting as connector to research groups within the school, they have many external contacts. We have not evaluated the connectivity to researchers outside the group of 75 in this study. However, some of the faculty members have come from other research institutions and also have joint appointments with other research institutions. As indicated in Table 1, maintaining joint or additional external appointments seems to contribute to both the establishment of external links and to the total number of publications for the school.
Finally, it may not be natural for faculty within departments to work together. But we have not concluded that internal collaboration is related to a high number of publications.
125
We see a low cluster coefficient across the college as a whole and this is because it may be seldom that someone in the Health department co-authors a paper with someone in the Informatics department. However, co-authorships are not chosen randomly, but by overlapping interests of research topics. In conclusion, the internal network does not have to be tightly linked to foster production, but faculty may look to this connectivity to make new contacts externally. Above all, it is important to make new contacts.
126
REFERENCES
Albert, Rka, Hawoong Jeong, and Albert-Lszl Barabsi (1999) Diameter of the World Wide Web, Nature, 401, 130-131. Albert, Rka, and Albert-Lszl Barabsi (2002) Statistical Mechanics of Complex Networks, Reviews of Modern Physics, Vol 74. Barabsi, Albert-Lszl (2002) Linked, Cambridge, Massachusetts: Perseus Publishing. Barabsi, A.L, H.Jeong, Z. Nda, E. Ravasz, A. Schubert and T. Vissek, (2002) Evolution of the Social Network of Scientific Collaborations, Physica A 311, 590-614.
@q Qhl and Alfrd Rnyi (1959) On Random Graphs I, Math. Debrecen vol. 6, 290297. (Eight papers on random graph theory by these authors are listed in Michal Karonski and Adrzej Rucinski, The Origins of Theory of Random Graphs, in The Mathematics of Paul Erd V, ed. R.L. Graham and J. Nesetril, pub. Berlin_ Springer, 1997.
Flake, Gary William, Lawrence, Steve and C. Lee Giles, Efficient Identification of Web Communities, Proceedings of the Sixth International Conference on Knowledge Discovery and Data Mining (Boston, Mass: ACM Special Interest Group on Knowledge Discovery in Data and Data Mining, August 2000), 156-160. Gladwell, Malcolm (2000) The Tipping Point. New York: Little, Brown. Grossman, Jerrold W. (2003) maintains a website that explains the Erd ir (Accessed, July 17, 2003.) http://www.oakland.edu/~grossman/erdoshp.html Molka-Danielsen, Judith (2004) FORSKDOK hit list a complete listing of researchers, publication count, and co-authors referral date August 27, 2004 can be found on my web page: http://home.himolde.no/~molka/bibsys-test-04.html Newman, Mark (2001) Clustering and Preferential Attachment in Networks,, Physical Review, E64, 25102. Watts, D.J., and S.H. Strogatz (1998) Collective Dynamics of Small-World Networks, Nature 393, 440-442. Watts, D.J. (1999) Small-Worlds. Princeton NJ: Princeton University Press.
127
Appendix
The following Figures demonstrate the clustering of co-authors. Only internal co-author links within the node set of researchers are shown.
K.Haugen 17 P. Stre 11 NJ.Berland 17 47
all ac e
7
S.W a all ce
K.Haugen
J.Odeck 2
Informatics
2 H. Arntzen 2
2 A.Olstad K.Jrnsten 4
77
O.Larsen 2 .Halskau J.Odeck 3
A. Lkketangen 15 K. Danielsen 6 26 10 B. Jger 6 LM Hvattum 15 O. Ohren 7 K.Borgen 4 6

1
S.
K.A. Olsen 1 1
1
16 4 5 5 D. Tipper
8 25
17 D. Woodruff 1 Olstad K.Haugen 4
J.Molka-Danielsen 2 1
M.Spring
O. B
1 10 A-K Wallace
HF Nordhaug 2
AG
S. W
M.Risnes 2 T.Crainic
Figure 5: Informatics Institute Cluster: researchers and co-author links.
128
al la ce
2 F.Lbersli
14
R. Hveberg
Li
um
S. Br t he n
3
H A.
T.Aarseth 5 I.Gjerde4 .Brekk1 O.Hauge1

ter H. Gammels 6 2 de ov nh 1 R L.
vik er
Helsefag
3 E.Jrgensen
30 22 D.M.Berge K.Westad Hauge
4 M.Lvlien 5 E.Lykkeslet Strmskag
1 4 E.Braute 1 3 20 S. Vatne 1 4 G-U.Stavik
4 R. Michaelsen
52 S. Bjrkly
4 J. Berg
2 T.Skrondal 1
E. Rekdal
els am m H. G ter
1 HK.Aass 1 AM.Botslangen 2 K.Dahl
2 AJ.Ory
1 T.Aarseth I.Gjerde 1
3 H.Sundal 1 2 H.Bakken
2 A.degrd
2 I.Kamsvg
Figure 6: Health Institute Cluster: researchers and co-author links.

A. Hervik A. Hervik 3 1 EG.Standal 3 N.Rudi 4 1 E.Rekdal 1 3 .Opdal 80 1 1 S. Vatne
Samfunnsfag
E.Rekdal
16 B.Jkupsstovu
DM Berge 1 5 A.Brekk
2 1 31 1 9 2 6 T. Aarseth
H. Gammelster
H.Bakken 4 1
I. Gjerde 2 3 11 2 5 L.Rnhovde 2
rge .Be D.M
3 4 28 K. Tornes
erge D.M.B
en a kk H.B
M D.
rg .Be
D.M.Berge 6
Figure 7: Social Sciences Institute: researchers and co-author links.
129
S. Br t he n
2 11
A.H ervik
39 D.Woodruff 4 5 A.Lkketangen 20 1 O.Larsen A.Lkketangen AK.Wallace 1 D.Woodruff 1 2 2 K. Haugen
A.H
NJ.Berland
1 2 2 Dauzre-Prs 5 7 A.Olstad
konomi/Logistikk
237 S. Wallace
A.Lkketangen 2 2
d 7 NJ.Berlan
ic T.Crain
2 4
2 2 A.Lkketangen 30 3 7 2 T.Crainic 1 2 3 AG.Lium 1 1 9 I.Gribkovskaia 69 A.Buvik 4 A.Lkketangen 2 98 K.Jrnsten 2 A.Hervik
. Halskau
Figure 8. Economics/Logistics Cluster: researchers and co-author links.

5 B.Guvag 2 5 88 S. Brthen 13 D. Woodruff 3 2 A.Lkketangen 1
DM-Ber ge 1
konomi
11 19 P. Solibakke
2 R.Rasmussen
13 B. Foss
erge DM-B
K.Bedrings 4 O.Sttem 8
19 J.Odeck
14 2 154 A.Hervik 1
26 H.Hjelle
.O
6 A.Dedekam 9
1 1
pdal
1 DM-Berge
H.G am me ls te
1-S.Vatne
O.Hauge 2 3 12 H.Bremnes 11 K.Haugen S.Wallace 1 2 K.Jrnsten 2 A.Olstad 1 A.Buvik

O.Lar se n
2 E.Rekdal
1-.Opdal
A.Hervik
r
e rv ik
ik erv A.H
2 1
1-H.Gammelster
Figure 9: Economics Institute Cluster: researchers and co-author links.
130
11 2 K.Bedrings 5
konomi
88 S. Brthen 14 154 A.Hervik 3 1 1 1 2 1 3 1 8 26 H.Hjelle
1 1 3.Opdal
Samfunnsfag
1 1
2 39 K. Haugen 2 20 2 O.Larsen 2 2 4 2 30 3 . Halskau 7 1 2 3 AG.Lium 2 69 A.Buvik 98 K.Jrnsten 4 2
1 7 A.Olstad 1
237 S. Wallace
5 1 1
1 30 D.M.Berge 1
1 R.4Michaelsen 1 1 3 20 S. Vatne 2 4 1 4 J. Berg G-U.Stavik 1 3 H.Sundal
Helsefag
9 1 I.Gribkovskaia 1
1 2 H.Bakken
Figure 10: Connected Network Tree of researchers and co-author links.
131
5 1 A.Brekk
konomi/Logistikk
2 6 9 I. Gjerde 3
12 3 H.Bremnes
4 16 B.Jkupsstovu 2
7 11
2 E.Rekdal
11 L.Rnhovde
9 3 1 O.Hauge
1 31 T. Aarseth
28 K. Tornes
19 13 J.Odeck 2
80 H. Gammelster
13 4 B. Foss
17 11 P. Stre 2 77 NJ.Berland 17 2 H. Arntzen A. Lkketangen 47 1 K.5 Danielsen K.A. Olsen 6 16 8 4 1 D. Woodruff 1 7 25 26 5 5 10 J.Molka1 B. Jger D. Tipper M.Spring 2 Danielsen 6 1 LM Hvattum 1 6 6 15 O. B 2 O.Ohren T.Crainic 1 10
A-K Wallace
4 7 K.Borgen
Informatics
Mobile Work Mobile ICT Supporting Secondary Work

Bente Skattr, Agder University College, Grooseveien 36, N-4876, Grimstad, Norway, bente.skattor@hia.no Per Hasvold, Norwegian Centre for Telemedicine, P.O. Box 35, N-9038 Troms, Norway, per.hasvold@telemed.no Lasse Berntzen, Vestfold University College, P.O. Box 2243, N-3103 Tnsberg, Norway lasse.berntzen@hive.no Tore Engvig, University of Oslo, Department of Informatics, P.O. Box 1080 Blindern, N0216-Oslo, Norway toreen@ifi.uio.no
Abstract
This paper addresses one area within mobile computing that has, so far, received limited attention by researchers: ICT support of mobile work among non-office workers. To understand mobile work among such workers, we have established a framework that groups work activities into primary and secondary activities. We have used this framework in three different settings: Nursing in hospitals, home health care, and construction. Our research shows that even if primary activities differ considerably, we identified a set of generic, secondary activities relevant to personnel working in all three areas. Such activities are supportable by ICT, and particularly mobile ICT. Keywords: mobile work, framework, mobile computing, primary and secondary activities
Introduction
The use of mobile ICT is rapidly growing. Mobile ICT has primarily been marketed as a tool for business professionals, but such technology is finding new application areas every day. One new, exiting area is the use of mobile ICT among mobile workers that normally do not depend on ICT to perform their work. Lately, the use of mobile ICT for computer supported cooperative work (CSCW) has had
132
an extensive growth, and several researchers have looked into ways of structuring and understanding use of mobile ICT in this context (Julsrud et al. 2002; Lyytinen and Yoo 2002; Kristoffersen and Ljungberg 1998). Their frameworks are focused on office workers, and we have found little help in using those frameworks when dealing with nonoffice workers. Our research focuses on mobile ICT support for non-office workers. Even though bluecollar workers were some of the first adopters of mobile ICT (Churchill and Munro 2001), there has been a lack of research on blue-collar workers and use of mobile ICT. Brodie and Perry (Brodie and Perry 2001) express concern about the lack of research when they state that Without this, how can designers and developers envision and create future mobile technology that will successfully support the dynamic, flexible and often urgent nature of the work that mobile blue-collar workers do?. Our research acknowledges that many mobile workers have primary activities that do not include the use of computers. This is clearly visible for non-office workers like nurses, home health care personnel, construction workers and electricians. Strictly speaking, the majority of these workers do not even need ICT to perform their primary activities. It is still interesting to find out if and how such workers may utilize mobile ICT to improve their work situation. We are therefore proposing our own framework to explore similarities and differences between such professions in order to understand and structure mobile work. We have tested this framework on three distinct case studies: Nurses Home health-care personnel Construction workers
In the following we explain how we developed our framework, including how we used the framework on the three different case studies. For all three cases we have found several opportunities for new applications by looking at secondary activities.
Related Work
Mobile work and the support of mobile work by ICT is a new research area. Despite this, a few models and frameworks have been suggested (Julsrud et al. 2002; Kristoffersen and Ljungberg 1998; Lyytinen and Yoo 2002). We have evaluated these with respect to our case studies. It is our experience that these models do not seem to support us substantially when we examine non-office mobile workers. The purpose of The model of IT-use in mobile settings by Kristoffersen and Ljungberg is to provide designers with a framework of concepts to understand and talk about how people use IT in mobile settings. The model emphasizes environment (e.g. physical surroundings, organisational constraints), modality (the patterns of motion e.g. stationary, walking, wandering, travelling or visiting.) and application (e.g. data, technology and program) to explain the particularities of mobile settings (Kristoffersen and Ljungberg 1998). In our case studies we found that the mobile patterns of the nurses and
133
construction workers could be regarded as wandering, e.g. working while being mobile locally. The hospital and the construction sites could be the local area. Further, we found that since home health-care personnel repeatedly drive from home to home and working at these homes for certain periods of time, their mobile patterns could be classified as travelling or visiting. In all our cases, mobility is the norm and not the exception. Hence, we find no additional information from using Kristoffersen and Ljungbergs model in order to elaborate the work of non-office workers. There is little or no input as to the characteristics of the individual or collective activity of the non-office worker to be gained from this model. We need a further and deeper elaboration of the work by non-office workers. In order to organize research topics in nomadic information environments, Lyytinen and Yoo present a framework that emphasizes four levels: individual, team, organisational, and inter-organisational. All these levels are comprised of both service and infrastructure development. They access the opportunities and challenges of IS research that deals with design, use, adoption and impacts of nomadic information environments. Lyytinen and Yoo focus on different factors that are important to services on different levels. The services should cover any functional application of the infrastructure resources to provide a computational solution to a clients needs (Lyytinen and Yoo 2002). The framework suggested by Lyytinen and Yoo seems to be more adept at identifying factors that will influence the development and adoption of new technologies, but does not provide much detail for actual specification of requirements and understanding of the needs of mobile or non-office workers. Julsrud et al. were focusing on nomadic users (salesmen, administration and office, craftsmen like carpenters and electricians) and these users work practises. This was done in order to understand the need of mobile communication solutions among nomadic working groups. Julsrud et al emphasize three conditions: the work processes, mobility and use of ICT. The work processes relate to repeated duties during the work, mobility concerns the types of travelling or movement during work, and the use of ICT is about existing habits of use of mobile ICT. During their research of work processes they focused on important tasks and how they collaborate (Julsrud et al. 2002). Compared to the model of Kristoffersen and Ljungberg and the research framework of Lyytinen and Yoo, Julsrud et al. we find their work more relevant for our cases, but again, we found that we need a further division or segmentation of the work processes and their contexts.
How the framework was developed

To develop this framework for mobile work among non-office workers, we have relied on experiences and empirical data from three separate, ongoing field studies. The cases take place at different arenas and are at different stages. However, all the cases concern mobile non-office workers and they focus on their work practice and developing ICT support of their work. In each case we have collected substantial empirical data. The case studies and their methods are further elaborated below.
134
Case Studies
Nurses
This case study examines mobility among nurses, secretaries and physicians at a surgical department at the University Hospital for North-Norway (UNN). More specific, we try to investigate if we can improve collaboration and coordination among healthcare professionals by using mobile information technology. If so, what are the critical factors? Which activities are suited for computer support? The case is based on ongoing work. Topics for discussion are: mobility caused by bridging discontinuities, problems with using modalities of mobility (such as local, micro, wandering, etc) as a means for providing inputs for systems design for actual applications, identifying and ranking primary and secondary working activities and finding out how the context of the activities tie up human resources (vision, hearing, hands, fingers, attention, etc). The methods used for studying the nurses and developing the applications rely on interviews, studies of descriptions of formalised procedures, and observations. In addition, grounded brainstorming sessions were used to discuss and interpret the observations. Two applications were designed in collaboration with a group of nurses and physicians and are in regular use today. Most of the ideas for these applications, the design and implementation came about as a result of the studies and discussions with the users.
Home health-care
Home health care is a municipal responsibility. Based on individual patient evaluations, home health care personnel provide necessary services in the homes of the patients. Team organisation is common, but this does not imply that members in the team work together on performing activities in patient homes. Some activities are directly related to the client (medication, feeding, personal hygiene), while other services are related to the client environment (cleaning, shopping). Other activities are related to transport, coordination between team members, reporting, and collaboration with other service providers. Like most public sectors, home health care services are affected by demands for efficiency and quality. Based on observations and interviews, we designed a prototype mobile phone application to support secondary activities. This prototype includes registration of start and end of visits, keeps track of where each team member are, team status, and includes functionality for ordering specialist services for the patients.
Construction workers
This case study covers mobile work in residential construction in Oslo, Norway. The project is a co-operation between Agder University College and Selvaag BlueThink in Selvaag Gruppen12. Compared to other manufacturing, residential construction is still marked by small
12
www.selvaag.no
135
amount of series, low degree of standardisation, custom-made requirements, and a large diffusion of competence and qualifications. In general, manufacturing is supported by ICT to a large extent. However, residential constructing is just supported by ICT in some areas. ICT is supporting the project management and the design process in construction, but ICT is little used for supporting the work and production at the building sites. Developing ICT to support the construction workers will influence their work and bring out new opportunities. There are many professions among construction workers, e.g. carpenters, concrete workers, electricians, plumbers, bricklayer, architect, digger drivers, crane drivers and dynamiter. This case study has focused on carpenters and concrete workers. The construction case has focused on facilitating a rich and extensive data collection. The study is based on an ethnographical approach at the building sites, using a mixture of qualitative techniques like observations, interviews, workshops, meetings, paper reviews (work process documents, checklists, and similar). In this project different types of dictaphones and one type of advanced mobile phone have been tested out at the construction sites while focusing on context of practices and reflection-in-action (Iacucci and Kuutti 2002). This work gave insight of mobile work, details about how, when and why mobile ICT is appropriate to use and the context of use at building sites.
Primary and secondary activities

The notion of primary and secondary activities in the framework is of great importance. Hence, we find it necessary to explain the development of these expressions. Initially they were proposed by one of the authors and then discussed and elaborated upon into the proposed framework as a group effort. The initial idea came from working with nurses to try and find out if mobile information and communication devices could be used to support the work and activities of nurses in a hospital department. The initial phase of that project consisted of a series of meetings and brainstorming sessions with a small group of nurses and the author, who was the project manager and responsible for any overall systems design. The interviews provided a picture of the work in the department that was very much in line with the general image of nurses and nursing. The nurses described their work as being about providing care, attention and being the humanistic element in an increasingly technical and advanced environment where physicians were more focused on technical issues of the diagnosis and the surgical procedures. The early ideas, based on this user input, were attempting to support these kinds of activities. However, the proposed services did not gain a lot of support, as the nurses were very sceptical of bringing computers into the patient room. Observing the nurses in action revealed several new aspects of why handling mobile devices in the patient rooms was impractical and also revealed that the nurses performed a lot of activities that was of a more supportive nature. Handling a mobile device in the patient room was difficult because a lot of the human resources of the nurse were preoccupied with activities related to providing care and attention to the patient. Hands, torso, eyes, speech and hearing were in use for the more important task of providing care and nursing and giving full attention to the needs of the patient. 136
So, from these experiences it seemed to be some activities that were regarded as nursing and some that were not. When confronted with the observations, the nurses interviewed acknowledged that they did a lot of activities that were not nursing and that just had to be done. These were activities related to logistics and the coordination of activities in the department and with other departments, receiving phone calls, handling of messages, etc. An interesting point was that many of these activities were well suited for computer support and could possibly provide more efficient systems than the existing in bridging discontinuities in the workflow. The key issue here is that there are some activities that are part of the identity of the user and what is important for the motivation and the priorities of that profession. Then there are other activities that are important, but has a different relation to what the worker views as part of their work and their identity as a professional. This led us to distinguish between these types of activities. We called the former primary activities and the latter secondary activities. From the notion of primary and secondary activities we developed the rest of the framework to help us understand the context and the nature of the work of practise and the priorities of the users.
Framework for Mobile Work

This section presents the framework for mobile work. The framework aims to contribute to find out more about how mobile work is done, the problems mobile workers face, and how they coordinate and collaborate with colleagues. Our assumption is that you have to understand their work in detail before starting to design and develop mobile ICT to support their work. It is our experience that studying the actual work activities and how these are supported will reveal opportunities for use of mobile ICT. When trying to understand and discuss mobile work, and structuring mobile work into the framework, we aim to contribute by helping answering questions like:
Which activities are suited for computer support? How can we improve collaboration and coordination by using mobile ICT?
Is it possible to utilize current functionality of mobile phones in mobile work? How can ICT facilitate knowledge creation and transferring/sharing of it?
Below is an illustration of the framework for mobile work. The framework focuses on the mobile work and it divides it into primary and secondary activities. Both primary and secondary activity is categorised into four categories: 1) Work procedures, 2) Complex activities and complex problems, 3) Information discontinuities and 4) Quality of work. Other issues are described under the section Attributes like organizational context, mobility context, and knowledge types. Aspects and details of all these issues are presented and explained below the illustration.
137
IrCrCrhyuphr8pv Qvhhpvvvr
xprqr pyrhpvvvriyr vshvqvpvvvr hyvsx
Trpqhhpvvr
xprqr pyrhpvvvriyr vshvqvpvvvr hyvsx
6vir
thvhvhypr ivyvpr xyrqtr r
Figure 1 Illustration of The Framework of Mobile Work
Primary and secondary activities

Activity Theory (Engestrm 1987; Kuutti 1991; Leont'ev 1978; Nardi 1996) place importance in the objectives we want to accomplish in selection of what activities we chose. This indicates that there is some sort of hierarchy or value to tasks and activities in relation to the objective the subject (or actor) is trying to achieve. Based on these thoughts and the early experiences from the nurse case, we define primary and secondary activities as follows: A primary activity has a direct association with the primary objective of the work, while a secondary activity has a more implicit or indirect association with the primary objective. To illustrate the differences between primary and secondary activities we present some examples: An example of primary activity is to prepare a patient before a surgery. This activity leads the nurse to move, clean and talk to the patient. Examples of secondary activities are coordinating appointments with other departments, the passing of messages, checking of information related to patients, checking of status of other activities, other administrative tasks, etc. Another example is a carpenter building a wall. The objective is to erect the wall. His primary activities are sawing, hammering and measuring. Examples of secondary activities while building this wall could be fetching materials and discussing drawings with the colleagues. Secondary activities are often supportive of the primary activity. Secondary activities are of particular interest in those cases where the primary work does not require assistance from a mobile information device, while some of the secondary activities do. Secondary activities are sometimes associated with discontinuities and bridging these discontinuities may be important to the overall performance of the primary activity. The use of mobile ICT might help to bridge these discontinuities. 138
The distinction between primary and secondary activities may provide additional insight in participatory design processes as it may help the designers to inquire into aspects of the work and the work activities that are not always portrayed by the user and their own descriptions of their work. When a nurse was asked if she could see any need for a computer system or some mobile communication device, the answer was very typical of the responses we would get: No, I dont think so. The most important thing for me is to be attentive to the patient needs, to provide care and to be there for the patient I dont need computers for that. The primary objectives of most workers are what define the scope of problems that we are conscious of and where most of the attention and problem solving creativity is focused. In other words: Interviews and other research techniques that do not involve in-situ observations and processes that discuss secondary activities are likely to miss out on activities that are often significant in the amount of time it takes to perform these tasks and activities. In fields where new technology, new services, and new organisation models are being proposed, it may be important to be observant of these hidden activities. Interestingly, there is a paradox in that these activities are hidden from consciousness when people are asked to describe their work and activities, but are often obtrusive and cause or are caused by discontinuities in their primary activities (see (Gershman, McCarthy, and Fano 1999), for examples of how bridging discontinuities is used as a resource for developing new services). To explore detailed differences between secondary and primary activities we have categorized the work activities and work situation into four categories: - Work procedures, - Complex activities and complex problems, - Information discontinuities and - Quality of work. These categories descended from discussions within the case study at construction sites, and they are meant to cover the breadth and the extensiveness of the activities and situations out in the field. Both primary and secondary activities are categorized into these. In the following, each category and their characteristics are presented:
Work procedures
Activities in this category are related to typical procedures and routines; how the activities are usually carried out. Important characteristics of this category might be are people working together in teams or are they working alone? Do they share working tasks individually or within teams? Is the responsibility clearly defined or not? Which tools and which facilities support these activities? Are the procedures formal or informal? Are the routines documented? Are there rules that regulate the work procedures?
Complex activities and complex problems

The activities are related to how exceptions are handled or what happens when complex problems arise. How are they solved and which actions are taken? These activities are often not documented. Are they solved individually or as a team? Whom are contacted and why? What is done when exceptions and deviation occurs?
Information discontinuities
139
This relates to what people do when they disrupt their work when lacking information. Do they search in certain places, or do they ask other people? This situation might happen when unfamiliar tasks are going to be performed. Who are distributing information? Do formal or informal routines exist?
Quality of work
Activities in this category relates to how quality of work is accomplished and how quality is measured. For instance, is the quality assurance focusing on processes or products, or both? How are quality problems discovered and how are they improved?
Attributes
Further, the framework also includes three attributes or aspects. These attributes are mobility context, organizational context, and knowledge types.
Mobility Context
Mobility and context are broad and complex topics. Looking closer at mobility in mobile work, different patterns of mobility exist. Some patterns are already specified and discussed, e.g. local mobility in collaboration in a design company (Bellotti and Bly 1996), micro-mobility in hospitals, remote mobility at construction sites, and both remote and local mobility at London underground (Luff and Heath 1998), walking, visiting, travelling and wandering (Kristoffersen and Ljungberg 1998). Other important aspects related to mobile patterns, is to consider the roles of space and place (Harrison and Dourish 1996). Space is the physical structures and artefacts, while place includes the cultural meaning and use. Working activities are performed at different spaces, but a space can be a different place at different times. Context might be any information that can be used to characterise the situation of an entity, where an entity is a person, place, or object that is considered relevant to the interaction between a user and an application, including the user and the applications (Dey 2001). Terms like context of use, contextual awareness, and context-awareness are often discussed. For instance, the latter is devoted a whole special issue of HumanComputer Interaction13. Some try to related mobility and context, like Kakihara & Srensen (Kakihara and Srensen 2001) considers context to be a separate and important dimension of mobility. Agre (Agre 2001) takes this a step further, and defines a conceptual framework for understanding context. Agre argues that the breakdown of the traditional mapping between institution and place complicates the analysis of context. Agres framework links the built environment (architecture), practice and institution. As we see, defining exactly what mobility and context are and how to deal with it is not a simple task, and we do not try to do that in this report. In this framework we consider mobility context as mobile patterns; where does what happen. This perspective might imply that people move and their work are performed at different locations during the day. We do not consider mobility as a question of distance travelled to perform the work,
13
Human-Computer Interaction, Volume 16, 2001.
140
but recognizes the fact that the worker is not stationary, at her or his desk with fixed infrastructure.
Organizational Context
Most workers do their work in some organizational context. Team organization is one common example of organizational context. Team organization does not imply that workers work together on the same task at the same time, but is more an organizational umbrella for individuals working on different tasks, often in different locations. Pinelle and Gutwin (2003) calls this loose coupling in mobile groups. Team members often possess different skills and formal qualifications, and tasks are often transferred to other team members as the activity progresses. Secondary activities are often used to bridge discontinuities between such team members, and between different teams working on different shifts.
Knowledge types
To support knowledge processes with ICT it is important to structure the different knowledge types and the knowledge processes (Alavi and Leidner 2001; Fagrell et al. 1999; Kucza 2001). Below is a description and examples of four knowledge types as (Alavi and Leidner 2001) classify them: Tacit knowledge: Knowledge that is rooted in actions/activities, experiences and engagements in a profession or a specific context (Polanyi 1966). This could be related to cognitive or technical tacit knowledge. An example is a process that requires a high degree of experience and creativity where the worker grasps and feels how to carry out the activity. Tacit knowledge is hard to communicate or distribute to other, and tacit knowledge is often learned through performing or action. Explicit knowledge: Knowledge that is expressed/articulated and generalized. An example is a clearly described process that states how to perform it. This could be based on a formal education, books, quality or work procedures, or similar. Often this type of knowledge is written and it can be communicated without actually having to perform or experience the knowledge. Individual knowledge: Knowledge that is created and inherent in a person. For instance it could be an activity that just one person in a company knows how to perform. Often, this is a type of tacit knowledge. Social knowledge: Knowledge that is created and inherent in collective actions of a group. An example is knowledge of how a team of people work and collaborate in curtain cases. Often this is a type of tacit knowledge.
Other types of knowledge14 exist, but we have limited our discussion to the four types of
Other type of knowledge types according to Alavi and Leidner: Know-how - procedural knowledge, for instance, knowledge about how to perform a process that is to be carried out. For instance, this knowledge is communicated through routine wor, e.g. how to administer a particular drug. Know-about
14
141
knowledge mentioned above. Even though we are well aware that the boundaries of all these categories and attributes not always are well defined and that they sometimes float through each other, our aim has been to test out this framework and see how it works with respect of the three different case studies.
Results and Discussions

After the group discussions of the framework, we have documented and placed empirical data from each case study with respect on the characteristics of our framework. Through this documentation, reading the others group members work of this documentation, and further discussions of the case studies, we discovered both similarities and differences between the case studies. Also, new interesting aspects became apparent. In the following we will present the results and discuss the headlines of this work.
Mobile work: Primary and Secondary Activities

The following table summarises the main results of our three case studies with respect to primary and secondary activities. They are divided into 1) Examples of activities, 2) work procedures, 3) complex problems, 4) information discontinuities and 5) quality of work. The comments are notated with N for Nurse, H for Home Health-Care and C for Construction.
Qvh Trpqh
1) Examples of activities
2) Work procedures
N: - Prepare patients for surgery, monitor health status, perform tests, provide medical care and medication, communicate medical information to patient. H: - Provide medical care and medication, cleaning and other household tasks, food preparation and feeding, help with personal hygiene. C: - Hammering, sawing, measuring, polishing, concrete filling, hammer drilling. N: - Both strict procedure by using checklists and follow her/his own judgement.
N: - Logistics (keep track of patients, equipment, test results), communicate with relatives, retrieve patient information, report on work, handling messages, (re-scheduling). H: - Travelling between client homes, coordination among team members, ordering specialist services, report on work, and training. C: - Report on work, report problems, logistics (keep track of equipment, tools and materials), search for information, and coordination among team members. N: - Disruptions, triggered by discontinuities, less prescribed procedures.
declarative knowledge, for instance, knowledge about which tool to use or what drug is appropriate for an illness. Know-why causal knowledge, for instance, knowledge about why a tool works or understanding why the drug works. Know-with relation knowledge, for instance, knowledge about how a tool works together with different kinds of materials or understanding of how a drug interacts with other drugs. Knowwhen conditional knowledge, for instance, knowledge when to use different types of tools or understanding when to prescribe the drug. Somehow related to tacit knowledge are also embodied knowledge (Mingers 2001).
142
3) Complex activities/ problems
4) Information discontinuities
H: - Detailed descriptions, including instructions on how to perform different tasks. C: - Sequences of activities and professions, starting at the bottom and go upwards, not documented, but built-in team. Safety rules are documented. N: - Individual judgement, clear plans, debriefing. H: - Change in medical state of the patient. C: - Starting with new buildings or minor construction that is new, buildability of drawings. Solved through secondary activities, individually or by team. N: - Arise through discontinuities. H: - Arise through discontinuities and exceptions. C: - Drawings, discovered during the activity and arise through discontinuities.
H: - Disruptions, triggered by discontinuities, formal routines. C: - Judgement, verbally, built-in, no documentation.
5) Quality of work
N: - Regular meetings at handover of shifts, debriefing, reports. H: - Formal quality inspections. C: - Checking the product/the construction.
N: - Memorise all details during busy periods, heavy workload and accidents might lead to improvisations. H: - Trigged by primary activities, order a specialist, intra-team communication. C: - Trigged by primary activities, discussion with colleagues, looking and discussions over the drawings of what and how do build, reporting. N Triggered by discontinuities in primary activities. H: - Trigged by primary activities. Information found in medical journal, home health care system, or by contacting other personnel. C: - Trigged by primary activities, discussion with colleagues, looking and discussions over the drawings of what and how do build, reporting. N Not described. H: Regular meetings at handover of shifts, debriefing, reports. Also triggered by primary activities, formal quality inspections of journals and documents C: - Controlling the activities with use of checklists.
Table 1 Findings of mobile work, primary and secondary activities. This comparison shows similarities and differences between the three different case studies. The primary activities are very different. Working with patients in a clean, tidy hospital environment is conspicuously far from working with tools and materials at a dirty, noise-filled, outdoor construction site. However, we find interesting similarities in the secondary activities. Secondary activities are often used to bridge discontinuities arising from work handover. The typical situation is reporting as a mechanism of information transfer from one worker to another. Secondary activities are also triggered by the need to perform complex primary activities, sudden discovery of unforeseen problems, or by information deficit. We also witness that many primary activities involve physical or manual work, while secondary activities involve communication, collaboration, reporting and handling papers or documents, or handling logistics. All three cases show use of formal procedures manifested as checklists, flow diagrams 143
and other documentation on how to perform tasks. Such formal procedures are based on paper documentation. Examples are procedures for making patients ready for surgery, procedures for performing client personal hygiene in the home health care, and safety procedures on construction sites. Such health and safety procedures are well documented and strictly formal. Although common procedures are learnt by hearth, it is sometimes necessary to review the paper documentation. Our case studies show that in spite of large differences in primary activities, the workers need mechanisms to collaborate and coordinate15, retrieve information (e.g. formal procedures, standards, background documents and drawings), exchange information, make reports, and order goods and services. This discovery implies the possibility of developing some kind of common toolkit for personnel doing mobile work, independent of their primary activities. We found that possible areas for introducing mobile ICT for all these cases might be related to communication, coordination, collaboration, knowledge processes and logistics.
Other findings
Mobility Context
A common characteristic of all the occupations in this study was that they almost continuously were on the move. E.g. nurses moved from patient to patient in different rooms, home health-care moved from patient to patient in different houses and construction workers moved from section to section in a building, or from building to building. These are classic examples of remote mobility, that is conventional sense of mobility; individuals who move around different physical locations who require access to patients, colleagues, or information (Luff and Heath 1998). Being constantly on the move, will require that mobile ICT must be easily available and wearable. That is, it must be embodied (Mingers 2001) and part of the workers embodied knowledge and interaction with the world. This is also suggested by the fieldwork studies of Fallman (Fallman 2003). In our case studies we see this clearly, especially when it comes to the carpenters. In the same way as the hammer, the inch ruler, the knife and the carpenter's pencil, the mobile ICT must become a part of their work wear. Another notion of Luff and Heath (ibid) is micro-mobility. This is related to use of artefacts. In their research, they exemplify this with a medical record, and describe how this medical record is mobilised and manipulated for various purposes at hand. This is also the case for both Nurses and Home Health-Care. Related to construction site, micro-mobility will certainly challenge mobile ICT if we are going to support processes that involve drawings. At the construction sites we often observed the carpenters discussing what they are going to build over the drawings, while concrete workers discuss how to build things. While doing this, they often use their pencils to
Following Brodie and Perry (Brodie and Perry 2001) we define collaboration as working on a shared task-based goal, whilst co-ordination involves a simpler form of information recall or articulation work (most usually agreeing where to meet, or breaking a task into parts).
15
144
draw and make notations. It could be relevant to ask whether it is possible to make use of handheld at all in these situations. The drawings are a good example of a fine-grained artefact (Luff and Heath 1998).
Knowledge types
All three cases show significant amounts of tacit knowledge (Polanyi 1966) and social knowledge (Alavi and Leidner 2001) and social relations between the workers. However, at the moment we are not able to tell whether there exist different types of knowledge connected to primary and secondary activities. In order to answer this, further investigation is necessary.
Experiences with the methods

Two basic methods for studying the nurses activities were applied: Interviews and studies of descriptions of formalised procedures, and observations coupled with grounded brainstorming sessions to discuss and interpret the observations. It is interesting to note that the interviews and formal procedures were almost only limited to descriptions of primary work activities, while it was only as a result of the observations we became aware of the extent and importance of the secondary work activities. The experiences at the constructions sites were quite similar. While observing their work, it became evident that they actually use a lot of effort performing secondary activities, e.g. searching or fetching materials or tools, examine drawings, discussing solutions with others. It is also our experience that the workers often were focusing on the primary activities when they were expressing their needs. Only rarely did they express their needs related to supportive activities (secondary activities).
Rationale for mobile ICT

In all three cases, workers spent significant time on secondary activities. If secondary activities can be done more effectively, workers can spend more time on primary activities. Mobile ICT may reduce time spent on secondary activities and improve quality. We found that possible areas for introducing mobile ICT for all these cases might be related to communication, coordination, collaboration, knowledge processes and logistics. Another common characteristic of all three cases is limited ICT skills among the workers. Thus, usability is a concern, and necessary steps must be taken to develop solutions that are very easy to use. Brodie and Perry (2001) supports this concern and report heavily use of mobile phones among blue-collar workers, but also that many of their subjects were extremely resistant to using other forms of technology.
Framework Evaluation
The development of the framework itself, as well as the actual application of the framework to the three cases, has certainly been a process of maturation. We cannot say that maturation is completed, but we are currently able to do a preliminary evaluation of advantages and disadvantages of the framework. We learned some important lessons. It was very useful to split activities into primary and
145
secondary ones. This helped us to discover an important pattern. In all three cases, we found that mobile ICT was inappropriate to support primary activities. On the contrary, it became evident that mobile ICT could very well support secondary activities. Mobile ICT can improve both efficiency and quality, and the mobile workers could be able to spend more time on primary activities. This observation could be an important contribution to mobile ICT design and development processes. Some shortcomings of the framework were also discovered, and we would like to discuss problems related to two important topics. First, we had some difficulties selecting the appropriate level of detail. Should our focus be on a general level, or should we dive into all the details? We ended up trying to use examples that capture typical characteristics of the work performed within each case study. In some cases, we also dived into the details. Thus, we became aware of the lack of depth concerning knowledge types and related knowledge processes. Therefore, future work should explore how to better integrate knowledge types and processes into the framework. Secondly, we have identified a set of attributes we found particularly relevant to our case studies of non-office workers. We believe that other relevant attributes exist, and hence, our framework should be examples of attributes. Further empirical work is needed, especially from other non-office occupations.
Conclusion
Our research has focused on the development of a framework for structuring and understanding mobile work among non-office workers. The main idea is to categorize work activities into primary and secondary activities. The framework was applied on three different occupations: nurses, home health-care personnel, and construction workers. We found that even while primary work activities between the three case studies are very different, there are considerable similarities related to secondary work activities. All three case studies showed secondary activities involved communication, collaboration, reporting and handling papers or documents, and handling logistics. Primary activities mostly focused on some kind of physical work, while secondary activities normally were non-physical work. Secondary activities were often trigged by primary activities. Based on our three case studies, we have evaluated the framework itself. We presented lessons learned, as well as some possible improvements. We found the division of work activities into primary and secondary activities very useful. We showed that mobile technology did little to support primary work activities, but had an important potential in supporting secondary activities. Based on the assumption that mobile technology may improve both efficiency and quality, mobile workers should be able to spend more time on primary activities. Therefore, this framework could contribute in the design and development of applications for mobile workers.
146
Bibliography
Lasse Berntzen is an Associate Professor at Vestfold University College. He is teaching courses on development of advanced Internet applications using the .NET framework, including mobile applications. He is also doing research on E-Government and EDemocracy. Per Hasvold is a PhD student at Norwegian Centre of Telemedicine, title Understanding mobility and computer-supported collaboration in healthcare. His background is from system development and implementation of ICT within healthcare. Bente Skattr is a PhD student at Agder University College. The title of her PhD Thesis Proposal is ICT Support of Mobile Work: Design and Usability for Exposed Physical Working Environment. Her background is from system development and implementation of ICT within the sectors of banking, insurance, financial and logistics. Tore Engvig is a Cand. Scient student at the University of Oslo working with separation of concerns using AOP in component based QoS systems. Tore started working as a system developer in the mid 90s. Hes experiences mainly relate to distributed real-time systems for financial system, but also includes banking, insurance, publishing systems and e-business.
147
References
Agre, Philip E. 2001. Changing place: Contexts of Awareness in Computing. HumanComputer Interaction 16 (2):177-192. Alavi, Maryam, and Dorothy E. Leidner. 2001. Review: Knowledge Management and Knowledge Management Systems: Conceptual Foundations and Research Issues. MIS Quarterly 25 (1):107-136. Bellotti, Victoria, and Sara Bly. 1996. Walking Away From the Desktop Computer: Distributed Collaboration and Mobility in a Product Design Team. Paper read at Proceedings of the 1996 ACM Conference on Computer Supported Cooperative Work, at Boston, Massachusetts, United States. Brodie, Jacqueline, and Mark Perry. 2001. Designing for Mobility, Collaboration and Information use by Blue-Collar Workers. SIGGROUP Bull. 22 (3):22-27. Churchill, Elisabeth F., and Alan J. Munro. 2001. Work/place: mobile technologies and arenas of activity. SIGGROUP Bull. 22 (3):3-9. Dey, Anind K. 2001. Understanding and Using Context. Personal and Ubiquitous Computing 5 (1):4-7. Engestrm, Yrj. 1987. Learning by Expanding: An Activity-Theoretical Approach to Developmental Research. Helsinki: Orienta-Konsulit. Fagrell, Henrik, Fredrik Ljungberg, Magnus Bergquist, and Steinar Kristoffersen. 1999. Exploring Support for Knowledge Management in Mobile Work. Paper read at The European Conference on Computer Supported Cooperative Work (ESCW'99), at Copenhagen, Denmark. Fallman, Daniel. 2003. Enabling Physical Collaboration in Industrial Settings by Designing for Embodied Interaction. Paper read at Proceedings of the Latin American conference on Human-computer interaction, at Rio de Janeiro, Brazil. Gershman, Anatole V., Joseph F. McCarthy, and Andrew E. Fano. 1999. Situated computing: Bridging the gap between intention and action. Paper read at The Third International Symposium on Wearable Computers, at San Francisco, USA. Harrison, Steve, and Paul Dourish. 1996. Re-Place-ing Space: The Roles of Place and Space in Collaborative Systems. Paper read at Proceedings of the 1996 ACM Conference on Computer Supported Cooperative Work, at Boston, Massachusetts, United States.
148
Iacucci, Giulio, and Kari Kuutti. 2002. Everyday Life as a Stage in Creating and Performing Scenarios for Wireless Devices. Personal and Ubiquitous Computing (6):299-306. Julsrud, Tom Erik, Birgitte Yttri, Eva Mjvik, and Petter Nielsen. 2002. Nomadiske brukere. Oslo, Norway: Telenor FoU. Kakihara, Masao, and Carsten Srensen. 2001. Expanding the 'mobility' concept. SIGGROUP Bull. 22 (3):33-37. Kristoffersen, Steinar, and Fredrik Ljungberg. 1998. Representing Modalities in Mobile Computing: A Model of IT-use in Mobile Settings. Paper read at Interactive Applications of Mobile Computing (IMC'98), at Rostock, Germany. Kucza, T. 2001. Knowledge Management Process Model: Technical Research Centre of Finland,VTT Publications. Kuutti, Kari. 1991. The Concept of Activity as a Basic Unit of Analysis for CSCW Research. Paper read at Second European Conference on Computer Supported Cooperative Work, September 25-27, 1991, at Amsterdam, Netherlands. Leont'ev, A. N. 1978. Activity, Consciousness, and Personality. Englewood Cliffs, NJ, USA: Prentice-Hall. Luff, Paul, and Christian Heath. 1998. Mobility in Collaboration. Paper read at Proceedings of the 1998 ACM conference on Computer supported cooperative work, at Seattle, Washington, United States. Lyytinen, Kalle, and Youngjin Yoo. 2002. Research Commentary, The Next Wave of Nomadic Computing. Information Systems Research 13 (4):377-388. Mingers, John. 2001. Embodying information systems: The contribution of phenomenology. Information and Organization 11 (2):103-128. Nardi, Bonnie A. 1996. Activity Theory and Human-Computer Interaction. In Context and Consciousness: Activity Theory and Human-Computer Interaction, edited by B. A. Nardi: MIT Press. Pinelle, David, and Carl Gutwin. 2003. Designing for Loose Coupling in Mobile Groups. Paper read at Proceedings of the 2003 international ACM SIGGROUP conference on Supporting group work, at Sanibel Island, Florida, USA. Polanyi, Michael. 1966. The Tacit Dimension. London, UK: Routledge and Keoan.
149
ICT-related change in complex organisations: The role of infrastructural and regulatory constraints
Margunn Aanestad Institutt for Informatikk Postboks 1080, Blindern 0316 OSLO margunn@ifi.uio.no http://heim.ifi.uio.no/~margunn and Nina Boulus Assessment of Technology in Context Design Lab School of Communication Simon Fraser University Canada nboulus@sfu.ca http://www.sfu.c/aticdl
ABSTRACT
The topic for this paper is the conditions for change in large and complex organisations, where information and communication technologies (ICT) are introduced to enable more efficient forms of communication and collaboration. Our research aim was to study the factors that shaped and constrained ICT-related organisational change, and in this paper we report from a process of introducing a digital patient record system into a hospital. The findings from the case study indicate that space for change and learning is significantly shaped by the existing institutionalised practises and information infrastructure, as well as by regulations, both formal laws and locally defined rules and
150
procedures. The significance of these aspects has been neglected in most studies of organisational change and learning, and should be given more attention.
151
Introduction
Grand visions are associated with information and communication technologies (ICT) in attempts to reform the public sector. These technologies are expected to enable more efficient flow and use of information. In the recently issued national strategic plan for IT in health care, Te@mwork 2007 (HD-SD, 2004) the issue of a well-functioning information flow between actors and levels of health care is central, and ICT is seen as central in allowing this to happen. Through linking health care enterprises and other actors together in electronic networks, more efficient division of labour and distribution of functions between the levels can be achieved. According to the plan, the parallel existence of paper-based and electronic systems, the multitude of applications and the lack of standards hinders efficient interaction between the actors today. But also other factors need to change: We believe in improved quality, increased effectiveness and cost saving through the use of information and computing technology. We also believe that the costbenefit will first be fully realised through changes in routines, organisational development, standardization and the management of processes for change both at the national and local level. (ibid., p.2) Managing this kind of change and learning processes will be crucial, both for health care institutions and generally within the public sector, which are under an increased demand for reform. The topic for our paper is the conditions under which such organisational change takes place, and our research aim is to study how the constraints from the setting affect the learning and change attempts.
152
ICT and organisational change

Visions of ICT-enabled reorganising for increased efficiency can be compared to ideas behind one of the major IT trends in the 1990s: Business Process Reengineering (BPR). The great promises of BPR have remained unfulfilled, and in particular the early emphasis on radical organisational change (Hammer, 1990) turned out to be problematic. However, another principle from Hammers message has had a long lasting effect: rather than automating an organisations existing structure, which is often based on functional divisions, reorganisation should aim at optimising central business processes by using ICT. Also in the visions for reforms of health care and the public sector in general, such a process focus is central and goes along with a client or customer orientation. For example, the core vision of the strategic plan for IT in health care is to establish continuity of care for patients and clients (HD-SD, 2004, p.5), which implies focusing on the patient trajectory as a central organising principle. This goal will require significant organisational change within the health care sector, which is currently divided into a multitude of actors and institutions. The sector spans actors and institutions belonging to the home care, social services and primary care sector and all the way to tertiary care institutions, e.g. specialised university hospitals. Within each of these institutions, reorganising to achieve patient-centred care would involve changes within and between multiple clinical and service departments. However, in contrast to
reorganisations in the business sector, such organisational change will demand significant learning on the site rather than buying ready-made solutions; as standard business processes do not yet exist for most of the patient-related work in hospitals. Moreover, the hospital cannot temporarily suspend the caring for its patients, so change processes must be non-disruptive and successful.
The need for tools for such learning and redesign processes around ICT in health care has been recognised by the Danish institute for health care informatics (DSI) and the EPR Observatory. Both institutions have issued handbooks in how to conduct work process analysis and redesign (Mabeck and Pettersen, 2002; 2003a; 2003b, Kopke and Lassen,
153
2002), where the relation to BPR ideas is explicitly mentioned (Kopke and Lassen, 2002: p. 8, Mabeck and Pettersen, 2003a: p. 11). These handbooks suggest techniques and method for how to organise and carry out organisational redesign, where the involvement of all the actors that belongs to the specific patient trajectory in question is regarded as central. These actors should meet and describe their own part of the trajectory and listen to the others descriptions. The work flow and document flow within the trajectory can be graphically depicted, so that redundancies, overlaps and gaps become visible. This joint construction of a shared representation is seen as a prerequisite for change, which may include elimination of parts of the documentation or shifting work tasks and responsibilities to a lower level in the organisational hierarchy.
A critique of this approach to organisational change within the health sector is presented by Yrj Engestrm (Engestrm, 1999). He claims that the BPR-like descriptions of work flows as linear and temporal sequences of task is too simple and fail to capture the richness of work. So-called practice- or action-oriented ethnographic studies of work has emphasised the socio-spatial dimension of work, manifested in discourse, in gaze, in gesture, in embodied situated action with material artefacts. (Engestrm, 1999: p. 64). Practice-oriented ethnographies of work shows that it is a multifaceted, distributed and collective accomplishment and that material artefacts like paper documents, boards and shelves may play crucial roles (Brown and Duguid, 1994; Luff and Heath, 1998, Lundberg and Sandahl, 2000).
Another objection to the BPR approach is the assumption that the logic of work can be adequately represented by a single, linear and temporally oriented representation. The complexity of medical work have been convincingly described in well known studies, most notably Strauss et al. (1985), where the notion of trajectory was suggested, as well as multiple trajectories intersecting in a grid. The trajectory notion has been taken up widely, and in reorganising attempts the patient trajectory focus is used as an organising principle that is assumed to result in a more efficient service provision. In a more recent
154
study of documentation practices within a hospital this assumption of using the patient as the sole organising principle is questioned (sterlund, 2002, p. 258). The spatio-temporal organisation of work varies between individuals, occupational groups and departments. There is a multiplicity of logics or organising principles, and this is reflected in a multifaceted production and use of documentation. sterlund found that some documents were for individual and private use, some for communal use across departments (e.g. used only by senior residents), some for local use within a physical space, e.g. a department, and some for trans-local use, e.g. documents following the patients. The seemingly redundant practices of documenting the same information in multiple places have their explanation in the demands from these multiple logics.
However, according to Engestrm (1999), the limitation of the BPR-like approaches is not just due to their one-dimensionality, but also to a trait shared with most of the richer, ethnographic depictions of work: Namely, that most of these accounts fail to explain what drives people in their work and how change and transformation is accomplished. With a theoretical basis in activity theory, where the internal contradictions and tensions within a so-called activity system is seen as the central impetus behind change, he has studied processes of organisational redesign, change and learning in Finnish hospitals and health care institutions (Engestrm, 1999; 2001). In the usual accounts of reorganising health care processes, the success stories, the phase of designing the new is described as a relatively straightforward activity. Engestrm, to the contrary, has his focus exactly on this redesign task, and describes the ambiguity and uncertainty that the participants experience in the collective effort to design new patient care processes. The diffuse nature of the problem and of the possible solutions was a significant characteristic of the quest for the new work process. The processes were also affected by power asymmetries and organisational politics. In addition to the explicit and verbalised disagreements between the persons involved, problematic issues (called systemic contradictions within activity theory) were manifested also in more subtle ways, like for example in hesitations during discussions, dissatisfaction formulated in fuzzy terms, and in the linking together of seemingly non-related issues in the discourse. The
155
researcher-facilitated change process brought the different persons and groups notions and conceptualisations into the open, as well as the inherent contradictions and tensions within the activity system. Engestrms studies thus clearly illuminate the non-trivial nature of learning as a collective and developmental activity. However, as the focus in Engestrms studies is on learning and how it is accomplished, the factors that may hinder learning is not given the same attention. Although it was probably not really the case, from the accounts it may seem that the staff designed the new work practices relatively free from any external constraints. What were redesigned were for example paper forms for the patient record and work routines (Engestrm, 1999). With an electronic patient record system, such local redesign would usually not be allowed or possible.
Our research aim has been to study the constraints on such learning situations. We believe that the force of the institutionalised practices, and the organisational power struggles are important, but so are also the local and material particularities. Other studies have shown how the existing information infrastructure (Monteiro and Hanseth, 1995, Hanseth and Monteiro, 1997) and the institutionalised work practices together form large, complex and highly entrenched socio-technical networks where the possibility for change is limited (Hanseth and Monteiro, 1998). In our study we build on these contributions, but also want to study also another category of constraining factors, which is how regulations of different kinds affected the change and learning processes. Our research aim is to describe these constraints as well as the strategies that were employed by the persons involved. Rather than presenting research results, we will suggest a direction for research and start to examine and open up a discussion on the available space for action, for making change within infrastructural constraints. We believe that an awareness if these issues will be crucial for several attempts to reform and reorganise organisations and institutions.
Method
156
The study employs qualitative research methods and is located within the interpretivist tradition of IS research (Myers and Avison, 2002; Klein and Myers, 1999; Walsham, 1993; Walsham, 1995). Our aim is thus not to test hypotheses based on a model of reality that predicts the way different variables or factors relate to each other. We are more interested in understanding phenomena and processes, not the least based on the involved actors own interpretations of them. Our empirical material is used to illustrate the issues of how constraints impact the redesign process. This work should be regarded as work in progress, where we aim to continue data collection as well as refine our analysis.
The work is based on a longitudinal research project, as researchers at the Department of Informatics at the University of Oslo since 1996 have followed the implementation and gradual expansion of an electronic patient record system (EPR system) at a major University hospital in Oslo. The empirical illustrations referred below are based on the authors fieldwork from October 2002 to May 2004. We have performed 22 observation sessions (in total 47 hours duration) focusing on use of the information systems, and 26 in-depth semi-structured interviews. In addition we participated in 5 project meetings as observers, as well as analysed various project documents.
Empirical illustrations
CASE VIGNETTE ONE: The Digital Internal Referral Letter
An upgrading of the digital patient record system scheduled for September 2003 included an internal referral letter, an electronic form for ordering internal services like physiotherapy, examination by a specialist etc. Previously these internal referral letters had been written on paper forms which were sent in the internal mail system. The department described here was selected to be a pilot department for this implementation during May 2003. User representatives from each of the major groups of employees, i.e. doctors, nurses and secretaries, were appointed to cooperate with the IT department. The
157
shift to electronic forms would require changes in the work practices of the doctors, who used to fill in the paper forms during the morning meeting. This was a relatively short meeting where all inpatients and new patients were discussed. If it was decided that a particular patient would receive physiotherapy, the responsible doctor was handed an empty form, filled it in and signed it. Another doctor found a sticker with the patient ID and bar code and stuck it onto the form. All in all the team of doctors had created a very efficient procedure for filling out such forms. After the meeting the forms were handed over to the mercantile staff for registration and copying and then sent with the internal mail to the Physiotherapy department. Filling in an electronic form would require logging onto the EPR on a computer, of which there were only one in the meeting room. That each doctor should log on consecutively was regarded as infeasible, as it would take too much time. As a solution the doctors suggested that the request would be dictated rather than written, and that after the secretaries had written them in the EPR, they would be signed (in the EPR) by the responsible doctor and sent to the Physiotherapy department.
These changes of work practices were central in the discussion in the project meetings with the project leader from the IT department. Both the secretaries and the doctors claimed that they were already overburdened with work. The policy of the IT department was to limit their role to a supplier of IT tools, and hand the responsibility for utilising the possibilities and functionalities over to the individual department. The project leader thus did not intend to decide on the departments internal work practices. The departments manager had appointed user representatives so that this decision could be made by those who were closest to the work. However, the user representatives resisted to be made responsible for this decision on behalf of other co-workers, who might be more affected than themselves, and kept insisting that we want someone from the management level to decide. The project leader did not manage to make them reach a conclusion, and eventually, just a few days before implementation, she was forced to ask the department management to reach a decision. It was decided that for in-patients the doctors would write the letters themselves, while for new patients the referrals could be dictated together
158
with the admission note and be written by the secretaries. During the morning meetings one doctor would be logged into the EPR system, and would write and sign the referral letters for everyone else present. This doctor would appear to be the responsible doctor for ordering the service, which was not correct. The solution to this was to write the responsible doctors name in a different field of the form. Then some other problems emerged. The EPR did not contain any lists that would give the department an overview over which request had actually been sent to and served by the Physiotherapy department. Previously the copies of the paper forms had served this role. There was no simple way to check that a request in a draft version (unsigned) had actually been signed and sent by the responsible doctor. Also in the physiotherapy department the introduction of electronic forms demanded changes in work organisation. There were six sections within the department; each comprised several physiotherapists with specialised skills. Incoming referral letters were distributed on tables and in mail shelves, first to the right group and then within the group. The incoming electronic referral letters would be printed on paper and included in this paper-based routine. Two of the reasons for this were that not all departments sent electronic requests, and there was no functionality in the EPR system for allocating jobs to individuals by transferring the request from the general list.
Although the problems encountered in this case are trivial and very similar to other transitions from papers to digital documents, they indicate our main point: Changing one element in an information infrastructure is not trivial. Substituting a single paper form with its electronic equivalent, showed how a single and simple element nevertheless has connections to other elements, and how these dependencies influence change processes. If we conceptualise IT in its organisational context as making up a complex sociotechnical web, we may say that both the position and the relations of the new element in this web may have to be negotiated. This is a process which may take some time, and it can also be politically charged, as it involves a redefinition of the division of labour; of how roles, responsibilities and risks are distributed. The case also shows how new relations need to be forged, e.g. in the form of gateways between the old, paper based and the new digital information infrastructure. New routines and work tasks were required to
159
connect them. Hence the employees at the department get more work to do and more routines to take care of, and still the realisation of the benefits still lies mainly in the future, when everything is digitised. The actors may see the increased speed of transfer (avoiding the mail system) as a benefit in this case, but besides that, not much has changed. However, the lack of benefits from this change was also closely related to another reason for choosing exactly this procedure as a pilot, namely that it was not critical. If the system did not perform as expected, the manual routines could easily be deployed.
CASE VIGNETTE TWO: Scanning Incoming Referral Letters

The EPR system was expected to be able to include most documents and forms that would be found in the paper based patient record. For several reasons (more fully explained elsewhere, see Hanseth et al., submitted) the development of the EPR system did not proceed as quickly as expected, and the result was a partly paper-based and partly digital patient record. Even if the EPR system would eventually cover most of the required functionality, some information might not be readily available in digital form. For example incoming referral letters from general practitioners, specialists or other hospitals would be on paper for the foreseeable future, until a sector-wide information infrastructure was in place. These factors, together with an impending crisis regarding space and staff in the paper record archive, led to the initiation of a project for scanning paper documents in the spring of 2003.
Initially the strategy for the project was two-tiered. The archive would perform centralised bulk scanning of whole patient records, and the departments should scan the incoming documents, for example letters coming with the ordinary mail, as well as the documents produced during the patients stay in the department. Due to technical problems with the central scanning equipment, the decentralised scanning was started first, and four pilot departments received scanners in November and December 2003.
160
A central aim of the pilot project was generating experience as input to the development of procedures, both the so-called level one and level two procedures, for the whole hospital and for the individual department, respectively. This would include rules on which documents should be scanned, which image quality that was considered acceptable, where in the EPR system they should be placed (the EPR has several chapters and sub-chapters), and when in the work process this should be done. To a certain degree the archive and the hospitals record committee wanted the hospital to have uniform rules on these issues, while the detailed procedures would have to be decided locally, based on the actual local division of labour and routines. The pilot departments were given a so-called skeleton of a level one procedure, and told that they needed to develop their own specific local procedure for the level two procedure. The issues of rules for document categorisation, i.e. placement in the EPR structure, were negotiated between the users and personnel from the archive department, who would work closely with the departments in reviewing the various documents in use.
The agreements reached on documents in general use, would be included in the scanning procedure as a hospital-wide rule. The questions that arose during the practical experience with scanning were fed into the centralised making of standardised procedures/practices. One example may be the scanning of external referral letters. External referral letters are letters sent by the patients general practitioner or another doctor, e.g. a specialist to the hospital with a request for examination and/or treatment. The letter usually indicates the problem, and it may include the results of tests that have already been performed. It may thus be one or more pages. When the pilot department started to scan these letters, several practical issues arose. It had to be decided whether the test results would be placed together with the referral letter, or in the EPRs chapter for tests results. It was decided that they should be scanned and placed together with the letter (as page 2, 3 etc.), as long as the tests were ordered by persons outside the hospitals. If the tests were ordered by personnel in the hospital, they would be treated as ordinary test results and placed in another chapter.
161
One problem that had not been foreseen by the project staff, was the fact that in some cases the referral letter would say that a test had been performed, but the result was not yet ready at the time of writing and would be forwarded to the hospital as soon as it arrived. When such individual, loose sheets of test results arrived, they should ideally have been placed together with the previously scanned referral letter. But the scanned documents were saved in the TIFF file format, and thus it was not possible append more pages to a previously scanned document, so the test results had to be included in the EPR as a separate document. This less than optimal situation would exist for the foreseeable future.
As was mentioned above, digitising and eliminating paper documents were partly done based on the limitations of space and staff in the patient record archive. However, realising the expected benefits was not easily achieved. The laws that govern patient records do not distinguish between paper and digital record systems, but the hospitals system must fulfil certain requirements, e.g. to security, long-term storage and archive organisation, and must be certified by the national archiving authorities (Riksarkivaren). Currently the paper record system was certified, and the process of application had been started also for the electronic record system. Until the EPR system was approved all paper documentation had to be kept. To comply with the law, the information entered into the digital record system were printed on paper and stored in the ordinary paperbased record. This means that the documents that were scanned still had to be kept and archived. Initially the plans were to just store them in boxes marked with the date and the department, and not in the paper records of the individual patients. This would make up a less ambitious storage system, but nevertheless yet another system that needed to be handled by the archive. However, this storage system seemed to not be used, as the department claimed that they could not easily eliminate paper documents from the work flow. This will be further described below, but here it suffices to point out that the archive was not experiencing any reduction in the amount of paper after the scanning project started. The reason for this was dual; on the one hand laws and regulations
162
impeded the elimination of paper documents, and on the other hand, the work processes using these paper documents had to be addressed and changed. The realisation of this fact arose as the pilot departments worked on the scanning procedures, and the process of discussions associated with this is the topic of the third case vignette.
CASE VIGNETTE THREE: Designing A Better Way to Work

Workflows were expected to change when digital documents were introduced. However, there was no central dictate on exactly how the reorganised work practices should look like, only very general guidelines, such that incoming mail should be scanned upon arrival. The responsibility for utilising the potential benefits rested with the individual departments. The IT department offered to help the departments to start to reorganise by conducting workshops. These workshops started with mapping the work routines related to incoming mail. In most of the larger departments several people were involved in the registration and processing of the mail before it reached the receiver, usually one of the doctors. The workshops where the authors took part included secretaries, patient coordinators, and the office leader together with persons from the IT department and the record archive. The IT staff did not consider it their responsibility to follow the departments through with the whole redesign, only to get the process and discussions going.
The discussions touched upon issues and problems relating to local particularities. One reason for scanning the incoming mail was to eliminate the paper as early as possible in its circulation within the hospital, and the incoming referral letters was supposed to be scanned when they arrived. However, before the scanned document can be linked to a patient record, the record must be created, and before a new patient record can be created, the patient must be registered in the hospitals patient administrative system. Therefore the task comprised two parts, associated with two separate information systems. The scanning software required Windows 2000 as its operating system, while the rest of the hospitals network was based on Windows NT. The scanners were thus supplied 163
together with Windows 2000 computers, and these computers could not be used for ordinary office work, like accessing the patient administrative system for registration. This fact limited the possibilities for redesign of the process, as these two tasks had to be separated, at least for the time being.
Another aspect of this linkage between the scanning software, the patient record system, and the patient administrative system concerned the incoming applications or referral letters specifically: The staff in this department objected to the idea of scanning when the letters arrived and before they were assessed by a doctor and a decision had been made. The reason for this is that the department receives more applications and requests than it can serve, and redirects some of them to other hospitals. All the applications were registered in the patient-administrative system, but a patient record was only created for those whose applications were accepted. Because of this the department chose to defy the central directive and scan the referral letters only after they were processed (i.e. evaluated by the doctor). In other departments where almost all referrals were accepted, this issue was treated differently, and referral letters could be scanned upon arrival. One significant factor here in regard to the incoming referral letters was the demand (imposed by law16) on the hospitals to reply to the patient within 30 days, the evaluation guarantee. This made it important to design routines which would make sure that a letter would not be stuck anywhere in the process, e.g. if a doctor was away for some time.
As noted above, the workshops were intended to map work processes associated with the handling of incoming mail. In the first workshop six different flows for different sections within the department were discussed. Charting the processes was a collective task, and in the discussions that followed there was not complete agreement as to what worked well and not so well. This uncertainty and ambiguity was naturally reflected in the decisions about what needed to be changed. It seemed that a lot of the meetings content and result depended on the persons present and their engagement, attitudes and
16
Forskrift om ventetidsgaranti. Fastsatt ved kgl res 27. juni 1997.
164
argumentative strength. It was also striking how far the discussions departed from the scanning project and flow mapping. Problematic aspects with several other information systems and the hardware in general, were topics in these discussions. It became larger than merely charting and redesigning linear depictions of specific work flows. Changes in these work flows affected the whole information infrastructure of the department.
One decision that was made by the hospitals central record committee was that a so called evaluation form should not be scanned. This form did not contain clinically relevant information, but was used as a process aid for the doctor to indicate the decision made, like whether the request would be rejected or accepted, and if accepted, what degree of urgency it had. The patient record is usually seen as an information repository, and its role as an actor in the work flow and organisational coordination is often overlooked. The EPR system had consequently been designed without explicit attention to work flow issues, and its main functionality was related to storage and retrieval of individual documents. This implied that to exclude the evaluation form from the record was problematic. As the EPR system did not contain any email-like functionality, there was no simple way to indicate to a doctor that there was a digital referral letter waiting to be evaluated. Instead of finding a pile of letters in the mail shelf, the digital version required the doctor to enter the patients ID number or name into the system before the letter emerged. Keeping the evaluation form would allow it to be used as a signal that there is a referral letter to be evaluated, and it also contains the information that is needed for finding the patients file in the EPR system. However, the doctors dont bring the referral letters to their office desk, but looks through them standing by the mail shelf, and tick the appropriate boxes on the forms before they are put back in another mail shelf for further processing by the secretaries. Going digital would mean that they had to go to their office to do this job. The secretaries suggested that a PC could be placed in the mail room, preferable a stand-up screen so that nobody would sit down there for long times. Another solution to this issue was also suggested. Since one section within the department did not use the evaluation form, the doctors would jot down their decision
165
directly on the referral letter. If this solution was chosen, the scanning would have to occur in the end of the process, not in the beginning as was said in the central guidelines.
SUMMING UP
The cases illustrate how the actors perceived the change process as open, and that no single actor possessed an overview over the whole processes. In a different kind of organisation, e.g. a hierarchically organised production plant, management would at least in principle have an overview over the tasks of the organisation. A hospital employs many occupational groups, many of them professional, which has traditionally enjoyed a high degree of autonomy over their work situation. In addition a hospital must be regarded as a knowledge-intensive organisation, where the diagnostic, therapeutic and care work are based on an evolving knowledge base. Coupled with the number of interacting and interdependent actors who often need to negotiate cooperation in an ad hoc manner, this makes detailed overview over the whole hospital organisation practically impossible. Not only does each employee only know parts of what goes on around her, this is the case also for other groups that may be expected to have general overview, e.g. for the management and the IT department. Consequently, in such change processes there are no expert who alone could design the new work processes. This has to occur in a collective process among the users. However, this local design process must occur in a manner which allows a scaling up to the organisational level. In studying these processes we have seen that these factors influence these change processes: The necessity of linking to the existing information infrastructure was a major shaping factor. It must continue to serve the ongoing work without interruptions, while at the same time the actors must consciously design for a different future. In the cases described above, some of the new practices and routines were clearly seen as preliminary and suboptimal procedures. Their main aim was to accommodate the current hybrid solution rather than to realise the ultimate benefits. Still it was an explicit concern that they should not hinder a later change
166
towards fully digital document processing. This illustrates the need to focus on the transition period as an object of concern and planning, not just the end result or goal. The crucial importance of the transition has also been well illustrated by Monteiro (1998). Technological characteristics of the existing information infrastructure or of the new technology placed constraints on the solution and the way work processes around it became organised. One example is how the hospitals existing standard for operating system would not be changed just to accommodate one new application, which implicated the separation of patient registration and the creation of a record file. Another example is how the file formats used by the scanning software restricted the possibility to design efficient solutions for appending documents that were sent later.
The role of regulations and standards was an important shaping factor. Regulation can be of different kinds in terms of formality, strength, and domain. The local division of work, roles and responsibilities is one kind of regulation that affected the change process, as the example with the doctors signing of internal referral letters shows. The national formal law that governs patient record systems and registers is another example of a different kind of regulation, which substantially impacted the path of the project. A third example is the hospitals guidelines regarding which documents belonged to the record and should be scanned. What is common with all three is that they have roots back to a paper era, so that they to some degree clashed with or impeded the attempts to change.
DISCUSSION
Change processes like these are long and complex as they involve changing a complex socio-technical web, where dependencies and links between the elements pose challenges to any attempts to change. The reverberations of change are not easily predicted. Speaking about realising benefits, in the sense of calculating return on ones investments, 167
is not particularly well suited here. The transition phase is important, and the result may very well depend on how this phase is planned and managed. Organisational learning needs to be facilitated. A transition strategy needs to address both the need to live with the current status, as well as the need to keep the goal in sight and not end up in a suboptimal state, with reified solutions, only half-way to the goal, where further change is not possible. Existing research has discussed so-called barriers to change, both individual and organisational, however, here we have emphasised that the technical and material aspects are important to examine and address when we want to make change.
The empirical material here originates from specialist departments within a tertiary hospital. The uniqueness of the work carried out here may account for the quite high local independence in the change process. This regards both the speed and the direction and content of the process. Neither the management nor the IT department tried to dictate the departmental redesign process. The relationship to hospital-wide rules was not taken on blindly, but was to some degree challenged and negotiated. This might very well be different in another hospital with a different organisational culture in regard to departmental autonomy, or where the work processes did not involve so many special aspects. Routine health care may be more amenable to the imposing of standardised processes, best practices. Further studies could illuminate questions like: How do such constraints play in when the change is forced by management? What does actually happen when there is not enough time allocated for the kind of slow processes described here? Is it in any situation possible to neglect these material, particular aspects that may affect the direction of change? Examining such processes should be relevant for researchers studying how ICT is implicated in the reforms of the public sector. Currently, several projects aiming at better integration of separate public entities are ongoing. A large and ambitious reform project is called SATS (Samordning av Aetat, Trygdeetaten og Sosialetaten), which aims at formal and practical integration of the three entities at the point of service delivery. The three entities are subject to different regulatory regimes, and the regulation may change according to political decisions. Studying the space for
168
action within such projects could thus be a useful complement to the study we have performed within one, relatively homogeneous organisation.
We will continue our own study with the aim of describing and analysing the change processes more substantially. Our aim is to illuminate further the issue of how to make change and learn in complex organisations within infrastructural and regulatory constraints, and to study what role such constraints play. We may for example ask whether the constraints also serve as a resource in change, in that they may serve to limit the action space from a too widely open to a more manageable action space.
ACKNOWLEDGEMENTS The generous and cooperative attitude of persons in the IT department as well as staff in the clinical departments is gratefully acknowledged.
REFERENCES
Engestrm, Y. (1999): Expansive Visibilization of Work: An activity-Theoretical Perspective. Computer Supported Cooperative Work, vol. 8, pp. 63-93. Engestrm, Y. (2001): Expansive Learning at Work: toward and activity theory reconceptualization. Journal of education and work, vol. 14, no.1, pp. 133-156. Hammer, M. (1990): Reengineering work: Dont automate obliterate! Harvard Business Review, vol. 68, no. 4, pp. 104-112. Hanseth, O. and Monteiro, E. (1997). Inscribing behavior in information infrastructure standards. Accounting, Management & Information Technology. 7 (4), pp. 183-211. Hanseth, O. and Monteiro, E. (1998): Changing irreversible networks: Institutionalization and infrastructure. In Kristin Braa and Eric Monteiro (eds.): Proceedings from 20th IRIS, p. 21 40.
169
Hanseth, O., Jaccuci, E., Aanestad, M. And Grisot, M.: Reflexive Standardization. Sideeffects and complexity in standard-making. Submitted to Special Issue on Standard Making of the MIS Quarterly. HD-SD (Helsedepartementet, Sosialdepartementet) (2004): S@mspill 2007. Elektronisk samarbeid i helse- og sosialsektoren. Statlig strategi 2004-2007. [Te@mwork 2007. Electronic Cooperation in the Health and Social Sector. National Strategy 2004-2007 for Norway] Klein, H.K and Myers, M.D (1999): A set of principles for conducting and evaluating interpretive field studies in information systems. MIS Quarterly, 23 (1), 67-93. Kopke, P. and Lassen, D. (2002): Metodehndbog i arbejdsgangsanalyser. EPJObservatoriet, Virtuelt Center for Sundhedsinformatik, Aalborg University, Denmark. Luff, P. and Heath, C. (1998): Mobility in Collaboration. In Proceedings of 1998 ACM Conference on Computer-supported Cooperative Work, p. 395-314. ACM Press, New York. Lundberg, N. and Sandahl, T. (2000): What do artifacts mean to us in work? In Kkl, T. (ed.): Proceedings from the 22nd Information Systems Research Seminar in Scandinavia (IRIS 22), p. 363-372. Jyvskyl, Finland. Mabeck, H. and Pedersen, P.G. (2002): Arbeidsgangsanalyse fase 3: Organisatorisk forandring nr IT-systemet kendes. DSI Institut for sundhedsvsen Mabeck, H. and Pedersen, P.G. (2003a): Arbeidsgangsanalyser I. Beskrivelse af metoden. DSI Institut for sundhedsvsen Mabeck, H. and Pedersen, P.G. (2003b): Arbeidsgangsanalyser II. Skemaer og eksempler. DSI Institut for sundhedsvsen. Monteiro, E. (1998): Scaling information infrastructure: the case of next generation IP in Internet. The Information Society, 14(3), 229-245. Monteiro, E. and Hanseth, O. (1995): Social shaping of information infrastructure: on being specific about the technology. In Orlikowski, Wanda J., Geoff Walsham, Matthew R. Jones and Janice I DeGross. Information Technology and Changes in Organizational Work. Chapman & Hall, 1995, p.325 - 343. Myers, M.D. and Avison, D.E. (2002): Qualitative Research in Information Systems: A Reader. London, Sage Publications.
170
Strauss, A., Fagerhaugh, S., Suczek, B., Wiener, C. (1985): The Social Organization of Medical Work. University of Chicago Press. Walsham, G. (1993): Interpreting Information Systems in Organizations. Wiley. Walsham, G. (1995): Interpretive case studies in IS research: nature and method. European Journal of Information Systems, vol. 4, pp. 74-81. sterlund, C. S. (2002): Documenting Dreams: Patient-centered records versus practicecentered records. PhD dissertation, Sloan School of Management, MIT, USA.
171
IT Infrastructures and Standards
SESSION 4: IT Infrastructures and Standards
This track discusses the important topic of IT infrastructures and the related standards and standardization processes. The first paper by Elevoll and Graumann quantitatively investigates the different factors that lead to successful implementation of a broadband infrastructure. Next, the second paper by Skogseid adopts a qualitative approach and focuses on the importance of alignment between local actors in successful development and operation of a local broadband infrastructure. Finally, in her paper, Blechar takes a closer look at a different type of standards (functional standards) for digital rights management in the mobile telecommunications industry.
1. "An empirical investigation of factors influencing success in broadband implementation" by Elevoll and Graumann 2. Local actors bridging the broadband divides" by Skogseid 3. "The emergence of functional standards: Digital Rights Management standards in the mobile telecommunications industry" by Blechar
An empirical investigation of factors influencing success in broadband implementation

Siv Elvevoll (sivelvevoll1@tiscali.no) Tempeveien 13, leil. 1401 7031 Trondheim and Conny Graumann (conny@hifm.no) Department of Sciences Finnmark University College Follumsvei 31 9509 Alta
Abstract
Succeeding in spreading broadband technology is important for the development of the Norwegian society, but there are only a few or none empirical studies on implementation of broadband technology. In this study different explanation factors, selected from a thorough literature review and a pilot study, were tested in a model mainly building on the work of Wixom and Watson (2001) on data warehouse implementation. A quantitative survey was done among project managers on 129 closed Hykom-projects with a response rate of 45 percent. The key informants provided information on important factors on an organizational level, and collected data were analysed by means of SEM/PLS. The study revealed that the most important explanation factors for project implementation success were Resources and Team skills. The factors User participation and Champion were not supported. Technical implementation success was explained by different characteristics of Technology, which after analysis was attempted split into three new variables: IT plans, Physical characteristics of technology and Knowledge about technology. Keywords: Implementation, Broadband, Key informant, Partial Least Squares, Formative measures
Introduction
Implementation of information technology in order to create possibilities for cooperation and professional development as well as for decentralisation of functions and development of competence in the districts has long been on the agenda in the Norwegian
173
society. The Norwegian IT-sector sold for more than NOK 200 billion in the year 2000, which is 6.7% of the total Norwegian economy (Statistisk sentralbyr 2002). The government has established several programmes to achieve the objective to establish Norway as a country being on a high level regarding use of the possibilities offered by the technology. One of the established programmes, i.e. the Hykom-program, which has its origin in different governmental reports, was first established in 1999 for a three year period. In connection with the governmental eNorway 2005- plan, the program was prolonged for the years 2002-2004. The projects can be divided in three main groups: Projects within the school/education area, within the health/social area and for local government. Hykom is not meant to realise the physical building of high speed communication, but to accelerate the building of fundamental infrastructure by increasing the demand for services and products. The Hykom-funding is only meant for public enterprises and all projects involve cooperation between different participants. A school project may be about developing new means of learning, using video-conferencing between institutions to share special competence personnel, or teaching pupils how to use new technology. Local government projects may be about solving cases better and more efficiently through collaboration, as well as to offering citizens and businesses a high speed network, and giving local government institutions better conditions for communication and teamwork. The health projects can also be about increasing teamwork and skills, through use of video for diagnosing and teaching. Some projects try to move special health services to places where people live, so the users dont have to travel, and thereby saving inconvenience and expenses. Unfortunately, a lot of these and other IT-projects tend to be unsuccessful (Srgrd 1996 and Linton 2002). We think that studying what separates successful implementations from less successful ones is important. The aim of this study is to identify factors that might help implementation projects succeed and thus our research question emerges as: What are critical success factors for successful implementation of broadband technology? This article first presents argumentation for theory foundation and a research model for broadband implementation success that was developed from literature review and a pilot study. Further, it describes the survey that was used to collect data and the results from the Partial Least Squares analysis of the data. The findings will be discussed in the concluding sections.
Broadband and infrastructure
174
Because there are few studies of implementation of broadband/infrastructure projects (Wixom and Watson 2001), it is important to consider what other studies can contribute in identifying potential explanation factors for our study. We assume that studies of infrastructure projects have so many equalities with broadband projects that the findings in these can be used interchangeably. We will dispute that broadband should be viewed as infrastructure. The online dictionary Bokmlsordboka17 has the following definition of the term infrastructure: The network of permanent installations that are the basis for an enterprise, for instance roads, harbours, airports, power plants etc. In the IT context, the construct infrastructure so far has been used to describe the communication paths, i.e. the data network. With time, the construct has been given a broader definition, including such aspects as basic data sources, server services, and intranet i.e. We draw the conclusion that broadband can be viewed as infrastructure, and that the construct infrastructure in this context signifies individual, structural/organizational, technical and other factors that influence the way broadband is implemented. In the following we argue that studies of implementation of other technologies can be used in our literature review because these technologies also are infrastructure. Infrastructure can be seen defined at two levels in the literature (Duncan 1995). The first level includes components from technology, platform, network and telecommunication, key data and the most important applications for data processing. The second level includes resource planning- and management factors, i.e. factors as alignment, architecture and the qualifications of the involved parts, consequently, all IT-resources, which are general in their aim or use. Wixom and Watson (2001) find that there are only a few studies of implementation of infrastructure projects. One of the few studies they find is the one that Parr et al (1999) conducted on ERP18-systems. We did find some other studies on ERP-systems among others Mabert et al (2003) and Jacobs and Bendoly (2003). The construct infrastructure include different technologies and the use of it is so general that booth tangible and intangible IT-resources, that are not attached to a special and confined aim, can be infrastructure. This concurrence in the definition of infrastructure and broadband, makes it highly adequate to use factors from literature review of studies that search for factors for implementation success of the different types of infrastructure.
17 18
www.dokpro.uio.no/ordboksoek.html Enterprice Resource Planning
175
Research Model
In order to develop the research model, the IT implementation -, infrastructure - and IS success literature were reviewed to identify factors that potentially influence success in broadband implementation. In addition a pilot study, where we interviewed 6 project leaders, was conducted. The interviews were carried out with both a list of critical factors, that the project leaders should consider as important or not, as well as an open round of questions where the project leaders should tell us about their problems. This combination of quantitative and qualitative methods was important for us. We felt more certain that the theory actually was relevant for our broadband approach to implementation. Our setting is special because Norway is a small country with few large organizations and a high level of labour democracy. Also because we had a lot of formative measures, face validity was very important, and we felt that the interviewed project leaders gave us a true picture of the problems in our setting, and of what constructs and language they used. In addition method triangulation has a validity value in it self. These findings, together with the literature review, confirmed that using parts of Wixom & Watsons model (2001) could be appropriate for our use. As a foundation for our model, we use Wixom and Watson (2001) who say that A data warehouse is an IT infrastructure project, which can be defined as a set of share , tangible IT resources that provide a foundation to enable present and future business applications. Figure 1 presents the resulting research model. One can find the factors that explain Project implementation success, that is Champion, Resources, User Participation and Team skills, in the study of Wixom and Watson (2001). The factors that explain Technical implementation success have as far as we know not been tested together earlier and regarding the Technology factor, we developed the content of it for this study.
176
Champion
Project implementation success
Resources
+ + +
Technical implementation success
User participation
Team skills
Technology
Figure 1 Research model Before presenting the rationale for the factors and the relationships between the factors, the constructs implementation and implementation success are being analyzed.
IMPLEMENTATION
One of the challenges in IS-Success research is that the dependent variable implementation success has different content from one study to another and that there are many different operationalizations. It is important to reach an adequate understanding of the construct, before we can get a good operationalization. We therefore clarify what we consider to be the content of the construct implementation before we can develop our theory and present our assumptions about the rationale for the factors and the relationships among the factors. For this purpose we developed a framework which was based on the framework which Bacharach (1989) developed for evaluation of theories and on Venkatramans (1989) 4 questions for construct content. We found 4 criteria for assessing a construct; Hierarchical level, scope, dimension and discriminant validity. First we must determine at what level we are studying the organization. In this case we are working at the organizational level as we are studying implementation of broadband technology for cooperation between individual organizations and using project leaders as informants. Second we must determine the scope of the construct implementation. In literature you find very different use of the construct, ranging from very narrow definitions (Andersen 1994) to all-embracing (Munkvold 1999) use of the construct. We choose to use a narrow definition of the construct (Linton 2002). Third it is important to
177
consider several dimensions and regard implementation as more than a pure technical task. If implementation is to be successful the system that is developed must not be a goal in itself, but must be meant for use in a context and for use by people in an organization. Fourth, we must consider other adjacent constructs to clarify your use of the construct implementation even more. After analyzing the construct implementation in the framework, we found that implementation can be defined as: The activities that take place after the decision about introducing broadband technology, and until the implementation has failed or is ready to be routinized in the organization. Implementation has three dimensions, project, organizational and technical. The level of analysis is organizational, and the evaluation of success is done from the project leaders point of view.
THE DEPENDENT VARIABLES

When choosing the dependent variable it can seem that ..one must choose between simplistic methods, or the subjective measures which relate more closely to definitions of success but are based on more subjective data (Voss 1985,1988 in Linton 2002). When can you say that an implementation project is successful? Is it when it finishes on time and within budget (Parr and others 1999), or should one use other measures for deciding how successful a project is? Implementation is a construct with multiple dimensions and multiple measures should be used to investigate the degree of success (Linton 2002). To embrace the different dimensions it can be useful to define successful implementation on 3 areas, the organizational, project and technical implementation success (Wixom and Watson 2001). We choose to use these definitions: Organizational implementation success: Taking successfully care of organizational issues such as change management, widespread support, and political resistance. Project implementation success: Success in completing the project on time and budget and with the proper functionality. Technical implementation success: Success in overcoming technical problems and limitations. For several reasons we choose to limit our investigation to Project implementation success and Technical implementation success in our study. Amongst others we find that studying Organizational implementation success is best done by means of a longitudinal study. Further our use of project leaders as informants will make the study of organizational processes difficult.
178
HYPOTHESES
In the following we present the variables found in our literature review. These are the ones we think are candidates for being critical success factors19. In addition we present the hypotheses one can derive from the research model. Champion Champions play an important role in implementation. Champions can push the project forward to overcome obstacles (for instance technical problems and limitations) for implementation and use (Beath 1991) and help the project keep up progressing for meeting its goals at the right time. It is also possible that a champion could see possibilities for using the technology in new and untraditional ways. Champions often have a strong desire to adopt new technology, and will often solve problems that may arise on their own. Champions may not like bureaucracy; this weakness could make their role less important in tight organized projects, as was the case in the Hykom-projects. It seems as if the importance of the champions is at the highest when they have the most difficult role (Premkumar 2003). This makes it difficult to measure the importance of the champion, because they dont work and are not visible in all projects. The charismatic and non-formal way of working may create problems in structures that demand formal communication. We assume that H1: Existence of a champion is positively associated with a high level of project implementation success
Resources The recognition of Resources as a factor that explain project implementation success seems both intuitive right and gets support both in our literature review and pilot study. A high level of resources is important to secure that the project meets the milestones in the project plan and gets the right functionality. A high level of resources means the right amount of money and time, and this is important both in the beginning of the project and throughout the process. Too many resources could increase the complexity in the project, and make it difficult to finish on time. Too many resources could further change the goals; the project participants could e.g. see new possibilities during the project. Its easy to be very creative in development projects and especially if the resources are in abundance. Its likely that there is a connection between a high level of resources and implementation project success. Especially since a high level of resources could make it easier to
19
Critical success factors are the few things, that must go well to ensure success for a leader or an
organization, and therefore they represent the different managerial or business areas which must be given special and continuous attention in order to give a high degree of success.. (Boynton and Zmud 1984)
179
overcome resistance to change and to create a high level of organizational commitment to the project (Beath 1991). H2: A high level of resources is positively associated with a high level of project implementation success.
User participation
Its accepted both in theory and in society that users should participate in development teams groups for new systems. It is also important for further use of a system, that users participate. It is general acknowledged that users are the best to evaluate and ensure that the new systems is developed in a way that insures that the future use of the system is optimal. The communication concerning the user needs is facilitated and helps make sure that the system is implemented successfully (Barki and Hartwick 1994). When users participate it is more likely that they accept the system when it is delivered. On the other hand user participation can make the process more complex, and therefore makes it more difficult to finish on time. H3: A high level of user participation is positively associated with a high level of project implementation success. Team skills Lately, there has been focus on how projects are organized and performed, both within IT and other disciplines. It is thus strange that we during the literature review found that this factor was not discussed in central studies. Rather this aspect was discussed within the factors user participation and management support (Swanson 1988 in Marble 2000, Thong; Yap, and Raman 1996). Adequate team skills can be both interpersonal and technical abilities, and teams with strong team skills are able to perform tasks and interact well with users. The team members can then associate with users way of thinking and insure that the work is on track and therefore meet the goals on time and within budget limits (Parr and others 1999). The danger with too large a team is that the communication could be ineffective and the complexity could rise. But it is important with strong technical skills, which can give realistic estimates on technical problems and limitations (Wixom and Watson 2001). H4: A high level of team skills is positively associated with a high level of project implementation success H5: A high level of team skills is positively associated with a high level of technical implementation success.
Technology
180
The factor technology, as it came forth after the literature review and the pilot study, has some important aspects that can explain technical implementation success. The result of the literature review is difficult to use, because these studies are based on other types of infrastructure. We were uncertain of the operationalization of this variable. One important aspect of Technology is physical properties. This can be the complexity of the technology, how compatible it is with old technology, how difficult it is to use, and whether broadband is already implemented. We assume that knowledge about the technology is important. It is important that the suppliers are giving support, that the participating organizations have the right IT- competence, and that training material is developed to make the testing of the technology easier. Finally we assume that it is important that the implementation project is aligned with existing IT-planes. H6: Technology is positively associated with technical implementation success.
Research method
The research question and the conceptual framework give opportunities and constraints when choosing the design and research method. Primarily choice of design shall enable us to give a (valid) answer to the research question and hypothesis. We can differ among explorative, descriptive and explanative (causal) studies and three types of quantitative design stand out; Experiment, survey and longitudinal design. In our causal study we were trying to explain the relations between independent and dependent variables. Our research question, research model and hypotheses ruled out experimental design among other things because manipulation of variables was out of the question mostly because we were working on an organizational level, where variables are hard to control and manipulate. Usually experiments take place at the individual level. Carrying out a study with a longitudinal design was difficult primarily because of our limited resources, especially the time available for this study. This left us with survey as the most obvious design choice, which has its strong side on covariation20 but is weak on sequentiality and isolation. This means that we must lean upon the theoretical fundament when arguing for directions of relations. Control variables were included in the study to meet the demand for isolation. By doing this we tested whether alternative explanation variables were important. These were primarily determined by the setting of the study (Mitchell 1985) and were the following: Management support, Commitment to change, Formalization and demographical variables. We also included questions aimed at addressing key informant issues (Petty; Cacioppo, and Shumann 1983) since we used project leaders as informants on an organizational level.
DATA COLLECTION
20
Bollens (1989) 3 components in the definition of causality
181
Our population consisted of all projects in Norway working with implementation of broadband technology, taking place within the public sector and having certain traits in common as regards to organization and funding of the projects. Our sampling unit consisted of all ended Hykom-projects21, which are projects concerned with implementing broad band technology in public enterprises in Norway and getting a certain support from Hykom. We did not draw a probability sample, but ended up with a convenience sample instead. After the first draft of the questionnaire was made it was submitted to a thorough pretesting22 (Churchill 1979) by co-students, project leaders from ongoing projects and different people from Hykom. This was done primarily to ensure the face validity and to test the translation to Norwegian as well as the operationalization of the Technology variable. Use of a questionnaire gives a high degree of standardisation (can eliminate random measurement errors and give valid data) and a low degree of nearness (Ringdal 2001). You make sure that all respondents get identical questions but you have few possibilities to solve any confusion about the questions. In addition to the questionnaire we composed an information letter. The purpose of this letter was to explain the background of the study, estimate the time needed to complete the questionnaire and give instructions regarding filling in, saving and returning it. The informants were also told that the answers would be treated anonymously. The questionnaire was sent by email to 129 closed HYKOM-projects with a result of 56 incoming answers, giving us a response rate of 45 percent. Many important experiences were made during this phase of the study.
OPERATIONALIZATION OF CONSTRUCTS
Development of measures is the process which links a concept to one ore more variables and after this links these to observed variables, e.g. the process of attaching theoretical concepts to empirical items (Bollen 1989). This will ensure that you measure the concept/phenomena you intend to measure. When discussing validity in connection with development of measures our main interest is in construct validity which is the degree of accordance between theoretical and operational measure and is about the content of sense in the concepts. According to Reve (1985) you can split construct validity in these subgroups: (1) Face validity, (2) Convergent validity, (3) Divergent validity and (4) Nomological validity.
by the end of 2003 A pretest is a preliminary trial of some or all aspects of the instrument to ensure that there are no unanticipated difficulties (Boudreau; Gefen, and Straub 2001)
22
21
182
Where possible we used earlier developed items, mainly those from Wixom and Watson (2001). These are items used earlier and appear to be well tested. Only for the variable Technology we attempted to develop a new measurement instrument. This development was based on existing literature and our pilot study. Items were measured based on a seven point Likert scale ranging from (1) strongly disagree to (7) strongly agree and the language of the questionnaire was Norwegian. Traditionally you can split measurement models in two: Formative and reflective (Bollen and Lennox 1991). A formative measurement model is a measurement model where indicators form/are causes to the latent variable. A reflective measurement model is a measurement model where indicators are effects of or reflects the latent variable. To distinguish between these two one can ask the following questions for each measurement model (Bollen and Lennox 1991 and Sandvik 1998): (1) Do you expect covariation among the items? (2) Are the items not concrete? (3) Do the items represent different approximations to the variable? (4) Do the items reflect the latent variable? If you can answer YES to all the questions you should expect to have a reflective measurement model. In our study we discovered that Project implementation success and Technical implementation success had reflective measurement models, while the rest of the variables had formative measurement models. In table 1 and 2 you can see the items for the constructs.
DATA ANALYSIS
For analyzing our data we employed Partial Least Squares (PLS Graph, version 3.00 developed by Wynne Chin23). PLS is a SEM (Structural Equation Method) which has several strengths favourable for our study, among other small sample size requirements (approximately 50) and the ability to handle both formative and reflective measures. For the descriptive statistics of the data we employed SPSS and examined the mean and standard deviation of the data. Items were also checked for skewness and kurtosis and were found to have unproblematic univariate distributional characteristics (i.e. values below 2.0). In our sample we had 5 items with no missing values, 29 items with between 1 and 5 missing and 5 items with between 6 and 9 missing. Before examining the results of the test we must consider the measurements models performance.
Measurement models results When testing the measurement models reflective and formative measurement models should be treated differently.
23
http://disc-nt.cba.uh.edu/chin/
183
Formative measurement models validation Larsen et al (2004) suggest a duplex approach to validation of formative measurement models, where the most important thing is to evaluate the face validity of the operationalizations of the concepts. This we attempted to achieve in the pretesting of the questionnaire. Further it is necessary to evaluate the nomological validity. This concerns whether the concepts with formative measurement models actually have explanation power in the model in which they appear and is evaluated after the analysis. Items in a formative measurement model may but do not need to correlate with each other (Bollen 1989). This makes factor analysis unsuitable as validation method. However the weighting for each item can be inspected to identify the relevance of the items on the research model (Wixom and Watson 2001). See table 1. As it is difficult (impossible?) to conclude anything from the weighting of the formative measures, we had to assume that the face validity was in order and checked the nomological validity after the analysis. Table 1 shows item/operationalization, mean, standard deviation, weight and t-Stat for the items with formative measurement models in our research model. For each concept the theoretical definition is shown. Table 1: Survey items formative measures
tStat24 Champion: The presence of a person within the organization (and/or the project group) who actively supports and promotes the project A manager from IT maintenance and services 3,75 2,27 0,52 0,78 supported and promoted the project A manager from the professional area supported and 5,62 1,67 0,76 0,81 promoted the project Resources: Adequate amount of time, money and personnel to successfully implement the project The project was adequately funded 5,33 1,48 0,08 0,14 The project had enough participants to get the work done The project was given time enough for completion 5,56 4,40 Mean 1,10 1,90 Std. Dev 0,67 0,74 Weight 1,81 Mean Std. Dev Weight
3,60 t25 Stat User participation: When users are assigned roles and tasks during the implementation of the project
t-stat tells us about the significance of the values. With our sample size values about 2 will be significant at the 5 % level. 25 t-stat tells us about the significance of the values. With our sample size values about 2 will be significant at the 5 % level.
24
184
IT-personnel and users worked together as a team in 4,49 1,71 0,99 2,98 the project Users were assigned full-time to parts of the project 2,28 1,92 0,64 1,70 Users performed hands-on activities during the 4,60 2,32 0,07 0,17 project Users were partly ransomed for participation on the 3,35 2,34 0,06 0,18 project Team Skills: The project is composed to contain both technical and interpersonal abilities Members of the project group had the necessary 5,64 1,41 1,03 18,22 technical skills for the completion of the project Members of the project group had good interpersonal 5,84 0,958 0,13 0,34 skills Technology: Issues regarding the different types of IT- and other technological products being used in the project or already in place in the organization(s) The technology used in the project was standardized 4,33 2,15 0,12 0,35 product Large adjustments in the technology had to be made 4,36 1,95 0,69 2,67 The threshold for starting using the new technology 4,67 1,40 0,20 0,77 was low The new technology was fully compatible with the 4,52 1,82 0,31 0,96 old one The suppliers contributed with the support agreed 5,08 1,70 0,47 2,59 upon There was an available high-speed net (broadband). 3,94 2,49 0,01 0,03 All competing organizations had necessary IT 4,45 1,76 0,14 0,51 competence Good training material for the technology existed 3,68 1,77 0,47 1,84 All competing organizations had an IT-plan 3,88 1,94 0,25 0,67 The project fitted well in in the different 4,79 1,63 0,26 0,94 organizations IT-plans
Reflective measurement models validation

We chose to follow Hullands (1999) three step procedure for validation of reflective measurement models which includes looking at: (1) individual item reliabilities26, (2) the convergent validity and (3) the discriminant validity. In our model we had two concepts with reflective measurement models Project implementation success and Technical implementation success. Individual item reliability is considered satisfactory when exceeding 0.5 and when the t-Stat in addition exceeds 2 this is confirmed. All items in these concepts, except one item representing Project implementation success (Prim2), had loadings (which are a measure for convergent validity) close to 0.7 or above, which is satisfactory.
26
Item reliability is the sqaured factor loading of the item
185
Table 2 shows item/operationalization, mean, standard deviation, loading, t-Stat and item reliability for the items in reflective measurement models in our research model. For each concept the theoretical definition is shown as well as the composite reliability27 values and AVE28 values. Table 2: Survey items reflective measures
Mean Std. Dev Loading t-Stat
Item reliability
Project implementation success: success in completing the project on time, on budget with the proper functionality (Composite reliability = 0,74, AVE=0,71) The project was completed within the final 5,32 1,92 0,88 15,30 0,77 deadline The cost of the project did not exceed its 6,05 1,54 0,43 1,61 0,19 budgeted amount The project provided all of the functionality 5,98 1,21 0,74 4,26 0,55 that it was supposed to provide (secondary goal were reached) Technical implementation success: success in overcoming technical problems and limitations (Composite reliability = 0,91, AVE=0,89) Many technical problems arose during the 3,55 1,73 0,91 18,49 0,83 implementation Numerous technical constraints were imposed 3,43 1,89 0,91 19,83 0,83 on the implementation (new or existing technology was different than intended)
Random errors have little influence on constructs with satisfactory composite reliability. The calculated value for composite reliability should exceed 0.70. This criterion is met for both constructs. Testing the discriminant validity implies uncovering whether there are no (high) correlations with measures from other concepts. This can be estimated through average variance extracted (AVE) which measures average variance between the concept and its items and shall be larger than 50% and larger than the variance between the concept and other concepts in the model. These conditions are met here and are shown in table 3 beneath. Table 3: Discriminant validity for variables with reflective measurement models Project Technical implementation implementation success success Project implementation 0,84 success Technical implementation 0,32 0,94
27 28
Composite reliability: (shpyhqvt2/ (shpyhqvt2+ standard error Average variance extracted: shpyhqvt2/ (shptor loadings2+ standard error
186
success The item The cost of the project did not exceed its budgeted amount from Project implementation success did not met the criterions on several occasions (high kurtosis, low individual item reliability and low convergent validity) and one can imagine the reason for this may be the minor possibilities for changing the budget during the Hykom project. We chose to remove this item from further analysis.
Structural model results

Figure 2 summarizes the structural model results. Standardized regression coefficients, which indicate the strength of the relation between the independent and dependent variables, are shown above each path with belonging t-Stats. R2 values, which indicate the variance explained by the independent variables, are shown above these. R2 indicate the predictive strength of the model and the values are interpreted in the same manner as in ordinary regression analysis.
R2= 0,51
Champion
0,02 (0,41)
Project implementation success
R2= 0,54
Technical implementation success
Resources
0,51 (3,47**) 0,20 (1,27) 0,25 (1,68*) -0,08 (0,52)
User participation
0,39 (3,81***)
Team skills
Technology
Note: * significant at the 10 % level (p<0.01) ** significant at the 1 % level (p<0.10) *** significant at the 0.1 % level (p<0.001)
Figure 2 Structural Model Results

Three of the six path coefficients are significant at least on the 10 % level and the total explained variance for the two dependent variables is for Project implementation success 51 % and for Technical implementation success 54 %, which is satisfactory. Regarding the results from testing against control variables and demographical variables and the key informant diagnostics we can conclude that our original findings seem robust and that they are not affected by these controls. 187
Conclusions and Implications

We did not get support for the hypothesis that the existence of a champion and User participation has a positive effect on Project implementation success. The hypothesis that team skill has a positive effect on Technical implementation success was likewise not supported. But we did get support for the hypothesis that the extent of the resources and suitable composition of the project group have a positive effect on Project implementation success and that Technology has a positive effect on Technical implementation success.
As regards the nomological validity of the variables Champion and User participation one can conclude that the construct validity of these concepts are undermined by our analysis since we were not able to confirm our predictions. As far as the variables Resources, Team skills and Technology are concerned, one can conclude that the construct validity was strengthened as we were able to confirm our predictions in the analysis. As usual one must point out the weaknesses of the use of cross-sectional surveys. Longitudinal design is superior in overcoming these weaknesses. One might question the suitability of our use of project leaders as key informants in the study as one might suspect that the project leaders might not have enough knowledge about the different processes in the implementation projects and that their engagement in the work was little. We did try to safeguard the study from these problems, but still we question this issue. Further research might improve this by including several key informants, for example project owners or other participators in the project. Another issue that we found difficult to handle during the study and one that might be a weakness in this study is the use of formative measures. We found very little literature with concrete techniques for handling formative measures, especially on how to evaluate results from analysis. The operationalizations of the concepts in this study are mainly based on the ones found in Wixom and Watsons study (2001). We were not entirely content with these operationalizations and suspect that the operationalizations for the concepts Champion and User participation might be possible explanations for the missing support for the hypothesis involving these concepts. The concept Technology was attempted developed from the bottom. When starting out we imagined being able to easily identify dimensions in this concept by doing factor analysis with SPSS, which of course not was possible since we were dealing with
188
formative measures. Never the less we try the following split and leave the closer investigation of this to further research.
189
Table 4: Results of testing with Technology-dimensions Relation with Technical Patht-stats implementation success coeffisient R2 = 0,523 IT-plans (item 9 and 10) 0,13 1,50* Knowledge about technology 0,41 2,43** (item 5, 7 and 8) Physical characteristics of 0,52 2,75*** technology (item 1, 2, 4 and 6)
* significant at the 20 % level (p<0,20) ** significant at the 2 % level (p<0,02) *** significant at the 1 % level (p < 0,01)
By IT-plans we mean the presence of well adjusted IT-plans in participant organizations. By knowledge about technology we mean the existence of and the possibility to obtain necessary knowledge and by physical characteristics of technology we mean the adaption of new technology to existing technology and to new demands. When summarizing our contribution, we find that we are able to recommend project leaders to focus on the importance of adequate resources in a project together with making sure that the project group has an adequate composition. We have used an model already developed for testing data warehouse implementation and found that the factors Champion and User participation are not important in broadband implementation projects while they are important in data warehouse implementation. We also find that the developed framework for analysing the use of the concept implementation can be of use in further work, both theoretical and empirical.
Bibliography
Siv Elvevoll graduated from Buskerud University College in 2004 as Cand. Merc. She is writing a proposal for a PhD scholarship on NTNU, were processes that make organizations work optimal is the subject. She has worked many years as teacher in Hammerfest secondary school. Her professional interests cover among others economy, organizations, people and work processes and strategic use of IT.
190
Conny Graumann graduated from Buskerud University College in 2004 as Cand. Merc. and is now working as lecturer at Finnmark University College, teaching courses in IT, systems development, social informatics and strategic use of IS. Her professional interests cover among others Structural Equation Modelling, use of IT in organizations and strategic use of IT.
References
Andersen, Erling. (1994) Systemutvikling. NKI Forlaget. Bacharach, S. B. (1989) "Organizational Theories: Some Criteria for Evaluation". Academy of Management Review. 14(4): 496-515. Barki, H. and Hartwick, J. (1994 Mar) "Measuring User Participation, User Involvement, and User Attitude". Mis Quarterly. 18(1): 59-82. Beath, C. M. (1991 Sep) "Supporting the Information Technology Champion". Mis Quarterly. 15(3): 355-372. Bollen, K. A. (1989) Structural Equations with Latent Variables. Wiley. Bollen, K. A and Lennox, R. (1991) "Conventional Wisdom on Measurement: A structural Equation Perspective ". Psychological Bulletin. 110(2): 305-314. Boudreau, M. C; Gefen, D., and Straub, D. W. (2001 Mar) "Validation in Information Systems Research: a State-of-the-Art Assessment". Mis Quarterly. 25(1): 1-16. Boynton, A. C and Zmud, R. W. (1984) "An assessment of critical success factors". Sloan Management Review. 25(1): 17-28. Churchill, Gilbert. (1979) "A Paradigm for Developing Better Measures of Marketing Constructs". Journal of Marketing Research. XVI: 64-73. Duncan, Nancy Bogucki. (1995) "Capturing Flexibility of Information Technology Infrastructure: A Study of Resource Characteristics and their Measure ". Journal of Management Information Systems . 12(2): 37-57. Hulland, J. (1999 Feb) "Use of Partial Least Squares (PLS) in Strategic Management Research: a Review of Four Recent Studies". Strategic Management Journal. 20(2): 195204. Jacobs, F. R. and Bendoly, E. (2003 Apr 16) "Enterprise Resource Planning: Developments and Directions for Operations Management Research". European Journal of Operational Research. 146(2): 233-240. Larsen, T. J., Sreb, A. M., and Sreb, . (2004). "The Role of Actual IS Utilization 191
Level: Extending the Expectation-Confirmation Model of IS Continuance Intention". Unpublished work. Linton, J. D. (2002 Feb) "Implementation Research: State of the Art and Future Directions". Technovation. 22(2): 65-79. Mabert, V. A.; Soni, A., and Venkataramanan, M. (2003 Apr 16) "Enterprise Resource Planning: Managing the Implementation Process". European Journal of Operational Research. 146(2): 302-314. Marble, R. P. (2000 Sep) "Operationalising the Implementation Puzzle: an Argument for Eclecticism in Research and in Practice". European Journal of Information Systems. 9(3): 132-147. Mitchell, T. R. (1985) "An Evaluation of the Validity of Correlational Research Conducted in Organizations". Academy of Management Review. 10(2). Munkvold, BE. (1999) "Challenges of IT Implementation for Supporting Collaboration in Distributed Organizations". European Journal of Information Systems. 8(4): 260-272. Parr, A.; Shanks, G., and Darke, P., (1999). Identification of Necessary Factors for Successful Implementation of ERP Systems. Ngwenyama, L. D Introna Myers M. D. DeCross J. I. "New information technologies in organizational process"Boston: Kluwer Academic Publishers: 99-119. Petty, R; Cacioppo, J, and Shumann, D. (1983) "Central and Peripheral Routes to advertising effectiveness; The moderating role of Involvement". Journal of Consumer Research. 135-146. Premkumar, G. (2003) "A Meta-Analysis of Research on Information Technology Implementation in Small Business". Journal of Organizational Computing and Electronic Commerce. 13(2): 91-121. Reve, Torgeir. (1985) "Metoder og perspektiver i knomisk-administrativ forskning". Reve, Torgeir. Validitet i konomisk-administrativ forskning. Oslo: Universitetsforlaget; pp. 52-72. Ringdal, Kristen. (2001) Enhet og mangfold. Fagbokforlaget. Sandvik, Kre. (98). "The Effects of Market Orientation". Unpublished work. Srgrd, Pl. Offentlige IT-fiaskoer. Dagbladet. Oslo; 1996 Apr. Thong, J. Y. L.; Yap, C. S., and Raman, K. S. ( 1996 Jun) "Top Management Support, External Expertise and Information Systems Implementation in Small Businesses". Information Systems Research. 7(2): 248-267. 192
Venkatraman, N. (1989 Aug) "Strategic Orientation of Business Enterprises - the Construct, Dimensionality, and Measurement". Management Science. 35(8): 942-962. Wixom, B. H. and Watson, H. J. (2001) "An Emperical Investigation of the Factors Affecting Data Warehousing Success". MIS Quarterly. 25(1): 17-41.
193
Local Actors Bridge the Broadband Divides

Ingjerd Skogseid Western Norway Research Institute Postbox 163, 6851 Sogndal, Norway
Email: ingjerd.skogseid@vestforsk.no HTTP://WWW.VESTFORSK.NO Abstract: This paper analyses how local actors are bridging the rural broadband divide. It presents the bottom-up development of broadband infrastructure in the rural region of Sogn & Fjordane. To bridge the broadband divide individuals, businesses and public sector in several rural communities have joined forces and taken the initiative to develop and operate local broadband infrastructure. Each initiative faces a different set of challenges depending on the local context. The paper open up the black box of the innovative process to get a better understanding of the process of developing infrastructure in rural settings. Two cases will be used to illustrating how the communities have assembled rural broadband assess using available knowledge relational resources and infrastructure. Keywords: broadband infrastructure, regional innovation, actor network. Acknowledgement: This work has been made possible through funding from the Norwegian Research Councils grant no. 150249/510.
Introduction
The use of information and communication technology (ICT) is changing the way we are organising society. Provision of services such as e-government, e-learning, e-health, and to enable a more dynamic e-business environment (eEurope 2002) are important to facilitate regional economic development. To be able to reap the benefit of these services it is necessary to have sufficient access to communication infrastructure that is to broadband infrastructure. At the time of market liberalisation of the telecom sector, the Norwegian government chose a demand driven strategy for development and delivery of broadband internet access. The competition among telecommunication providers was believed to be
194
sufficient to ensure provision of broadband services to everyone. A study carried out in the county of Mre & Romsdal (Molka-Danielsen, Ohren & Jger 2003) found that there are clear differences between rural and urban areas in Mre og Romsdal with regard to the development of broadband infrastructure. The aggregation of demand of users in rural communities is often not great enough to justify the investments. (ibid) The government have put in place mechanisms that can increase the demand, particularly in rural areas; they are supporting regional and local governments by partially funding the acquisition of broadband services. This is a welcome contribution, but the local and regional government still have to provide the major part of the investment, which can be difficult. When the local and regional government are providing services requiring broadband capacity it is also necessary that inhabitants and businesses have equal access to be able to utilize these services. Having access to broadband infrastructure is one way of making the conditions for enterprises in rural and urban areas more equal; the same kind of infrastructure at the same cost. It is important to ensure that both the public and private sector have access to ensure a living community. The national strategy encouraged local and regional governments to systematically work together with local businesses to aggregate a demand and share the cost for broadband access. This paper analyses how local actors are bridging the rural broadband divide. It presents the bottom-up development of broadband infrastructure in the rural region of Sogn og Fjordane. To bridge the broadband divide individuals, businesses and public sector in several rural communities have joined forces and taken the initiative to develop and operate local broadband infrastructure. Each initiative faces a different set of challenges depending on the local context. The paper open up the black box of the innovative process to get a better understanding of the process of developing infrastructure in rural settings. Two cases will be used to illustrating how the communities have assembled rural broadband assess using available knowledge relational resources and infrastructure.
Theoretical Background
This paper will use theoretical contributions from literature on regional innovation systems and actors network theory. The literature can help us understand the processes of building infrastructures. The theories from regional innovation and diffusion theory helps to clarifying the characteristics of regional economies which are important in understanding the spreading of broadband access. In particular the paper will use concepts from actor network theory to explore the diffusion process. In relation to the case studies are the developments a simple diffusion process or a more complex translation process (Latour 1987). The linear innovation and diffusion model, which dominated for a long time, is now widely accepted as being too simplistic to explain all situations (Rogers 2001). Further it has revealed that such processes are influenced by a number of factors, both at a micro and more aggregate levels. The term regional innovation system has thus been introduced to explain part of this complexity. Evolutionary research in the fields of regional economics (Morgan 1997) has defined the main characteristics of regional innovation 195
systems as being learning and innovation, with both individual and collective innovations. Collective innovations are seen as interactive processes in which the firms networks are important aspects of their collective innovative capability. This term captures the trend of building regional organisations and networks to strengthen the innovation capability of enterprises (Cooke 1998). It also includes collaboration on innovative activities (knowledge development and diffusion) between enterprises and knowledge organisations such as research institutes, colleges, libraries, consulting companies etc in the region (Asheim & Isaksen 1997). Results from studies show that the innovative capability of enterprises is highly dependent on their ability to come into contact and to co-operate with other actors, such as customers, suppliers and R&D organisations (Gregersen & Johnsen 1997). Latour (1987) describes two models to describe these processes, the model of translation and model of diffusion. The regional innovation system can encompass both these models. The model of translation describes the innovative processes leading up to a stable product or fact. That is the actions needed to enlist and interest human and non-human actors in the development of an artefact. When these actions are being successful then there is no way around the implementation and diffusion of the artefact. The model of diffusion then describes the actions and strategies needed to diffuse the artefact, that is technical artefacts are equipped with an inertia moves the artefact which seem independent of human action. In the translation process the activity of enlisting and interesting actors and keeping them interested is important. In doing so access to knowledge resources is vital for the innovations taking place in a community. The most sited knowledge resources are; intellectual, social, cultural and institutional capital (Putnam 1993; Bourdieu 2001; Healey et al. 1999). Social capital refers to the social relations and duties that the individuals in a society have toward each other. It takes time to develop this kind of capital, but it can later be taken out as surplus to gain personal or collective goals in a translation process. Healey et al. (1999) uses the expression institutional capital to refer to a combination of the knowledge resources, relational resources and mobilisation capability in a community. Access to knowledge and relational resources are a prerequisite to being able to develop a sustainable institutional capital. To achieve common social goals, these resources have to be deliberately mobilised through use of common arenas and networks using mobilisation techniques and change agents (Healey et al. 1999; Amdam & Veggeland 1998). Local communities are currently facing challenges both from within their community and from a global perspective. As a result of technological development a local community is today much more than before influenced and challenged by things happening in a global perspective. How inhabitants, companies, and organisations react to these challenges and how they co-operate and compete will to a great extent influence their capacity to develop efficient responses to challenges (Amdam 2000; Healey et al. 1999; Putnam 1993; Sthr 1990). A communitys capability of facing challenges is particularly dependent on how the various actors or stakeholders manage to produce and exploit competitive knowledge (Diez 2000). At the same time, challenges must be countered 196
through strategies and tasks that the whole community is involved in and which are adapted to suit the actual situation and challenges of the community in question (Sthr 1990). Healey (1997) points out that to develop a community there is a need to create alliances across special interest groups and ethnical divides to be able to face the challenge together. When there are established arenas where inhabitants meet and discuss, they can learn from each other. The common platform which may emerge through this collaboration can be a basis for meeting future external challenges. By operating in this way the inhabitants and businesses are better equipped to transform the external challenges to their communitys advantage. Further Healey points to the fact that individuals take part in different social networks with different form and reach but at the same time the networks are somehow connected. Social and cultural networks connect individuals, social groups and different types of knowledge. In the networks ideas and impulses are communicated, just like the densely coupled nerve system of a community. What connect such networks are common interests and relations. Trnqvist (1997:102-103) shows that accumulation and development of knowledge in the networks happen through personal contact and communication. According to Healey (1999) these knowledge and relational resources must be deliberately mobilised in the development of a community.
Research Method
This paper explores the process of establishing heterogeneous networks consisting of both a technical and organisational actors. The aim is to understand the process; how and why (Yin, 1994) things happened. An interpretative research approach is used to get a better understand the unfolding of the development. This approach takes both sociotechnical and socio-cultural aspects of the construction into consideration. Interpretive research is concerned with the development of a deep and contextual understanding of the phenomenon. The focus is on the subjective processes surrounding the social construction of the network, using qualitative research methods (Walsham 1993). The aim of such interpretive analysis is to understand rather than to predict. The interpretive perspective helps to focus on the formal and informal parts of the process that took place. This paper reports from a longitudinal process, which covers the years from 1998 till today, which involve activities at different levels; at both regional and local levels. At the regional level the process has been carried out as a participant observation process (Cole, 1991; Baskerville and Wood-Harper, 1996). As researcher Ive not tried to stay distanced or remain a silent observers, I have on the contrary participated in the group activities, had access to minutes and to the discussions taking place. Data has been gathered as part of the process through the interaction, participation and observation of meetings and activities. Data has been gathered through at a local level as part of the regional process, observing and interacting with the local levels, through the web sites of the initiatives, project reports and through interviews with the two main actors in the local initiatives to complete the picture.
197
To explore the cases in dynamic and historical perspective (Granovetter & McGuire 1998 p. 149) concepts from actor network theory (ANT) (Latour 1987; Bijker, Hughes & Pinch 1987; Monteiro & Hanseth 1995; Callon 1999) is applied. In describing the development of the network and the roles and relationships of the actors, ANT is used as an analytical tool. The creation of an actor-network involves linking a number of heterogeneous things (Monteiro 1999) through the enrolment of actors and the translation of their interests and ideas into the network and thereby aligning them (Callon, 1999, Walsham 1997). This approach opens up for a degree of uncertainty as the events are interpreted and the interpreter bases their interpretation on the knowledge they have about the case, this again can open up for different interpretations. In this case, the interpretation is carried out as a critical reflection by one of the actors in the network. The aim is to analyse the development of the network exploring the development process. In doing so, it enables the researcher to get a detailed understanding of the development process, given the specific context.
Broadband development in the Sogn og Fjordane region

At the time of tele-liberalisation in Norway in 1998, access to broadband infrastructure was not widespread. In the Sogn og Fjordane region broadband access was very limited, the regional college and research institution along with the regional hospital had access to broadband capacity while some enterprises had access to leased lines for specific purposes and used ISDN for email and internet access. At the same time urban areas were offered a number of broadband alternatives from both local and national providers. The national providers had no concrete plans to develop broadband access to these rural areas. The development of the infrastructure was demand driven and the capacity and cost could not be guaranteed. The Sogn og Fjordane region is located in the western part of Norway, covering an area of 18,634 square km and with a population of approximately 110 000 inhabitants, that is an average of about 6 persons per square kilometre. The region is characterised by a harsh nature with glaciers, mountains and fjords dividing the populated areas. The inhabitants are spread over large parts of the region with a few agglomerated areas. The industries are to a large extent based on natural resources and are therefore located close to a source for one of its input factors. The main industries include fish farming, fishing, agriculture, food industries, shipyards, mechanical industries, foundry and metal industries, hydro power/energy and tourism. The region has the largest export per capita in Norway (4.6 % of the total export form Norway). Together this makes the region very challenging to develop and in particular with regard to broadband infrastructure. For commercial providers the cost of cabling the whole region is too high and the demand is too small.
198
To meet the challenge a number of regional and local initiatives have been taken, utilizing existing relational and organisational network and establishing new where needed. A total of 13 local and regional initiatives are ensuring that large areas of the region now have access to broadband infrastructure. Below is a list of the initiatives with a short description of main characteristics (BBF 2004). Aurland & Lrdal breiband established in January 2003. Owned jointly by two municipalities and two electricity suppliers. Using a combination of existing cable-TV infrastructure and wireless Datapart as established in 1993 located in Luster and Sogndal. From 2002 providing wireless broadband access to users in seven localities in these two municipalities. Enivest as established in 2000, the largest broadband provider with a regional basis. Provide a combination of ADSL and wireless broadband access in least nine municipalities Jlster Breiband established in 2002, provide wireless broadband access in a number of localities mainly in the municipality Jlster but also in neighbouring municipalities ViaVest established in 2000 serving businesses in Stryn. ViaVest collaborate with Telenor renting capacity on copper wire but are mostly providing access via wireless technology. Vik IT-Partnar established in 2002, in collaboration with the electricity provider Sognekraft AS they provide wireless broadband access to large parts of the municipality Vik. They also serve central parts of the neighbouring municipality Balestrand. Vik IT Partner is the only commercial providers in both these municipalities. Zet.no was one of the first companies in the region that offered internet access, initially using fixed line connection and dialup modems. The company was the first to provide broadband access using wireless technology following the liberalisation of the telecommunication market in 1998. They offer wireless broadband access in five towns across the whole region rdalsnett established in 2001 with 12 local owners. Utilizing existing cable TV infrastructure and wireless technology to deliver broadband access. Kommunenett Nordfjord a collaborative initiative between seven municipalities Bremanger, Eid, Gloppen, Hornindal, Selje, Stryn and Vgsy delivering broadband services to public sector organisations in these municipalities at the same time as they are collaborating on a number of services built on top of the broadband network. Kommunenett Sogn a collaborative initiative between the municipalities Sogndal, Vik, rdal, Lrdal, Luster, Leikanger, Hyanger, Balestrand and Aurland and Sogn og Fjordane county municipality. The initiative includes a
199
common purchase agreement with regard to ICT services and delivery of broadband access to public institutions in the municipalities. Sunnfjord nett a collaborative effort by the municipalities Solund, Naustdal, Jlster, Hyllestad, Gaular, Frde, Flora, Gulen and Fjaler and Sogn & Fjordane county municipality. The initiative includes a common purchase agreement with regard to ICT services and delivery of broadband access to public institutions in the municipalities. Firdanett further described below Kapasitetslaget further described below
All these initiatives cover each only a part of the Region Sogn & Fjordane and all are the result of local initiatives. In addition to the local levels there is a regional initiative called Broadband-forum Sogn & Fjordane29 (BBF); a network organisation, initiated in 2001, in response to the increasing interest and need for broadband infrastructure in the region. BBF is a task force that places focus on establishment, development and utilization of broadband infrastructure in local communities. BBF initiated and participated in a number of activities, such as; information meetings and an annual conference, surveying the availability of broadband infrastructure, and participated in a number of externally funded broadband projects focusing on the utilisation of the network for provision of services or education. The national broadband plan says that infrastructure development must be based on market forces. By ensuring the information flow and exchange of experience the forum is supporting the demand side and there through strengthening the demand for broadband service, the aim is to have a competitive market in most parts of the region. In the following I will focus on two of the local initiatives to in an attempt to look at the factors contributing to the development of the infrastructure.
KAPASITETSLAGET30 SOGNDAL MUNICIPALITY

Sogndal is one of the main agglomerations in the region. The municipality has about 6,600 residents and an additional 2,000 students during the school year. Sogndal has a diverse business structure. Agriculture is important, and so are the manufacturers who use local raw materials in their products, such as Lerum Conserves, which turns fruit and berries into preserves. The community has one of the largest upper secondary schools in the country and is home to the Sogn og Fjordane College and the Western Norway Research Institute. Sogn og Fjordane College is located in many buildings spread throughout Sogndal; a broadband infrastructure between these buildings has been built over the last 10 years.
29 30
http://www.it-forum.no/pub/menu/mid9-40.asp http://www.kapasitetslaget.no/
200
This infrastructure consisted of leased lines from the national telecom provider Telenor, and lines owned by the college itself and the municipality. When tele-liberalisation occurred in 1998 the whole infrastructure was purchased by the college and the municipality. In the period from 1996 to date the two largest public organisations in the region have participated in several European research and development projects together with the local research institute and college. The focuses of these projects were on a number of issues related to diffusion of ICT to local businesses and use of ICT to improve delivery of public services. In this process they got in touch with several small local technology companies. In 2000 the need for a better, faster infrastructure became apparent both to be able to provide services and for interaction with the technology sub-suppliers. For the individual organisation the cost-benefit was too high, but as an integrated group they saw that there could be an opportunity to obtain broadband access at less expensive rates. The two regional public sector organisations31 applied for public funding through a national program (HYKOM) and managed to secure the initial funds for the initiative, focusing on delivery of public services and on connection to providers of tools and services. Toward the end of the project the joint venture company Kapasitetslaget i Sogn AS (KL) was established, to continue the work. Twelve private companies joined efforts with the local energy company Sognekraft, the college, the research institute, and regional public sector organisations in this effort. The new companys main purpose was to negotiate inexpensive internet access and to stimulate the building of regional and local broadband infrastructure. They aim at coordinating the development of the ICTinfrastructure, and negotiating common internet access for all partners. The common infrastructure improves the communication between companies and offers a platform for collaboration and joint efforts among the small enterprises. The broadband network was extended to two neighbouring municipalities Leikanger and Luster in 2002 and presently more than 30 companies receive broadband internet access through KL. The 20 km line to Leikanger from Sogndal is rented un-terminated fibre lines (dark fibre). A new 30 km power and fibre line has been installed to connect Luster, and two local energy companies are renting out un-terminated fibre to KL. In summer 2003 a new piece of infrastructure was installed to connect Kaupanger, and it is owned by a local power company. One spin-off activity is the national portal norge.no which was located in the area because sufficient broadband capacity was available, in addition to the availability of skilled workers and the low cost of office space. Several new companies have been able to locate in the area due to the network capacity; examples of these are Asplan Viak Internett and ArtsPages. KL negotiates access to the national infrastructure at regular intervals. In negotiating they achieve sufficient access at all times at an acceptable cost, dependent on the needs of the customers. The experience is so far that working jointly in this way has led to cheaper and faster access for all the companies connected through KL.
31
The Sogn og Fjordane County Municipality and the Sogn og Fjordane County Governors office
201
FIRDANETT32 GLOPPEN MUNICIPALITY

Gloppen is one of 26 municipalities in the region; it has a population of 6000. The modern history of Gloppen is one of decline. In the past employment was in the agricultural sector and in the textile industry. The textile industry has been closed down. Farming is still an important industry (Statistisk Sentralbyr, 2002), but currently it is the public sector that employs the most people. The development of the local broadband initiative Firdanett is the result of several interacting processes. In 1996 the chamber of commerce was revitalised and a new strategy was published with the aim of changing the development of the community (Gloppen Nringsorganisasjon, 2001). This strategy included a reorganisation of the chamber of commerce, employing a full-time manager, and a process leading to a joint understanding of the current status and of the future challenges. As a part of this new strategy the chamber of commerce initiated co-operation with the municipality. This was welcomed by the local politicians, who also looked for new opportunities for developing the municipality. A Gloppen based ICT firm, Datainstituttet as, had several customers who wanted to buy additional ICT services such as backup, printing and server space. Several of these customers were located in the same building as Datainstituttet, and to these customers the services were made available through a Local Area Network (LAN) established in 1996. Due to regulations in the law, it was not possible to expand the LAN to other surrounding buildings. When the tele-liberalisation took place in 1998 the laws were changed and the LAN was expanded to other customers outside the building. This LAN and its services formed the basis for the development of the broadband network. To be able to improve the handling of the investments needed to develop the infrastructure, a new company, Gloppen IT nett was established. Gloppen IT nett was a collaborative effort between Datainstituttet, the chamber of commerce, the municipality, and the local energy company (Gloppen kommune, 2000), and was equally owned by the partners. The company continued to develop the LAN into a full broadband network. In collaboration with the municipality they secured public funding through a national program (HYKOM33) to be able to develop the network to include all public offices. The project was important to expand the network to get better coverage throughout the municipality. The development of Gloppen IT nett is considered an exemplary model (Samferdsledepartementet, 2002) for public-private partnership when it comes to development of broadband infrastructure. The operation of the Gloppen IT nett was handled by Datainstituttet. After about a year, the owners could not agree on a joint financial and strategic plan for the company. Datainstituttet sold their shares to the regional energy company. Gloppen IT nett was refinanced and its name was changed to Firdanett. Firdanett is mainly owned by the local energy company with the chamber of commerce and the municipality as minority shareholders. Firdanett has two employees, one technician and one salesman. In March
32 33
www.firdanett.no http://www.hoykom.no
202
2004 Firdanett merged with another broadband company Enivest. Enivest is in part owned by the same energy company; the effect of this merger is not yet clear and will not be further addressed in this paper. Firdanett offers broadband internet access and IP telephony, and Datainstituttet offers their ICT services to the businesses connected to the network along with other service providers such as accounting. The local broadband infrastructure consists of a mix between fibre technology and radio transmitters / receivers. Firdanett has access to national infrastructure through a 6 Mb connection through the national provider Telenor. As a result of this local initiative, 55 companies and 85 households receive broadband internet access [2003 figures]; despite the absence of the national provider Telenor, which at that time did not offer broadband to the households and the small businesses in the community. Co-operation between the public and the private sector, with the private sector as initiator, is characteristic for Gloppen (Gloppen kommune, 2000). Some spinoff activities are the Sandane Business Garden, which is offering an innovative environment for knowledge based firms and incubator facilities for start-ups. The establishment of a call centre with initially 40 jobs (it has been downscaled quite a bit recently), and a company providing internet based services related to digital images, Eurofoto, with eight new jobs are direct results of the broadband initiative in Gloppen.
Discussion
The following section will discuss the two cases in relation to the theoretical background to get a better understanding of the factors contributing to the development of broadband access in these rural communities.
THE DEVELOPMENT OF KAPASITETSLAGET

The initiators of KL had a common problem; they needed high capacity access to the internet to provide services at a national or international level. These services were generating heavy traffic both in the form of large files and the number of requests, e.g. norge.no, which has approximately 30,000 hits on its website every day. Having this focus made it easy for the initiators to recruit new participants and align them and their interests into the network [relational resources (Healey et al., 1999)]. In the enrolment process delivery of services were discussed with the result that this was defined as being outside the domain of KL, the important aspect was to keep it simple to ensure the support of as many actors as possible. Actors were enrolled based on a number of reasons, the main being expressed interest and need, access to knowledge and infrastructure. The actors represented private companies, public sector, third level institutions and research, having this combination was important to strengthening the innovative capabilities (Cooke, 1998) and the ability to succeed with the development. The enrolments of these actors have also made it possible to build on the installed base in the community both with regard to access to knowledge resources and infrastructure. Across the organisations taking the initiative, there was access to the knowledge and skills (Healey et al., 1999) needed to plan and establish the infrastructure. Together these 203
actors had the knowledge and relational resources (Healey et al., 1999) necessary to move forward and they were able to mobilise and realise the plans, by finding ways of funding parts of the development through a project. With the enrolment of the owners of existing infrastructure, KL was able to build on the installed base because they got access to the existing infrastructure. KL has been able to rent and make use of existing fibre infrastructure; this has been an important foundation for the success. The existing infrastructure that is owned by the municipality and the college was the basis for the development. Both these owners did not see any problems in renting part of the cable to KL for provision of broadband access to other actors. Furthermore the infrastructure that made it possible to connect Leikanger and Luster to KL has been built by other actors (the two local energy companies) and KL are renting these lines to provide services to its customers. The collaboration between KL and the local owners of the infrastructure has been a win-win situation. The win-win situation is based on the best possible utilisation of already available fibre infrastructure and the skills and technology needed to connect this infrastructure to the national internet infrastructure. KL provides the backbone fibre infrastructure to the internet; they do not provide any other services. Provision of other services has been left to the users of the infrastructure. The organisations that want to connect to the infrastructure can do so using radio link in some areas or by connecting to the backbone fibre at their own cost. Furthermore KL does not own the infrastructure but rents it from others, the only work necessary is to operate the switches which in normal situations are not very time consuming. To solve these issues the organisations owning the company have provided a pool of resources and skills so that they can be drawn upon in the operation of the infrastructure. These factors have made it possible to operate the organisation with a minimum of organisational resources and costs. The company does not have any employees, but are renting services from partners in the network and the competence to run the broadband network along with the infrastructure. Being a not for profit organisation the customers have to pay for the rent of infrastructure and for access to the internet.
THE DEVELOPMENT OF FIRDANETT

Prior to the establishment of the broadband network, key actors in the community had joined efforts and agreed on a joint strategy to stop the long-term decline of the municipality. The climate for innovation and change and the trust between the actors was improving. Development of a broadband infrastructure was one of the initiatives taken to achieve this. The initiative was taken by Datainstituttet who started to build an infrastructure to be able to offer services to their customers in a LAN. The liberalisation of the telecom market opened up for an extended network. Almost the entire infrastructure had to be established from scratch, only the LAN was available as an installed base. They wanted to provide fibre to customers in the centre of Sandane while other parts were connected using
204
wireless access. The cost of putting a fibre infrastructure in place was high and there was no short-term payback from the investment. Compared to KL they managed to enrol fewer actors in their network, both with regard to knowledge and infrastructure. The support in the community was narrower; the organisations involved had less economic power and represented relatively small organisations without a knowledge base that could be used as a pool in solving problem situations. They needed to have sufficient income to employ the necessary staff; this put a stress on the organisation. The ideas and targets of the different owners of Firdanett were not aligned and when the company got into problems this caused organisational stress which led to a reorganisation of the company. Another way of observing this development is that the relational resources and the mobilisation capability was sufficient, Datainstituttet sold their part of the company to the local energy company. As part of this reorganisation they aligned the different interests, separating access from services. The services were transferred back to Datainstituttet who included then in their line of products and services. Now Firdanett is offering competitive broadband internet access and IP telephony as their only services. The number of customers is increasing. In providing services to the community it is important to stress that both KL and Firdanett are successful developments from the communities point of view, while KL have had a much easier development Firdanett have struggled and overcome troubles several times, further it has proven difficult to make money from providing infrastructure.
Conclusion
This article presented two cases of local bottom-up initiatives to provide broadband access in rural communities. In the two cases different factors have been important for the development, these are summarised in the table below. Kapasitetslaget Regional public sector Enterprises College & research Experience running broadband networks Large organisations with a knowledgebase No employees - Human resources borrowed at no or minimum cost Renting infrastructure Not for profit Less expensive broadband capacity Firdanett Local public sector Enterprises General ICT skills Many small organisations with no knowledgebase 2 employees Building own infrastructure For profit Selling services Broadband capacity
Actors
Knowledge
Economy
Motivation/ alignment
205
Well aligned network
Infrastructure Developed infrastructure
Developing region At first, network not aligned after restructuring well aligned No available infrastructure
Table 1. Important factors contributing to the two developments Both these cases are examples of successful developments, they both reached the aim; providing broadband access in their local communities. The purpose of presenting both cases is not to compare them and say that one is more successful then the other, rather it is to illustrate that there are many, and different ways to fulfil such an aim. It is clear that if there had been no local initiative then it would have taken much longer to get broadband access in these communities. The extraordinary action taken to speed up the innovation process is not a diffusion process but a translation process (Latour, 1987). Because the development required that an unstable period when the artefact broadband access was assembled by patching together the installed base with new links using the skills and knowledge in the community. This assembly process is an example of a translation process as described by Latour (1987), in which the artefact broadband access has been patched together and fitted to the local context. By looking at the processes as translation processes we become more aware of the actions needed to build the infrastructure required to bridge the gap between the national infrastructure and the local needs. This analysis can not be used to make predictions in similar development but can be used to as a benchmark to improve the understanding and scale the expectations relating to this kind of translation processes. These processes are unpredictable; they are dependent on the kinds of actors involved. That is, the human actors; knowledge and relational resources, and the non-human actors; availability of infrastructure that can be utilized are important in determining the development path. The approach taken opens up for a degree of uncertainty as the events are interpreted and the interpreter base the interpretation on the knowledge they have about the case. A different researcher may come to a different interpretation given their knowledge about the case.
References
Amdam, J., (2000) Confidence Building in Local Planning and Development. Some Experiences from Norway. European Planning Studies, Vol. 8, No. 5, pp. 581600. Amdam, J. and Veggeland, N., (1998) Teorier om samfunnsplanlegging: kommunalt, regionalt, nasjonalt, internasjonalt. Oslo: Universitetsforlaget Asheim B.T. and Isaksen A. (1997) Location, Agglomeration and Innovation: Towards Regional Innovation Systems in Norway? European Planning Studies, Vol. 5, No. 3, pp. 299330. 206
Baskerville, R. and Wood-Harper, A. T. (1996). A critical perspective on action research as a method for information systems research. Journal of Information Technology, 11, 235-246. BBF (2004) Breibandsstatus Sogn og Fjordane, dated: 05.03.2004 http://www.itforum.no/pub/menu/mid9-40.asp Bijker, W. E., Hughes, T. P., & Pinch, T. J. (Eds.). (1987). The Social Construction of Technological Systems. Cambridge, Massachusetts: The MIT Press. Bourdieu, P. (2001). The forms of capital. In The sociology of economic life (Second Edition), eds. Mark Granovetter and Richard Swedberg, pp 96-111 Boulder, CO: Westview Press. (The original article, written in German, was published in 1983) Callon, M. (1999). Some elements of a Sociology of Translation: Domestication of the Scallops and the Fishermen of Saint Brieuc Bay. In M. Biagioli (Ed.), The Sciencer Studies Reader (pp. 67-83). New York: Routledge. Cole, R.E. (1991) Participant Observer Research. In Whyte, W. F. (1991) (ed) Participatory Action research, Sage Publishing Cooke P. (1998) Introduction. Origins of the concept. In Braczyk Hans-Joachim (ed.): Regional Innovation Systems. UCL Press. London, pp. 225. Diez J. R. (2000) The Importance of Public Research Institutes in Innovative Networks Empirical Results from the Metropolitan Innovation Systems Barcelona, Stockholm and Vienna. European Planning Studies, Vol. 8 No. 4, pp 451463. eEurope (2002) eEurope: Benchmarking national and regional e-business policies for SME, http://www.europa.eu.int/information_society/topics/ebusiness/godigital/benchmarking/d ocs/de_2002_06_12_final_%20benchmarking_%20report.pdf [Accessed 29.10.2004] Gloppen kommune (2000) Gloppen-modellen utbygging av digitalt allemannsnett i distrikta (Hykom rapport 1019/240). http://www.hoykom.no/hoykom/web_hoykom_prosjekter.nsf/677cffd27235dae8c1256d7 a0041af6f/efcfd7951176e01d41256823003ce603/$FILE/P1019%20Gloppen.doc [Accessed 29.10.2004] Gloppen Nringsorganisasjon (2001), Vedtekter Gloppen Nringsorganisasjon 27.3.2001. Available: http://www.gloppennaerings.org/vedtekter.doc. [Accessed 29.10.2002]
207
Granovetter & McGuire, "The Making of an Industry: Electricity in the United States". Pp. 147-173, in Michel Callon, editor, The Laws of The Markets, Oxford: Blackwell, 1998. Gregersen B. and Johnsen B. (1997) Learning Economies, Innovation Systems and European Integration. Regional Studies Vol. 31, No 5, pp. 479490. Healey Patsy (1997): Collaborative Planning - Shaping Places in Fragmented Societies, Macmillan Press Ltd, Houndmills, Basingstoke, Hampshire and London. Healey P., Magalhaes, C.de and Madanipour A. (1999) Institutional Capacity-Building, Urban Planning and Urban Regeneration Projects. Futura. Magazine of Finnish Society for Future Studies., Vol 18 No 3, pp 117137. Latour, B. (1987). Sience in Action: How to Follow Scientists and Engineers through Society. Cambridge, Massachusetts: Open University Press. Molka-Danielsen, J., Ohren, O., & Jger, B. (2003). Broadband infrastructure in the County of Mre and Romsdal and Rural Regional Development. Paper presented at the NOKOBIT 2003, Norsk konferanse for organisasjoners bruk av IT, Oslo, Norway. Monteiro E. and Hanseth O. (1995) Social Shaping of Information Infrastructure Development: On Being Specific about Technology. In Orlikowski Wanda. J., Walsham, G., Jones M.R. and DeGross J.I. (eds.): Information Technology and Changes in Organisational Work: Chapmann & Hall. London, pp. 325343. Monteiro, E. (1999). Actor Network Theory and Information Infrastructure. In M. Ciborra (Ed.), From Control to drift. The Dynamics of corporate information infrastructures.: Oxford University Press. Morgan K. (1997) The Learning Region: Institutions, Innovations and Regional Renewal. Regional Studies, Vol. 31 No. 5, pp. 491503. Putnam R. D. (1993) Making Democracy Work. Civic Traditions in Modern Italy. Princeton University Press. Princeton. Rogers E. M.(2001) Diffusion of Innovations. 4.edition ed. The Free Press. New York. Samferdsledepartementet (2002) Breiband i kommunane - Kva er breiband, kvifor er det nyttig og korleis kan det nyttast? http://www.dep.no/sd/norsk/publ/veiledninger/028021120002/dok-nn.html [Accessed 29.10.2004] Statistisk Sentralbyr (2002) Arbeidsmarkedsunderskelsen, 2. kvartal 2002. Statistisk Sentralbyr. Oslo 2002.
208
Sthr W.B. (1990) Global Challenge and Local Response. Initiatives for Economic Regeneration in Contemporary Europe. The United Nations University: Mansell. London. Trnqvist, G. (1997). Mnniska, teknik och territorium: NordREFO, Stockholm. Walsham, G. (1993). Interpreting information systems in organisations. Ch3. Organizational change: Context and process: John Wiley & Sons. Walsham, G. (1997). Actor-Network Theory and IS Research: Current Status and Future Prospects. In A. S. Lee & J. Liebenan & J. I. Degross (Eds.), Information Systems and Qualitative Research. (pp. 466-480.). London: Chapman Hall, London. Whyte, W. F. (1991) (ed) Participatory Action research, Sage Publishing Yin, R. K. (2003). Case Study Research. Design and Methods (3rd ed.): SAGE Publications.
209
The Emergence of Functional Standards: Digital Rights Management Standards in the Mobile Telecommunications Industry
Jennifer Blechar University of Oslo Department of Informatics P.O.Box 1080 Blindern 0316 Oslo, Norway jennifjb@ifi.uio.no
ABSTRACT
Current discussions of standards view standards primarily as the technical specifications necessary to enable large-scale infrastructural arrangements. Categorizations such as the International Classification of Diseases (ICD) or standards such as the Global Standard for Mobile Communication (GSM) are all essentially related to the manner in which objects (or people) technically communicate. While many of these discussions do emphasize the situated and local nature of standards development and implementation, little research seems to be focused on the nature or content of the standards themselves beyond the recognition that standards embody certain characteristics related to the environment in which they evolved. This paper investigates the Digital Rights Management (DRM) standards evolving for mobile devices to illustrate a new view of the nature of standards. These standards involve more than just technical specifications of transmission frequencies, protocols, bandwidth, etc. and represent, if stabilized, a new type of standard emerging in the mobile telecommunications sector, namely functional standards. This paper will also investigate the manner in which these standards are evolving in relation to previous standardization efforts in the industry. Based on this exploration, this paper will inquire into some possible implications of both the nature of the DRM standards and the manner in which they are evolving, in relation to wider issues of standards development and technology adoption.
INTRODUCTION
Research on standards is not new. As standards have become increasingly important, research regarding standards has grown. Typical areas of research regarding standards are related to concepts such as lock in, path dependence, network externalities and the
210
socio-technical nature of standards development (for example: Arthur 1989; Arthur 1984; David 1985; Katz and Shapiro 1985; Katz and Shapiro 1986; Star and Bowker 1999; Hanseth and Braa 2001; Hanseth and Monteiro 1997). This research has highlighted that standards are not purely technical artifacts but embed certain aspects related to the environment in which they were created (Star and Bowker 1999) and in which they are used (Timmermans and Berg 1997), and has been useful in identifying the sociotechnical complexity involved in the entire standardization process (Hanseth and Monteiro 1997). In this literature, even though the development and use of standards are viewed as non-technical and as being shaped by the environment in which they evolve, the standards themselves are viewed primarily as being composed of technical elements, for example, the International Classification of Diseases (ICD) or the Bridge Infrastructure in Norsk Hydro. Within telecommunications this is especially the case with most standards being related to bandwidth, radio frequencies, etc.. This paper will explore the Digital Rights Management (DRM) standards evolving in the mobile telecommunications sector. These standards seem to be extending beyond the definition of purely technical elements to include functional requirements related to the use of the technology/devices adopting them. Because of the inclusion of functional use specifications in the standards themselves, these standards illustrate, if stabilized, a new form of standards evolving in the telecommunications industry, namely functional standards. This paper will also explore the manner in which the DRM standards are evolving in the mobile sector, the processes that are being followed, the actors that are involved, etc. This process of standards development being followed for DRM standards will be compared to previous standardization processes in the mobile telecommunications sector, such as those which were followed for NMT and GSM. Some of these previous standardization efforts (for example NMT in the Nordic region) were successful while others, such as the first generation mobile systems in other regions were not. Lyytinen and Fomin (2002) attribute the relative success or failure of these 1G standards to, in part, the different institutional environments related to their creation in their respective regions. Therefore, this paper will explore the relative environment and approach for the development of DRM standards in order to highlight some possible consequences of the approach being used. Thus, the overall goal of this paper is twofold. First it will illustrate that the standards being developed in relation to DRM in the mobile telecommunications sector are a new form of standards within the industry and secondly that the development process being followed for the creation of these standards is different than previous standardization attempts in the mobile arena. Based on this exploration, this paper will inquire into some possible implications of the DRM standards and the standardization process in relation to wider issues of standards development and technology adoption. Information on the DRM standards being developed within the mobile telecommunications sector was primarily obtained from the current industry leader in the development of standards for mobile devices, the Open Mobile Alliance (OMA). Since the OMA is an open standards organization, information regarding the standards developed related to OMA DRM is freely available. This paper is based on the factual content of the OMA DRM standards, openly published through various OMA DRM
211
white papers and technical descriptions of enabler releases. The information regarding previous standardization efforts within the mobile sector has primarily been collected through documented research on these areas (a similar approach used by others in this arena such as Kiel 2002; Lehenkari and Miettinen 2002). The next sections of this paper will discuss the nature of standards according to current literature on the topic and the manner in which standards typically evolve, especially in relation to telecommunications. Here, standard setting processes which were followed for previous mobile telecommunications standards such as NMT, GSM, etc. will be presented. Following this, the OMA DRM standards and standards development approach will be introduced. The discussion will follow, highlighting the DRM standards and standardization approach currently being followed in the mobile telecommunications industry in relation to previous standards and standardization efforts. Possible implications of this will also be discussed. A short conclusion will then be presented.
WHAT IS A STANDARD?
According to Websters dictionary, a standard is something set up and established by authority as a rule for the measure of quantity, weight, extent, value, or quality. Bowker and Star (1999) define a standard as (among other aspects) being a set of rules for the production of objects that are deployed to make things work across distances and may be enforced by legal bodies. They focus on categorizations as a form of standards that are, on a greater scale, the key to underlying infrastructures in society. These categorizations (or standards) are viewed as boundary objects, enabling the communication within and between societies. Bowker and Star emphasize the ubiquity of classification schemes which surround our every day lives, such as, for example, the many underlying standards present in the simple act of writing a letter the content of the paper, the size of the envelope, the stamp, etc. (p. 37 ibid). They also highlight the material as well as symbolic nature of standards, or the need to view the context surrounding the creation and use of standards along with the standards themselves. Hanseth and Braa (2001) also discuss the context surrounding the creation and use of standards illustrating the manner in which corporate standards that are supposed to be universal are indeed localized when implemented. Thus, they point that standards which are to be universal, or complete satisfying the needs of all users, are rather local universals (Timmermans and Berg 1997). Meaning, universal standards are linked and integrated in the local environments in which they are implemented, thus these standards are not usable as is by all but rather embedded within the local context (Hanseth and Braa 2001). Hanseth and Monteiro (1997) also highlight that the standards underlying information infrastructures are part of a larger actor network and thus that they are not neutral and can embody certain use principles. The collection of works in Brunsson and Jacobsen (2000) view standards in the broad sense as being related to ideas and rules in society and address standards and standardization within the larger concept of control and regulation. For example, Brunsson (2000) argues that standards are a form of coordination and control, similar to 212
markets and organizations in that they form and organize peoples behavior. He points that standards surround much of what we do and that without them, cooperation and coordination would be much different. Other works in the book discuss various standards (such as ISO 9000 in Furusten 2000) and standard setting processes (Hallstrom 2000, Sahlin-Andersson 2000) in order to provide an overall introduction to the perspective of standards as rules in society. However, while these works, which embody a very broad concept of standards (including standards on such diverse subjects as how handball should be played, how advertisements should be designed, and how threatened bird species should be protected p. 3 ibid) do suggest that standards can be more than just technical specifications, they do not delve into the specific contents of the standards themselves and rather focus on the consequences of their creation and adoption. Standards have also been discussed from an economics standpoint. Grindley (1995), for example, discusses standards in relation to a firms strategic choices and competitive advantage in the market place. He specifies two distinct classifications of standards, namely quality standards and compatibility standards. Quality standards encompass certain product characteristics and minimum attributes such as weights, measurement, etc. and compatibility standards encompass complementary goods and services such as auto parts and user training and direct networks such as railways. His focus is on how these technical standards should be set, when they should be set and what actors should be involved in order to maximize and increase revenue to ensure success. Within mobile telecommunications literature, discussions of standards are typically related to debates on the implementation of the technical standards related to mobile telephone networks such as NMT, GSM or CDMA. This literature primarily focuses on the manner in which standardization was used for the creation of these systems (for example Lehenkari and Miettinen 2002) or the way the standards themselves evolved (e.g. the actors, motivations, processes, etc. involved) (for example Haug 2002; Funk and Methe 2001). While the development of the standards is seen as a socio-technical process involving the negotiation and alignment of the interests of various actors (see Lyytinen and Fomin 2002), the standards are primarily viewed and are related to technical specifications. For example, Fomin et al. (2003) discuss how design, sensemaking and negotiation interacted in the standardization processes of NMT, GSM and Bluetooth, yet they conceive standardization to be a process of creating a technical standard (p.2 ibid) and focus on the standards as the technical specifications they are. One exception to this purely technical view of standards is Nielsen and Hanseth (forthcoming) discussion of the standardization process involved in the development of the CPA platform in Norway. However, telecommunications literature and the standards they discuss have up to this point been primarily concerned with the technical specifications necessary to allow separate technical components to communicate. In general it is evident that the standards discussed in the literature are mainly described as technical interfaces. Categorizations such as the International Classification of Diseases, standards such as the Bridge infrastructure standard in Norsk Hydro and the Global Standard for Mobile Communication are all essentially related to the manner in
213
which objects (or people) communicate. These standards are also primarily related to the technical specifications necessary to enable large-scale infrastructural arrangements, whether this be related health or mobile communication. This literature is also primarily focused on understanding the context surrounding the standards the manner in which they are developed, the control they imply or the localized nature of their implementation.
THE DEVELOPMENT OF STANDARDS

From the literature above, it is evident that the development of standards involves many actors including governments, consumers, development organizations, firms, markets, etc. These actors can be involved in the standard setting process in a variety of ways. In general, standards are typically said to develop in one of three ways. Some standards are set by the market. Here, competing technologies are introduced and it is the market which decides the de facto standard (Keil 2002). Standards can also be negotiated through a voluntary consensus process between various industry and governmental actors or de jure standards can be set by the government through a regulatory process of some sort (Blind et al. 2002). The consensus process for developing standards usually involves standards development organization (SDOs) or other ad-hoc organizations and can be said to follow either and consortium model, where a group of vendors join forces to develop a standard, or a not for profit model. Some current SDOs involved in setting de facto standards include the American Standards Institute (ANSI), the International Standards Organization (ISO) and the Institute of Electrical and Electronic Engineers (IEEE). Keil (2002) also suggests a hybrid consensus-market model where a number of standard options are introduced by organizations into the market and the market decides the standard. In the development of mobile telecommunication standards (such as NMT, GSM, etc.), Funk and Methe (2001) discuss the hybrid system of markets and committees which has in the past lead to the development of these standards. For the development of NMT (1G) in the Nordic region, for example, a committee referred to as the NMT committee was formed. This committee involved Telecom Administrations of the Nordic Region (see Haug 2002) and developed the standard for 1G across all the participating regions (Denmark, Finland, Iceland and Norway). Because the operators within the NMT committee all had a monopoly in their respective market, they were not in competition with each other and thus could only gain by participating and collaborating in the development of NMT standards (ibid). This created a unique and cooperative atmosphere. The operators also had a lead role in the standardization process in that they analyzed and approved the standards which were to be used, thus increasing competition and forcing innovation among the competing manufacturers (Lehenkari and Miettinen 2002). This created a learning and innovative environment while also creating a competitive market for pricing (ibid). Although the development of NMT did involve various political, legal and economic negotiations, the legal aspect only pertained to licensing issues where the Nordic 214
countries needed to liberate their licensing practices related to the use of radio transmitting devices (ibid) and for the regulators (which were in the same administration as the operators) to essentially grant access to the market or make room for the new technology (Lyytinen and Fomin 2002). Otherwise, the NMT committee was able to work fairly autonomously (ibid) as there was nearly no national pressure on the NMT standardization process nor was there any pressure from standardization bodies (Haug 2002). The development of 1G mobile telecommunications standards in other regions, such as the US or Japan, were less successful than NMT in the Nordic region. These systems also involved a different developmental path than the committee based NMT. King and West (2002) describe the US operator, AT&T, as a monopoly (although technically it was not) and thus that it very closely resembled the governmental operators in the Nordic regions. AT&T was the leader in the development of wireline technology in the US and its priority was to advance the wireline telecommunications systems. Bell Labs (under AT&T) was a leader in technological development for technologies essential to cellular telephony, however, AT&T continued to be focused on wireline services, not seeing the need or demand for mobile telephony (ibid). The monopoly AT&T enjoyed in the US seems to have been a major factor in the innovation related to mobile technology in that this one organization focused on their core mission to provide wireline services and were not pushed to innovate due to lack of competition in the market (ibid). King and West (2002) describe this as the competency trap in that AT&T was trapped in their focus on their successful wireline services and were not able to realize or willing to adjust their focus to the new potential of wireless technology. The Federal Communications Commission (FCC) in the US was also not focused on mobile telecommunications and thus AT&T worked fairly autonomously during the early stages of wireless development. During the 1970s however, AT&T and the FCC began to face regulatory conflicts which resulted in a significant lag in the development of the US mobile telecommunications standard, AMPS (ibid). Therefore, even though the US had the technical lead over other global markets with the development of a test mobile system in 1977, it took nearly 6 years for this standard to be approved and mandated as the single standard by the FCC (in 1982) (Lyytinen and Fomin 2002). This delay in the approval of the standard caused the US to lag behind operators in other markets such as the Nordic region and thus despite their initial lead, the US was and still is behind in the advancement of mobile telecommunications (ibid). The 1G situation in Japan was very similar to that in the US as the Japanese operator Nippon Telegraph and Telephone (NTT) was also a monopoly (in addition to a government corporation). NTT was actually the leader in mobile telecommunications with its release of the worlds first commercial wireless service in 1979. Lyytinen and Fomin (2002) discuss how NTT was also similar to AT&T in that it did not believe there would be a large demand for wireless service. The NTT standard it produced was intended to be used by high state officials, thus, prices were high and the capacity of the system was very low (ibid). During the 1980s competing operators entered the market
215
(with competing standards) thus opening the market to increased competition. However the initial design of the NTT system and the capacity limitations of it prevented widespread diffusion of the technology (ibid). The above discussions have all focused on the standardizations efforts related to the first generation of mobile service. However, the standardization methods followed in the first generation in each of the respective regions seems to have continued for the most extent in the development of second generation systems. In the Nordic region, a committee similar to the NMT committee was formed for the development of the GSM standard34. Although the organization of the committee was slightly different and it was larger and involved more influence of the national and industrial industries, the concept surrounding the development of the standard was the same. Namely the standard was developed in the form of a cooperative multi-national committee resulting in one standard adopted by all (Haug 2002). This is in contrast to the US, which continued with its ad-hoc company based standardization process thus resulting in several standards in the market (King and West 2002). Japan also continued with the same standardization process as it followed with NTT in that the one standard which was developed was driven by the dominant mobile operator in the market (Kano 2000). In general, Europe and Japan were more oriented to the development of one common standard set through a committee (in Nordic) or through a dominant market leader (Japan) than the US, which relied more on the market for the selection of the standard (ibid). The development of standards for third generation systems has changed slightly from the development of first and second generation systems, one reason being the need for global rather than national standards. Here, standardization bodies are taking an increasingly active role and standardization forums such as the 3GPP35 have joined efforts with the European ETSI (Kano 2000), which accepted the CDMA standard as the global standard for 3G (Funk and Methe 2001). Governments have also become an increasingly important actor in the standardization process for third generation systems through their influence in the market, by for example deciding how many and to which organizations to award 3G licenses (Funk and Methe 2001). However, although some countries such as the US have moved more towards the use of standardization bodies for the development of standards, the way in which these standards are being released has not changed. Namely, Europe is focused on the development and implementation of one standard common for all (de jure standard) whereas the US continues with its strategy to release multiple standards to the market (de facto standards) (Kano 2000).
In 1989, the European Telecommunications Standards Institute (ETSI) was established. Some say the GSM standard was developed by this group (Kano 2000; Funk and Methe 2001), however others say that the committee formed to develop the GSM standard was able to work fairly autonomously from ETSI (Haug 2002). 35 3GPP is a partnership project which was established in 1999 and is comprised of national standards bodies (such as in the US and Japan) which are planning on adopting the UMTS or WCDMA standard for 3G (Kano 2000).
34
216
DIGITAL RIGHTS MANAGEMENT STANDARDS FOR MOBILE DEVICES

Having given an overview of the standards and standardization processes typically discussed and developed within the mobile telecommunications sector, the standards related to Digital Rights Management for mobile devices will now be presented. Digital Rights Management (DRM) essentially deals with the management and control of digital content and many committees, organizations and standardization bodies have been formed in an effort to develop standards for DRM. Within the mobile industry, one of the largest standardizations bodies dealing with DRM is the Open Mobile Alliance (OMA). The Open Mobile Alliance began in 2002 and developed its first candidate version of OMA DRM standards in November 2002. It is comprised of nearly 200 companies and includes members from all levels of the value chain, from mobile operators to device manufacturers to IT companies and content providers. Among the members of the OMA are companies such as Ericsson, Nokia, Microsoft, SK Telecom, T-Mobile, Vodafone, TeliaSonera, AOL TimeWarner, Sony, etc. The OMA is organized around working groups (see figure 1) which are aligned according to functional areas (although they are under the technical plenary), thus it aims to create mobile service enabler releases which are fully interoperable, independent of the underlying wireless networks, platforms, devices, etc. Maintaining an open organization is key to the OMA and they wish to have broad industry participation and adoption. Thus, most of the standards developed by the OMA are open and available to the public thus allowing both members and non members of the OMA to be able to comment on early version of the documents and provide feedback during the standardization process. Once the standards are developed, these are made publicly available through enabler releases. The OMA DRM standards are one of several enabler releases and standardization efforts undertaken by the OMA and is organized in the sub-working group called Download and DRM, which is part of the Browser and Content working Group (http://www.openmobilealliance.com/tech/wg_committees/bac.html).
Board of Directors
Technical Plenary
Technical Plenary Sub-committees
Requirements Group
Sub-Working Group
Architecture Group
Sub-Working Group
Security Group
Sub-Working Group
Interoperability Group
Working Group
Sub-Working Group
Working Group
Sub-Working Group
Sub-Working Group
Sub-Working Group
Sub-Working Group
Sub-Working Group
Sub-Working Group
Sub-Working Group
Sub-Working Group
Figure 1: OMA Organizational Structure (OMA 2003)
217
The OMA has thus far released two versions of their OMA DRM standards. The first release supported three levels of functionality - forward lock (to prevent content from being forwarded), combined delivery (where the content message contains the content and the rights) and separate delivery (which provides content encryption and supports content and rights arriving on the device separately) (see Figure 2).
Figure 2: OMA DRM V1 Standards (Larssen 2003 from OMA)
OMA DRM V1 was developed shortly after OMA was established and was intended to be an initial release of DRM standards to support limited value content. OMA released a candidate for OMA DRM V2 in February 2004. OMA DRM V2 will support, among other things, superdistribution of content, which is specified such that content and the rights to use the content are managed separately thus enabling peer to peer distribution of content while controlling its use by requiring consumers to obtain usage rights. This will open for many exciting and new business models in the market (see figure 3). The specification process for the OMA DRM V2 is currently underway. In addition to this second release, OMA announced the initiation of the Content Management License Administrator (CMLA), which is a licensing entity associated with OMA DRM V2 responsible for licensing and distributing the encryption keys and certificates to device manufacturers and service providers which are OMA DRM V2 compliant (Smith 2004). The CMLA consist of Intel, mm02, Nokia, Panasonic, RealNetworks, Samsung and WarnerBros. Studios (http://www.internetnews.com/ec-news/article.php/3307621).
218
The activities within the OMA, Download and DRM sub working group are essentially coordinated by the Technical Plenary. The Technical Plenary decides which deliverables are due, when (OMA 2004a). Any sponsor or member of the OMA can participate in the working group and it is up to each to decide the amount of participation. Thus, there is no minimum or maximum to how actively representatives can participate. The group also tries to coordinate with standardization bodies such as ETSI, although the extent of this coordination is not clear from openly available information. In relation to the development of the OMA DRM standards, one overall DRM requirements document was created which specified the use scenarios for OMA DRM V1 and V2 and the requirements. The requirements were divided into subsections of market requirements and engineering requirements. The market requirements were fairly short (10 requirements for V1 and 13 requirements for V2) whereas the engineering requirements were more robust, encompassing areas such as security, charging, streaming, privacy, etc. (OMA 2004b). Based on this requirements document, several specifications documents were/are being created related to the overall DRM architecture, rights expression language, content format, etc. Once these documents are created, they are submitted to the Technical Plenary for approval and then are released to the public. In addition, various test fests and other activities are then planned.
Content Issuer
Content Encryption Key
Rights Issuer
Browse to website and download protected content
Purchase rights and establish trust
Deliver protected rights object
Establish trust; Purchase and deliver rights object
Super-distribute content
Figure 3: Example OMA DRM superdistribution deployment model (OMA DRM 2003)
DISCUSSION
Standardization Process From the above, it is evident that that OMA DRM standards are developed through a process of aligning the interest of those actors (including content owners, device manufacturers, etc.) involved in the download and DRM sub working group of the OMA. At a high level, this approach seems to be similar to the committee based approach to 219
standardization used for the first and second generation mobile systems in the Nordic region, however the details of the interactions within this committee seem to be quite different. For the development of first generation mobile systems, the mobile operators themselves had control of the standardization process. In the Nordic region, the operators were responsible for coordinating and approving the standards produced and had the intention that the manufacturers had to be held at arms length (p. 103, Haug 2002) while the basics of the system were designed. These operators also had a monopoly in their respective markets thus shaping their coordination effort. In the US, AT&T was the only major player initially developing the standard as was NTT in Japan. Thus in each of these 1G cases the dynamics of the actors involved in the standard setting process were different than the ongoing negotiation of various actors interests that are being undertaken for the development of the OMA DRM standards. In addition, the NMT and GSM committees were responsible for developing one de jure standard which was accepted and adopted by all regions. This is also different than the OMA DRM standards. Here, the development of the standards does not mean that they must be adopted and the decision to participate in the standards development process coordinated by the OMA is a voluntary process and a strategic decision for firms (Grindley 1995). Thus, while the OMA is working to develop one de jure standard following an industry consortium model, there are other standardization bodies developing other DRM related standards (many of which include the same members as are involved in the OMA) and thus ultimately the standard will be selected in a de facto manner by the market. In general, the OMA DRM standardization approach seems to contain aspects of the committee based approach used in the Nordic region and the market based approach adopted in the US (or a hybrid model as Funk and Methe 2002 suggest). However, the actors involved in the committee and their dynamics seem to differ due to the consensus process being used in the OMA DRM approach and the openness of the standard (in the sense that any member of the value chain can participate with equal rights). Standards Standardization work and standards do clearly have consequences related to use. The meanings of standards are translated according to the context in which they are used (Star and Bowker 1999) and the actual content of standards can be modified as they are localized in the environment in which they are applied (Hanseth and Braa 2001; Timmermans and Berg 1997). In general, standards also can and do have use implicitly embedded within them. For example, one of the requirements of NMT was that the system had to support the use of a mobile phone outside of its home region (Haug 2002). Thus, the NMT standard can be said to have controlled the manner in which consumers were able to use their devices by stating that the devices should support International roaming. The same is true for standards related to the layout of the mobile phone, the size of the screen, etc. All these standards implicitly dictate the manner in which mobile phones are used.
220
However, the OMA DRM standards are quite different than the types of standards previously experienced within the mobile telecommunications sector as these contain specific functional use requirements. These standards very much control not only how devices or distinct technological components should communicate or the manner in which the devices are physically used, but also explicitly specify how consumers can use the devices adopting them. An example is the Forward Lock requirement in the first version of the OMA DRM which states that The Device is allowed to play, display or execute the content, but it cannot forward the object. (OMA 2003). This standard is not just specifying the technological interfaces between devices but is rather specifying how the consumer can use the device. This standard is essentially stating that a user will not be able to forward content to another user. The standard is thus explicitly placing controls on the user. This is much different than the NMT committees task to define interfaces between building blocks in the system and between the system and the telephone network (p. 102, Haug 2002), for example. Possible Implications The manner in which the OMA DRM standards are being developed and the fact that these standards contain many functional use requirements, can have some major implications related to their relative success in the market. Some of these possible implications will be discussed in this section. Lyytinen and Fomin (2002) highlight that successful development of wireless systems in the past has been dependant on correctly integrating and aligning the concerns of the systems builders and the entrepreneurs within the specific environment (with the correct weight for each). The question remains if the OMA DRM standardization approach is maintaining this correct balance. Since all actors in the value chain (except consumers) have the opportunity to participate in the OMA DRM standardization process and thus influence the resulting standards, it is not clear which perspectives will be inscribed in the final result. If too much emphasis is being placed on the perspective of the content owners, for example, the risk is that the OMA DRM standards will be too restrictive and will potentially fail to be diffused in the market. In fact, many consumers and consumer groups are very concerned that DRM will impose too many restrictions on a persons ability to use their legally obtained content (see for example http://europe-againstdrm.com/en/index.php?id=1, http://www.eff.org ). Thus, the manner in which OMA DRM standards are being developed runs the risk of placing too much weight on the perspectives of the most active participants in the group. This suggests that the OMA should be careful to consider the standards being developed and consumers interests in addition to industry leaders interests. The lack of consumers perspectives in the OMA DRM standards development process is important to note. The need to incorporate the views and perspectives of users has been noted in more than one discussion regarding the process of standardization (see for example Hallstrom 2000). However incorporating users in the standardization process is problematic and has been found to complicate and slow the entire evolution of the standards (Funk and Methe 2001). On the other hand, users must be willing to accept the
221
standards developed in order to allow for them to become a market success. Because the DRM standards contain many functional use specifications related to the manner in which users are able to use the devices adopting them, this again suggests a pragmatic approach is necessary for the development of the OMA DRM standards. Another important aspect to consider with regard to the actors involved in the OMA DRM standardization effort is that the decision for a firm to participate in a standardization activity is a strategic one (Grindley 1995). Thus, standards are an opportunity for firms to influence the market in a way which is beneficial for them. Many attribute Nokias involvement in the OMA, for example, as their attempt to secure their corner of the market by influencing the development of mobile DRM standards (Wireless Watch 2004). Naturally the same is the case for all firms participating in the standards development and thus the perspectives which are inscribed in the standards which are produced could again be disproportionately associated with a select few of the actors perspectives involved in the process. Actors may also be involved in more than one standardization effort (and indeed usually are in order to ensure they are not locked out of the market (Schilling 1998)), thus the OMA DRM standards may just be one of many DRM standards different organizations are involved in creating. The strategic mix that Lyytinen and Fomin (2002) discuss also highlights the role that the emphasis on service provisioning has played, thus that a strong emphasis on service provision combined with an appropriate approach relative to the specific environment can positively influence the commercial success of a system. In the development of OMA DRM standards it appears that the regulatory statutes are being given a higher priority than service. Regulation is one item which has not been explicitly discussed in this paper, however, the very concept of digital rights management stems from copyright laws and the need and desire for content owners to control their content in relation to these laws. Thus, regulation is a fundamental aspect of digital rights management if there were no copyright laws and all content was free and open to all, there would be no need for DRM in the form it is evolving today. DRM standards are thus very much based on regulation and in many ways can be viewed as a manner of regulating through the standards and technology being developed. It remains to be seen if the mobile services being adopted which use DRM technology will be accepted and become a commercial success in the market. However, the amount of regulation being translated into the standards and the manner in which the standards actually control the use of the technological devices and services adopting them could be an important factor in their eventual market success. The OMA DRM standards are also, as of yet, being developed as open and voluntary standards, thus essentially de facto. Since de jure standards which are appointed and recognized by national governments have a higher chance of becoming recognized and adopted as global standards (Funk and Mette 2001), it is questionable if the OMA DRM standards will be able to be successfully diffused. Thus, the current OMA DRM standards development approach has implications for the eventual ability of these
222
standards to generate the bandwagon effect (see Farrell and Saloner 1985 for a more detailed description of the bandwagon effect) which is related to the ability of the standards and the technologies and services adopting them to develop momentum or inertia (see Hughes 1987). If the standards are unable to develop and initial installed base of users, less products will be developed which support them meaning less services will be available which will mean less consumer demand, which will again reflect back on the products available and the installed base (see figure 4). With this regard, the nature of the OMA DRM standards, in that they contain functional use requirements, could play an increasingly influential role. Another issue which could relate to the successful diffusion of the OMA DRM standards is the existing freedom of use people have with their phones and their use of content on the internet. P2P networks on the internet, for example, have allowed digital content (especially music) to be easily shared and accessed by all those part of the network (for example with Kazaa and Napster). Technical savvy consumers have become comfortable with the concept of being able to share their content (even if much of this was illegally shared). OMA DRM standards, however, are now threatening to take away this freedom and prevent users from performing a function many have been able to do before with other devices such as laptops. This could have interesting consequences for the acceptance and eventual adoption of the devices containing DRM standards which impose such functional limitations.
Larger Installed Base More Complements Produced Further Adoption
Greater Credibility of Standard
Reinforces Value to Users
Figure 4 Standards Reinforcement Mechanism (p. 27 Grindley 1995)
CONCLUSION
223
This paper has illustrated that the DRM standards being developed in the mobile telecommunications sector represent a new form of standards in that they involve mainly functional use requirements. The development of these standards also appears to be following a new developmental path as compared to previous standardization efforts in the mobile telecommunications industry. Possible implications of these standards and the standards development approach have been discussed related to the general eventual adoption and diffusion of these standards in the market. While a direct treatment of adoption and diffusion of technology is out of the scope of this paper, the implications of this exploration suggest the need for further investigation with regard to the potential impacts of the OMA DRM standards and standardization process on the eventual adoption of the devices and services implementing them. Further research could be performed with regard to the OMA DRM standards themselves and the technologies and services adopting them (in terms of gathering empirical data) as well as with the diffusion of the technology in general. This will allow for a more thorough treatment of the topic and the ability to draw more concrete conclusions to the implications suggested in this paper as well as to provide suggestions for possible manners in which to address these issues.
REFERENCES
Arthur, W.B., Competing technologies, increasing returns, and lock in by historical events, The Economic Journal, 99, pp. 116-131, 1989. Arthur, W.B., Increasing Returns and Path Dependence in the Economy, The University of Michigan Press, 1994 Bowker, Geopffrey and Susan Leigh Star (1999), Sorting Things Out: Classification and Its Consequences, MIT Press, Cambridge, MA USA Blind, Knut, Bierhals, Rainer, Thumm, Nicholaus, Hussain, Kammal, Sillwood, John, Iversen, Eric, Van Reekum, Rik and Bruno Rixius (2002), Study on the Interaction between Standardisation and Intellectual Property Rights, Fraunhofer Institute Systems and Innovation Research, http://www.isi.fraunhofer.de/ti/Downloads/kb_final_report.pdf Brunsson, Nils (2000), Organizations, Markets, and Standardization, in Brunsson, Nils and Bent Jacobsen eds., A World of Standards, Oxford University Press, Oxford, UK. pp. 21-39. Brunsson, Nils and Bent Jacobsen (2000), A World of Standards, Oxford University Press, Oxford, UK. David, Paul A. (1985), Clio and the Economics of QWERTY, American Economic Review, 75(2), Papers and Proceedings of the Ninety-Seventh Annual Meeting of the American Economic Association (May, 1985), pp. 332-337
224
Farrell, Joseph and Garth Saloner (1985), Standardization, Compatibility and Innovation, The RAND Journal of Economics, 16(1), pp. 70-83. Fomin, Vladislav; Keil, Thomas; Lyytinen, Kalle (2003), Theorizing about Standardization: Integrating Fragments of Process Theory in Light of Telecommunication Standardization Wars, Sprouts: Working Papers on Information Environments, Systems and Organizations. Funk, Jeffrey and David Methe (2001), Market- and Committee- based Mechanisms in the Creation and Diffusion of Global Industry Standards: the Case of Mobile Communication, Research Policy, 30, pp. 589-610 Furusten (2000), The Knowledge Base of Standards, in Brunsson, Nils and Bent Jacobsen eds., A World of Standards, Oxford University Press, Oxford, UK. pp. 71-84. Grindley, Peter (1995), Standards Strategy and Policy: Cases and Stories, Oxford University Press, New York Hallstrom, Kristina Tamm (2000), Organizing the Process of Standardization, in Brunsson, Nils and Bent Jacobsen eds., A World of Standards, Oxford University Press, Oxford, UK. Pp. 85-99. Hanseth, Ole and Eric Monteiro (1997), Inscribing behavior in information infrastructure standards, Accounting, Management & Information Technology, 7(4), pp. 183-211 Hanseth, Ole and Kristin Braa (2001), Hunting for the Treasure at the End of the Rainbow: Standardizing Corporate IT Infrastructure, Computer Supported Cooperative Work, 10, pp. 261-292 Haug, Thomas (2002), A Commentary on Standardization Practices: Lessons from the NMT and GSM Mobile Telephone Standards Histories, Telecommunications Policy, 26, pp. 101-107 Hughes, Thomas (1987), The Evolution of Large Technological Systems in The Social Construction of Technology eds. Bijker, Hughes, Pinch, The MIT Press, Cambridge MA USA, pp. 51-82. Kano, Sadahiko (2000), Technical Innovations, Standardization and Regional Comparison a Case Study in Mobile Communications, Telecommunications Policy, 24, pp.305-321. Katz, Michael and Carl Shapiro (1985), Network Externalities, Competition, and Compatibility, The American Economic Review, 75(3), pp. 424-440
225
Katz, Michael and Carl Shapiro (1986), Technology Adoption in the Presence of Network Externalities, The Journal of Political Economy, 94(4), pp. 822-841 Keil, Thomas (2002), De-facto Standardization Through Alliances-Lessons from Bluetooth, Telecommunications Policy, 26, pp. 205-213 King, John Leslie and Joel West (2002), MA Bells Orphan: US Cellular Telephony 1947-1996, Telecommunications Policy, 26(3-4), pp. 189-203. Larssen, Sissel Henrietta (2003), Mobile Content Market & Mobile Digital Rights Management (DRM), ETSI Workshop, Mobile Commerce Transactions, SophiaAntipolis, 9th 10th of April 2003. Lehenkari, Janne and Reijo Miettinen (2002), Standardization in the Construction of a Large Technological System the Case of the Nordic Mobile Telephone System, Telecommunications Policy, 26, pp. 109-127 Lyytinen, Kalle and Vladislav Formin (2002), Achieving High Momentum in the Evolusion of Wireless Infrastructures: the Battle Over 1G Solutions, Telecommunications Policy, 26, pp. 149-170 Nielsen, Petter and Ole Hanseth (forthcoming), Standardizing public content service provisioning for mobile phones, submitted to MISQ special Issue on Standards Ole Hanseth and Eric Monteiro. Inscribing behavior in information infrastructure standards. Accounting, Management & Information Technology. Vol. 7, No. 4, pp. 183211. Open Mobile Alliance (2003), OMA Organization and Processes, OMA-ProcessesV1_1_2-20031028-A, http://member.openmobilealliance.org/ftp/Public_documents/OMA_Process_Documents/ OMA-Process-V1_1_2-20031028-A.pdf OMA DRM (2003), Open Mobile Alliance, Digital Rights Management, Short Paper, http://www.openmobilealliance.com/docs/DRM%20Short%20Paper%20DEC%202003% 20.pdf Open Mobile Alliance (2004a), OMA Work Programme and Release Handling Processes, OMA-ReleaseHandling-V1_1_1-20040127-A Open Mobile Alliance (2004b), OMA DRM Requirements, OMA-RD_DRM-v2_020040420-C
226
Sahlin-Andersson, Kjerstin (2000), Arenas as Standardizers, in Brunsson, Nils and Bent Jacobsen eds., A World of Standards, Oxford University Press, Oxford, UK. pp. 100-113. Schilling, Melissa (1998), Technological Lock Out: An Integrative Model of the Economic and Strategic Factors Driving Technological Success and Failure, Academy Management Review, 23, pp. 267-284. Schmidt, Susanne and Raymund Werle (1998), Coordinating Technology: Studies in the International Standardization of Telecommunications, MIT Press, Cambridge MA, USA Smith, Brad (2004), Music Industry Eager to Exploit Wireless, Wireless Week, February 15, 2004, http://www.wirelessweek.com/index.asp?layout=articlePrint&articleID=CA381615 Timmermans, Stefan and Marc Berg (1997), Standardization in Action; Achieving Universalism and Localization in Medical Protocols, Social Studies of Science, 27, pp.273-305 Wireless Watch (2004), Nokia Leads Bid to Control Mobile DRM Standards, The Register, 9 February 2004, http://www.theregister.co.uk/2004/02/09/nokia_leads_bid_to_control/
227
IT/IS Learning and Education
SESSION 5: IT/IS Learning and Education

Also this year, NOKOBIT has a track devoted to issues of learning and education. The first paper to be presented, is a paper written by Godejord, and presents a study of the consequences of using ICTs in high school education. Next, Jaccheri and sterlie investigate a course syllabus for empirical software engineering education.
Papers are presented in the following order: 1. "IKT skaper bde variasjon og lring - De frste tanker fra et prosjekt om elevers oppfatning av IKT som lringsverkty" by Godejord 2. Empirical Software Engineering Education" by Jaccheri and sterlie
IKT skaper bde variasjon og lring - De frste tanker fra et prosjekt om elevers oppfatning av IKT som lringsverkty
Per A. Godejord Frstelektor Samfunnsinformatikk Hgskolen i Nesna Seksjon for Informatikk Studiested Mo i Rana Midtre gt. 4 8600 Mo i Rana
http://it-mo.hinesna.no/~pag/pag/
Resym
Dette paperet legger frem de frste resultatene s langt fra mitt doktorgradsprosjekt om hvilken betydning IKT har som lringsverkty i den videregende skole og grunnskole, slik det oppfattes av elever i disse skoleslagene. Paperet sitt hovedfokus er p den videregende skole. Studien som her er beskrevet var en kvalitativ studie med Grounded Theory som hovedmetode, og fokuserte p oppfatninger av IKT som lringsverkty og ikke kvantitative mlinger. Underskelsen omfattet samtlige elever ved to videregende skoler i Nordland, samt en klasse ved en videregende skole og en 7. klasse ved en grunnskole i Sr-Trndelag . En stor takk til Nordland Fylkeskommune som har stttet prosjektet. De viktigste funnene s langt er at lrerens kvantitative bruk av IKT i undervisningen ikke ser ut til ha noen effekt p elevens oppfatninger om IKT har noen betydning for deres lring. Et knapt flertall av de spurte elevene i de utvalgte videregende skolene mente at IKT frte til at de lrte mer og lettere, og at dette skapte god variasjon i undervisningen. Men det var likevel en ikke ubetydelig andel som mente at dette riktignok skapte variasjon men gjorde ingen ting for deres egen lring. Begrepet variasjon var det mest fremtredene hos samtlige elever i den videregende skole. Videre kan det se ut som at elevens valg av hvordan de benytter seg av IKT som verkty har sammenheng med den enkelte elevs egen, individuelle, vei til lring. Fra januar 2005 vil 10 - 20 grunnskoler i Vesterlen og 3 - 5 videregende skoler i Nordland bli omfattet av prosjektet, samtidig som observasjonen i en klasse ved en videregende skole i SrTrndelag viderefres.
229
Innledning
Dette paperet beskriver frste del av et doktorgradsprosjekt og tar i hovedsak for seg bruken av IKT, med Internett som det dominerende verkty, i videregende skole. Paperet har ogs med data fra en 7. klasse sin bruk av pedagogiske spill i matematikk og geografi. Paperet beskriver en underskelse foretatt dels ved sprreskjema, og dels ved observasjoner og intervju. IKT har siden midten av 90-tallet vrt beskrevet som et viktig verkty for lring. I Statssekretrutvalget for IT sin rapport Den norske IT-veien Bit for Bit beskrives informasjonsteknologien som noe som kan gi betydelige endringer i mten vi lrer og henter inn informasjon p.36 Mitt prosjekt tar sikte p utforme en teori om IKT sin betydning som lringsverkty for elever i den videregende skole og grunnskole, basert p elevenes egen oppfatning av dette. Prosjektet pgr stadig og i lpet av vren 2005 vil 10 - 20 grunnskoler i Vesterlen og 3 - 5 videregende skoler i Nordland bli trukket inn.
De frste resultat
Det er viktig vre oppmerksom p at dette er de frste teorier, eller snarere frforstelse, rundt et begrenset resultat fra et prosjekt som enda ikke er ferdig. Kun respons fra totalt 620 elever er analysert, og det kun i en underskelsesperiode p et halvt r. Videre er det kun tre videregende skoler og en grunnskole som er involvert, hvorav kun en klasse fra videregende i Sr-Trndelag og en klasse fra en grunnskolen i Sr-Trndelag. Hovedvekten av respondentene er fra de to videregende skolene i Nordland. Fra januar 2005 vil 10 - 20 grunnskoler i Vesterlen og 3 - 5 videregende skoler i Nordland bli omfattet av prosjektet, samtidig som observasjonen i en klasse ved en videregende skole i Sr-Trndelag viderefres. De frteorier som jeg her fremlegger skal alts danne grunnlaget for videre studier og en endelig teori, og er ikke ment vre ferdige teorier som kan gi grunnlag for generalisering i den grad resultat fra kvalitativ forskning kan generaliseres. Ut fra de data jeg s langt har samlet og analysert, har jeg formulert tre hovedlinjer som jeg mente avtegnet seg nr det gjaldt IKT i den videregende skole, slik jeg tolket det: 1. Hvor ofte lreren benytter IKT i undervisningen ser ikke ut til ha noen sammenheng med elevens egne synspunkter p om IKT har noen betydning for dennes lring. Imidlertid ser det ut til at lrere med god kunnskap om hvordan man effektivt kan benytte IKT som verkty, for eksempel det ske p Internet, og som lrer sine elever dette, har en noe frre andel elever som gir opp bruken
36
Statssekretrutvalget for IT (1996): Den norske IT-veien Bit for Bit
230
av Internet og heller gr p biblioteket. 2. Det er et flertall av elever som mener at IKT frer til at de lrer mer og lettere, og at dette skaper god variasjon i undervisningen. Men det er likevel en ikke ubetydelig andel av de spurte som mener at dette riktignok skaper variasjon men gjr ingen ting for deres egen lring. Begrepet variasjon i forbindelse med IKT-bruk er det klart mest fremtredene blant elevene i den videregende skole. 3. Elevens valg av hvordan de benytter IKT som verkty kan se ut til ha sammenheng med graden av frigjring fra verktyet, den enkelte elevs egen, individuelle, vei til lring, samt de rammer lreren setter for kunnskapsbygging i forkant av IKT-bruken.
Underskelsen
Underskelsens frste del, som dette paperet omhandler, omfattet i utganspunktet 764 elever ved to videregende skoler i Nordland, men kun 592 responderte via det tilsendte sprreskjema. I tillegg ble det foretatt ikke-deltakende observasjon i en grunnkursklasse i samme fylke. Data ble ogs samlet inn fra observasjoner i en mindre klasse p 16 elever ved en videregende skole i Sr-Trndelag, og en 7. klasse p 12 elever ved en grunnskole i samme fylke. Siden mitt ml med dette prosjektet var, og er, teoribygging rundt IKT som lringsverkty slik det oppfattes av elever, valgte jeg variablene <Oppfattet lring>, <oppfattet variasjon> og <hyppighet av IKT-bruk fra lrerens side>, som mitt hovedfokus ved prosjektets start. Variabelen <oppfattet lring> krever en definisjon av hva jeg legger i lring. Det finnes en rekke definisjoner p lring, og avhengig av pedagogisk ststed vil lrere og forskere ha ulike oppfatninger av dette begrepet. Elevene hadde kanskje, slik jeg tolket det, et noe enklere syn p hva som er lring. Og slik jeg s langt oppfattet de elever jeg var i kontakt med, ble det satt likhetstegn mellom lring og faglig utbytte. Mange elever benyttet seg av forklaringen jeg lrer nr jeg har lrt noe nytt og knyttet dette da til det enkelte fag nr det var snakk om lring i skolesammenheng. Jeg har enda ikke kommet s langt at jeg har ftt forsikret meg om at dette faktisk er en dominerende holdning, og det vil vre et av flere viktige fokuspunkter i det videre underskelsesarbeidet. En slik definisjon vil antakelig ikke vekke srlig begeistring i pedagogiske forskningsmilj, men siden elever ogs rent faktisk mles etter hvor mye de har tilegnet seg av kunnskap i det enkelte fag er det ikke unaturlig underske elevens oppfatning av IKT som lringsverkty i forhold til begrepet oppfattet faglig utbytte. Det er etter min mening ikke interessant i denne sammenheng se etter lringsutbytte i forhold til det rent teknisk kunne beherske en datamaskin som sdan. Det betyr ikke jeg mener at alle elever behersker det tekniske, men jeg mener at dersom det skal ha noen hensikt hevde at IKT fremmer lring m dette sees i sammenheng med de fag elevene
231
skal gjennomg i skolen. Med andre ord: Oppfatter elevene at de lrer mer engelsk nr de bruker IKT? Dersom vi skal teoriforankre min tolkning av hva elever jeg har snakket med legger i begrepet lre, ligger Piaget sin definisjon av lring nrmest. Piaget hevdet at lring skjer nr det skjer en vekst i erkjennelsen og dermed i strukturen. Lringer er meningsfulle aktiviteter hvor det utvikles forstelse og ferdigheter gjennom handling. Og jeg finner det derfor naturlig tolke elevenes bruk av Jeg lrer nr jeg har lrt noe nytt som et uttrykk for en vekst i deres erkjennelse. Dette begrepet, og elevenes forstelse av begrepet, er imidlertid ogs et viktig moment som m avklares i det videre studiet. Min forstelse av lring er knyttet til den kognitive tradisjon, men er min forstelse den samme som elevenes? Da prosjektets fokus alts var p <oppfattet lring> og holdninger og meninger ikke alltid er like mlbare, var jeg i stor grad vrt avhengig av motta beskrivelser og fortolkninger fra de involverte for f kunnskap om prosessene som trer i kraft nr IKT benyttes, og for kunne danne meg en oppfatning av dette. Jeg fant det derfor hensiktsmessig velge en fortolkende tilnrming gjennom kvalitativ metode. Ved velge kvalitativ metode nsket jeg tilegne meg fyldige og varierte fortolkninger av hvordan IKT som lringsverkty ble oppfattet av elevene. Informantenes beskrivelser og tolkninger skulle s i neste omgang danne grunnlag for min analyse og tolkninger p samme mte som i "grounded theory": "A grounded theory is one that is inductively derived from the study of phenomenon it represents. That is, it is discovered, developed, and provisionally verified through systematic data collection and analysis of data pertaining to that phenomenon. Therefore, data collection, analysis, and theory stand in reciprocal relationship with each other. One does not begin with a theory, then prove it. Rather, one begins with an area of study and what is relevant to that area is allowed to emerge".37 Det finnes selvsagt ulike mter angripe det kvalitative studie p rent metodisk, og forskere av i dag gjr bruk av en rekke ulike metodiske fremgangsmter, der noen ogs blander kvantitative med kvalitative metoder i det som kalles triangulering. Dette at forskere benytter seg av ulike metoder og sledes blir en metode-shopper nevnes av flere forfattere. Wolcott 38 skriver i sitt pningskapittel Posturing in Qualitative Inquiry: Even in educational research, where we seem on the verge of canonizing it, qualitative research is not a field of study, and there is no clearly specified set of activities or identifiable group of specialists who practice them. To claim competence in qualitative research is, at most, to claim general familiarity with what is currently being done, coupled perhaps with experience in one or two particular facets (e.g., to be good
Strauss, A., and Corbin, J. (1990): Basics of Qualitative Research: Grounded Theory Procedures and Techniques, SAGE Publications 38 Wolcott, H.F. (1992): Posturing in qualitative inquiry, fra M.D. LeCompte, W.L.Millroy, & J. Preissle (Eds), (1992): The Handbook of qualitative research in education
37
232
at collecting and interpreting life histories, or to be a symbolic interactionist). Wolcott understreker dette ved fye til: The advice I offer to anyone new to qualitative research is to think like a shopper as you approach the dazzling panoply and claims assembled before your very eyes. Noen hevder at forskeren skal vre som en tabula rasa i forhold til sitt prosjekt 39og s bygge opp en ny teori basert p de innsamlede data. Andre hevder at forskeren tvert i mot m ha noen ideer p forhnd om hva han skal se etter.40 Jeg har s absolutt et sett med bde fordommer og synspunkter med meg i bagasjen, men har likevel forskt flge Glasers versjon hvor den pne, oppdagelsesorienterte tilnrmingen understrekes. I hans versjon er det "no veiled goal to tell them what to find and how to force it out of the data. The goal is to do research that allows the emergence of " what is going on".41 Men samtidig flger jeg Wolcott sitt rd og er en metode-shopper, som velger fra flere retninger, men ser p Grounded Theory som en god sttte i mitt eget arbeide. For samle inn data valgte jeg benytte sprreskjema, intervju, samtaler, observasjon og elevlogg. Dataene ble analysert parallelt med innsamlingen, og sentrale prosesser ble forskt identifisert ved hjelp av koding av data og utvikling av begrep, slik Grounded Theory beskriver. Denne prosessen er naturlig nok ikke ferdig og vil pg inntil prosjektet er avsluttet. Det frste jeg merket meg nr jeg var ute i observasjonene var elevenes reaksjoner nr lreren varslet at de skulle f benytte datamaskiner. I grunnskoleklassen var reaksjonene udelt begeistring, mens reaksjonene i klassen ved den videregende skole var mer avdempet. Jeg valgte kode dette som <forventning>. Lrernes oppfrsel var naturlig nok ogs viktig og jeg merket meg store forskjeller i hvordan de aktuelle lrerne forberedte elevene p IKT-bruken. Noen gav kun elevene en oppgave eller prosjekttema og lot deretter elevene fritt sette i gang med informasjonsinnhenting, noen organiserte timen stramt med klarlagte oppgaver og gruppesammensetninger, mens en lrer hadde i god tid fr IKT-bruken gjennomgtt tema som kildekritikk og hvordan man effektivt kunne ske p Internett. Jeg valgte kode dette som <rammer>, i betydningen de ulike rammer for IKT-bruken som lrerne la opp til. Nr s timene var i gang ble det klart at elevene hadde ulike mter forholde seg til IKT som verkty p. I den observerte 7. klassen var jeg tilstede gjennom en hel uke, der matematikk var satt i sentrum som en del av den skalte Abel-uka. Noe senere deltok jeg i en ordinr geografitime.
Glaser, B., and Strauss, L. (1967): The discovery of grounded theory: Strategies for qualitative research, Aldine De Gruyter/ New York 40 Strauss, A., and Corbin, J. (1998): Basics of Qualitative Research: Techniques and Procedures for Developing Grounded Theory, 2nd edition, SAGE Publications 41 Glaser, B.G. (1998): Doing Grounded theory: Issues and Discussions Mill Valley, Ca.: Sociology Press
39
233
Under Abel-uka skulle elevene f kunnskap om matematikk via en rekke ulike aktiviteter, som inkluderte bde bruk av matteboka, skattejakt og konstruksjon av varmluftballonger og papirfly. I denne sammenheng avvek jeg fra rollen som ikke-deltakende observatr og introduserte PC-spillene Matteraketten og Mattemysteriet. Jeg deltok imidlertid ikke i selve undervisningen og overlot til lreren foreta instruksjoner i bruken av spillet. Ganske raskt avtegnet deg seg et mnster der noen fokuserte p det lse matematikkoppgavene, mens andre fokuserte p det spilltekniske der rask klikking etter prv og feilmetoden gav dem adgang til neste niv. Dette siste var srlig utpreget blant guttene, mens jentene var mer opptatt av lse de matematiske oppgavene. Samtaler underveis gav varierte svar p om de syntes at de lrte brk, som var det matematiske tema, eller ei. Noe senere var jeg tilbake i klassen og denne gang var det geografi som stod p planen. Denne gang var det lreren som introduserte et PC-spill, Globetrotter, for elevene. Igjen foretok jeg passive observasjoner, samt samtaler med de ulike elevene og samlet inn loggboka med elevenes notater i etterkant. Ogs her var det varierende fokus p verktyet og ulike uttalelser underveis om hvorvidt de syntes spillet hjalp dem lre mer geografi eller ei. I denne klassen var det, slik jeg tolket det, en klart positiv holdning til IKT som verkty for egen lring. Her mente flertallet, i de avsluttende samtaler med meg, at de faktisk lrte mer matte og geografi ved bruke dataspillene Matteraketten, Mattemysteriet og Globetrotter, men her ligger kanskje oppfattet lringsutbytte svrt nrt nske om variasjon fra boka, og det er forelpig usikkert om elvene la det samme i begrepet lre mer som det jeg gjorde. I sine loggbker var de mer variert i sitt syn p lringseffekten av spillene. I de to videregende klassene var det Internett som var selve verktyet, og ogs her forholdt elevene seg ulikt til bruken av verktyet. I den klassen jeg selv var til stede som observatr, i faget engelsk, delte klassen seg nesten etter kjnn. Samtlige elever skte uten noen forutgende vurdering eller diskusjon av skemetoder eller bevisste valg av skemotor. Til tross for at faget var engelsk, benyttet de fleste KVASIR som skemotor. Etter kort tid gikk gav de fleste jentene opp benytte Internett og gikk p biblioteket. Noen av guttene fulgte etter til biblioteket, noen fortsatte lete p Internett, mens andre fant frem VG eller ringetoner til mobilen. I den andre klassen der faglrer var observatr, i faget samfunnskunnskap, delte elevene seg p en helt annen mte i forhold til IKT som verkty. Elever som foretrakk samarbeide med andre for lse en oppgave, valgte biblioteket og diskusjonen med sine medelever som sitt fremste verkty og datamaskinen kun som "skrivemaskin" nr oppgavebesvarelsen skulle ferdigstilles. Elever som foretrakk jobbe alene valgte derimot IKT, blant annet i form av Internet, som sitt fremste verkty nr oppgaven skulle lses. Disse elevene valgte kun g i dialog med lreren, mens de mer gruppeorienterte gikk primrt i dialog med sine medelever. Jeg har valgt kode disse ulike prosessene som <forholde seg til>, i relasjon til IKT som verkty. Dermed ble det viktig forske analysere hvorfor elevene i den ene klassen i SrTrndelag forholdt seg til IKT-bruken s vidt forskjellig fra klassen i Nordland. Basert p
234
det jeg kunne observere av lrer- og elevaktivitet i klassen i Nordland, kontra det min lrerobservatr rapporterte fra sin klasse, var det klare forskjeller i rammene for IKTbruken. Lreren i Nordland gav ingen opplring eller instruksjoner forut for IKTbruken, mens lreren i Sr-Trndelag ikke lot elevene f ta i bruk IKT, i betydningen Internett, fr de hadde grundig kjennskap til bde skemetoder og kildekritikk. Videre virket det som om elevene i klassen i Nordland var mye mer bunnet av IKT-verktyet enn elevene i klassen i Sr-Trndelag. Dette kodet jeg som <grad av frigjring>. Ut fra dette formulerte jeg flgende frforstelse: Elever som helt blir overlatt til sin egen kunnskapsbygging nr IKT i form av Internett benyttes, klarer ikke frigjre seg fra verktyet. Med dette mener jeg at de benytter det ukritisk, helt til de gir opp og gr p biblioteket. Elever som fr en stram innfring p forhnd og dermed ny kunnskap om effektiv bruk av verktyet, frigjr seg fra verktyet i den forstand at de n kan benytte det ut fra sine genuine lringsstiler. For f inn flere data sendte jeg ut et sprreskjema til to videregende skoler i Nordland som var rettet mot samtlige 764 elever p disse to skolene. 592 av disse leverte inn skjema. Her var det gjennomgende usikkerhet om de selv fikk kt forstelse for faget nr IKT ble benyttet. P sprsml om de syntes IKT frte til kt lring svarte likevel 307 av elevene ja, mens resten var negativ (181) eller usikker (36). De resterende besvarelser mtte kasseres pga. feil utfylling/ manglende utfylling. Jeg valgte kode dette som <oppfattet lringsutbytte>. De fleste av elevene var enig i at bruken av IKT i alle fall frte til variasjon i undervisningen. 17 av respondentene mente at IKT gjorde det artigere lre, men at dette likevel ikke hadde noen egentlig betydning for deres lring. Dette er et s vidt lite antall at jeg er usikker p om dette er en interessant trend, eller kun et utslag av at elevene har misforsttt sprsmlene. Jeg valgte kode dette som en utvidelse av begrepet <forventning> til <forventning om variasjon>. Den klareste trenden var her at det ikke virket som om det spilte noen rolle for elevens oppfatning av IKT som lringsverkty, om hvor ofte lreren brukte dette i undervisningen. En elev som svarte at lreren sjelden eller aldri brukte IKT, kunne likevel svare at bruk av IKT gjorde at eleven lrte mer og lettere, mens en annen elev kunne svare at IKT ikke hadde noen betydning for dennes lring selv om dennes lrer benyttet IKT ofte. Jeg kodet lrernes IKT-bruk som <kvantitativ bruk>.
Ut fra de data som til da var innsamlet fra observasjon, samtaler og sprreunderskelse, avtegnet det seg linjer som jeg valgte formulere p flgende mte: 1. Den <forventning> elevene har til IKT er avhangig av de <rammer> lreren setter for undervisningen og IKT-bruken. 2. Elevens <forventning om variasjon> er sterkere enn elevens forventning om <oppfattet lringsutbytte> 3. Elevene vil <forholde seg til> IKT-verkty p ulik mte, avhengig av de <rammer> lreren p forhnd har satt for undervisningen og IKT-bruken.
235
4. Elever som helt blir overlatt til sin egen kunnskapsbygging nr IKT i form av Internett benyttes, klarer ikke frigjre seg fra verktyet 5. <kvantitativ bruk> av IKT har ingenting si for <oppfattet lringsutbytte> Hvordan er s dette i forhold til andre forskningsprosjekt omkring IKT i lring? I det flgende vil jeg kun behandle funnene fra den videregende skole, siden det var altfor f respondenter fra grunnskolen med i prosjektets frste del til at de innsamlede data derfra kan tillegges vekt forelpig.
Diskusjon
Jeg skal i det flgende diskutere mine forelpige funn, eller snarere frforstelse, i lys av noen av de forskningsprosjekt som har vrt utfrt p dette omrdet, og gir frst en kort oversikt over de prosjekt jeg har valgt som bakteppe for dette. Prosjekt Elektronisk Ransel Dette prosjektet ble startet i 1997 som en del av omstillingsprosessen i Narvik kommune. Prosjektet varte i tre r, med eksperimentering i tre ulike klasser. Oppsummeringen i etterkant viste at prosjektet ikke hadde frt til vesentlig strre interesse for undervisningen enn tidligere. Etter at den frste nysgjerrigheten og gleden over f bruke data hadde gitt seg, falt de fleste tilbake til sine opprinnelige holdninger til undervisningen .42 I ITUs evalueringsrapport konkluderte man med at resultatene kunne tolkes i retning av at man hadde indisier p at IKT skaper et mer variert lringsmilj for elevene. Nr det gjelder mer negative forhold, ser det ut til at IKT kan forsterke forskjeller mellom svake og sterke elever. Videre konkluderer ITU med at IKT kan bidra til en bedre skole, men bare under gitte forutsetninger. Disse forutsetningene er knyttet til utvikling av skolenes pedagogiske praksis.43 PILOT-prosjektet PILOT-prosjektet ble igangsatt i 1999 med en intensjon om at skolene skulle prve integrere IKT i undervisningen p en fornuftig mte. 120 grunn- og videregende skoler deltok i dette prosjektet. Noen interessante tendenser fra dette prosjektet er at lrere og elever utvikler nye, komplementre samarbeidsrelasjoner, elevene fr nytte sin digitale kompetanse inn mot fagene og de blir kunnskapsbyggere gjennom at personlige opplevelser i lringssituasjonen blir til digitale skildringer som s distribueres p Internett. Dermed ble det en kende elevmedvirkning, som igjen skapte mer begeistring og bedre eiendomsforhold til lringsprosessene. Rune Krumsvik, en av forskerne tilknyttet PILOT, uttalte i bladet Utdanning at: Det har dermed vakse fram ei erkjenning
Johnsen, Thor (2001): Foredrag om Prosjekt Elektronisk Ransel, NKUL 2001 Ludvigsen, S., m.fl. (1998): Prosjekt Elektronisk Ransel, En evaluering av et pilotprosjekt med innfring av informasjon og kommunikasjonsteknologi i den videregende skole. ITU-rapport
43 42
236
av at skulane sin prosjektberedskap og den sosiale konteksten formar bruken av teknologien, og dersom dette skjer i det tradisjonelle klasserommet, har IKT ein hgst diskutabel verdi for elevane sin lringsprosess.44 IEA Second Information Technology in Education Study (SITES) 12-13 juni 2003 ble det avholdt en strre WorkShop i Budapest. Mlet var se p hvordan IKT kunne utjevne ulikheter og bedre lring blant elever innenfor EU-omrdet. En rekke studier beskrives og ulike syn p IKT som lringsverkty legges frem. Akkurat som i PILOT-prosjektet blir det fremholdt at IKT frst kan fungere nr lreren gjr noe med sine undervisningsmetoder: that from the available evidence it seems unlikely that ICT will be primarily beneficial for improving achievement per-se. Rather it seems more likely that the use of ICT flourishes when teachers change their pedagogical approach from primarily teacherdirected to more student-directed. 45 Impact 2 Impact2 var en omfattende studie i Storbritannia som omfattet 60 skoler. Prosjektet varte fra 1999-2002 og studien fokuserte p en bred bruk av IKT og konsekvensene av integrere IKT i skolene. Studien kom frem til en rekke resultater, men for korthetsskyld nyer jeg meg med sitere professor Angela McFarlane sin sammenfatning av studiet fra rets ITU-konferanse: ICT can make a contribution to educational outcomes for some learners under certain circumstances.46 Det jeg oppfatter som en av fellesnevneren for disse studiene er tolkningen av at lrerens undervisningsmetode er sentral. Hvordan ligger s mine frste funn, eller frforstelse, i forhold til de prosjekter som er nevnt over? Pstanden som er sitert fra professor Angela McFarlan om at IKT kan fre til lring for noen elever, under visse omstendigheter, ser jo i alle hyeste grad ut til samstemme med respondentene sitt syn, i den forstand at det er noks tett mellom de som mener at IKT ker deres lring, og de som mener at det kun gir variasjon i undervisningen. McFarlans bruk av begrepet visse omstendigheter er ogs interessant. Hva s med pstandene om at den tradisjonelle klasseromundervisningen ikke gir lringsutbytte av IKT-bruk, slik det beskrives fra PILOT-prosjektet 47og at nytten frst og fremst er knyttet til en overgang fra lrerstyrt til elevstyrt? 48 Slik jeg tolker disse studiene, virker det som om det er en felles konsensus at den konstruktivistiske metode er best i forhold til IKT-bruk, og at IKT som verkty frst gir lringsutbytte nr en fr elever til engasjere seg i en lringssituasjon. Den amerikanske
44 45
Utdanning Nr. 7 12.mars 2004, side 52 - 53 Pelgrum, W.J. (2004): Promoting Equity Through ICT in Education: Projects, Problems, Prospects 46 McFarlan, Angela (2004): ICT and Learning: Dust or Magic? ITU-konferansen Digital Agenda 2004 47 Utdanning Nr. 7 12.mars 2004, side 52 - 53 48 Pelgrum, W.J. (2004): Promoting Equity Through ICT in Education: Projects, Problems, Prospects
237
lringsforskeren Seymour Papert hevder at IKT frer til at elevene kan lre ved gjre (learning by doing)49 og dermed ta kontroll over sin egen lringsprosess. I grunnskolen og den videregende skole er det frst og fremst bruken av tekstbehandling, regneark og Internett som dominerer, samt en viss bruk av pedagogiske spill, og i den senere tid LMS. Det er srlig Internett som trekkes frem som det mest revolusjonerende innen bruken av IKT i lring, fordi dette muliggjr kontakt med andre, rask tilgang til ufattelige mengder informasjon og muligheten for at elevene selv kan lage fagportaler der de presenterer sin kunnskap. Og det er her vi ser, slik jeg tolker betraktningene fra blant annet PILOTprosjektet, at lreren fjernes som kunnskapsperson til fordel for den informasjon eleven kan finne p verdensveven, noe som ytterligere gir nring til forestillingen om at det er IKT som frer til endringen av lreren fra sage on the stage til enguide on the side.50 Sprreunderskelsen min gir ingen svar p dette, men mine observasjoner i klassene gir etter min mening et interessant, om enn svrt begrenset, bilde av dette. Engelsklreren i observasjonsklassen ved den videregende skolen i Nordland var svrt konstruktivistisk i sin tilnrming nr elvene skulle benytte IKT, og da srlig Internett. Elevene fikk som beskrevet tidligere kun et prosjekttema og ingen instruksjon i det ske effektivt p forhnd. Denne klassen skulle da, i henhold til pstandene i andre studier, ha vist en kt lringsglede fordi de selv fikk nesten total kontroll over sin egen lringsprosess og fikk benytte IKT. Dette var, slik jeg tolket situasjon, imidlertid ikke tilfelle. Situasjonen ble snarere preget av oppgitthet enn lringsglede. Sammfunnskunnskaplreren i den videregende skolen i Sr-Trndelag regner seg ogs som konstruktivist, men hadde en atskillig strammere regi p sin undervisning. Ingen av hennes elever i samfunnsfag fikk benytte datamaskinene fr de hadde gjennomgtt en grundig innfring i effektive skemetoder p Internett, hvilke skemotorer som gav best resultat avhengig av tema og opplring i kildekritikk. Deretter fikk elevene i frste omgang kun lov til benytte seg av ferdige skestrenger som gav begrensede, og faglig relevante, treff. Lreren hadde p forhnd sjekket disse skene. Frst mye senere i perioden fikk elevene lov til selv formulere sk. Dette alts i sterk motsetning til engelsklreren i den andre klassen. I utgangspunktet skulle man tro at lreren var fageksperten i klassen og den som skal besitte det faglige begrepsapparatet som kan hjelpe elevene til strukturere sine informasjonssk p Internett, ikke minst i forhold til hva som regnes som gyldig kunnskap i skolen. Lreren i samfunnsfag fylte, slik jeg tolket det, en slik rolle i forhold til elevenes bruk av IKT, mens lreren i engelsk ikke gjorde det. Dette med at lreren enten trer aktivt frem, eller holder seg i bakgrunnen, som ressurs nr Internett benyttes, berres ogs i PILOT-prosjektet, der en viss ambivalens kan spores hos elevene nr det
49 50
Papert, Seymour (1993): The Childrens Machine- Rethinking school in the age of the computer Pelgrum, W.J. (2004): Promoting Equity Through ICT in Education: Projects, Problems, Prospects
238
gjelder f hjelp til skeprosessen i form av pstanden om at lrerne heller ikke er s gode til ske. og er bra nr lreren forteller hvilke sider de kan g inn p.51 Det var alts i klassen der lreren var en klar fagperson at elevene, slik jeg tolket det, frigjorde seg fra selve verktyet og valgte forholde seg til IKT-bruken i henhold til sine individuelle lringsstiler. Jeg tolket dette som at bde kunnskap om det ske effektivt er viktig, i tillegg til kunnskap om det fag de jobber med. Dette sprsmlet berres av PILOT-forskerne fra prosjektet i Agder, der man antar at ogs fagkunnskap m ligge til grunn for at eleven skal kunne ske frem relevante sider p Internett.52
Konklusjon
Det er etter min mening vanskelig, og svrt tvilsomt, skulle konkludere ut fra de data som ble innsamlet og analysert i prosjektets frste del. Formulering av sprsml, samtalestrategi og observasjoner var i utgangspunkt designet for gi meg momenter til en frforstelse som igjen kunne danne basis for dypere underskelser i prosjektets andre del, som frst starter i januar 2005. Jeg synes likevel det er en spennende tanke at elevers IKT-bruk muligens kan kobles til deres egne, individuelle, veier til lring. Hvis denne tanken tler mtet med en strre underskelsespopulasjon i neste del av prosjektet, hva vil det i s fall ha si for hvordan vi tenker og handler omkring det bruke IKT i skolen? Dette blir kanskje legge hodet under bde en og flere giljotiner, men la meg likevel forske. Hvis vi godtar tanken om at IKT som verkty fungerer best i undervisning preget av konstruktivistiske tanker, blir det kanskje viktig stille sprsml ved hvordan lrerens konstruktivistiske holdninger kommer til uttrykk i rammene for undervisningen. Kanskje m lrerne, nr IKT skal benyttes, tre sterkere frem som fagpersoner som har kontroll med tilretteleggingen og strukturen p bruken, og da srlig nr Internett skal benyttes? Det kan virke, slik det beskrives i ulike forskningsrapporter og slik jeg opplevde det i mine frste observasjoner, som at noen lrere legger et altfor stort ansvar over p elevene straks IKT skal benyttes i undervisningen. Hvis dette er tilfellet, hva vil det ha si for hvordan elevene nyttiggjr seg av og oppfatter IKT som lringsverkty? Frer IKT ikke til konstruktivisme og problemorientert undervisning, men heller en faglig tilbaketrekning fra noen lrere sin side? Hvis vi ser p lringsteoretikeren Dewey, s mente han at elever p egenhnd var ute av stand til planlegge prosjekter og aktiviteter, og at de derfor trengte hjelp fra en lrer som kunne srge for kontinuiteten i lringsprosessen. I flge Dewey mtte et prosjekt vre et felles arbeid mellom elev og lrer. Dewy understreket at lreren mtte ha hovedrollen i denne type lringsmetode, i form av kunne gi bde veiledning og stake ut en fornuftig kurs for elevene.53
51
Simonsen, P.A.A. og Valvik, R. (2004): Elevenes fortellinger om IKT i skolen, PILOT Delrapport Agder Simonsen, P.A.A. og Valvik, R. (2004): Elevenes fortellinger om IKT i skolen, PILOT Delrapport Agder
2
52
2 Knoll, (1997). The Project Method: Its Vocational Education Origin and International Development, JITE Vol. 3.
53
239
Dersom de observasjoner jeg s langt har gjort viser seg ha gyldighet, vil vi kanskje fremover mtte understreke sterkere den betydning lreren har som fagperson i undervisningen, og at IKT-bruk m forutsette en aktiv veilederrolle fra lrerens side i trd med Dewey sine tanker. Dette betyr kanskje ogs at lrerne, nr Internett er det fremtredende verktyet, m gjre en innsats for skaffe seg kunnskap om det benytte dette p en effektiv mte, slik at flest mulig elever kan frigjre seg fra IKT-verktyet og utnytte det effektivt som et verkty for egen lring i de ulike fag og ikke primrt et middel til variasjon.
Litteraturliste
Glaser, Barney G., and Strauss, Anselm L. (1967) The discovery of grounded theory: Strategies for qualitative research, Aldine De Gruyter, NY Glaser, Barney (1992). Basics of Grounded Theory Analysis, Sociology Press Glaser, Barney (2001) The Grounded Theory Perspective: Conceptualization Contrasted with Description, Sociology Press Johnsen, Thor. (2001) Elektronisk ransel p godt og vondt , Magasinet Klikk 2001 Johannessen Tor Aa., mfl. (2002) Constructs Used by 17-19 Year Old Students in Northern Europe When Informally Evaluating their Teachers, European Educational Research Journal, Volume 1, Number 3, 2002. Karpati, Andrea. (Ed) (2004) Promoting Equity Through ICT in Education: Projects, Problems, Prospects. OECD/ Hungary Workshop, Budapest, Hungary, 12-13 June 2003 Knoll, Michael (1997) The Project Method: Its Vocational Education Origin and International Development, Journal of Industrial Teacher Education Vol. 3, 2002. Krumsvik, Rune (2004) PILOT inn for landing, Utdanning Nr. 7 12.mars 2004, side 52 53. Ludvigsen, Sten R., m.fl. (1998) PROSJEKTET ELEKTRONISK RANSEL, En evaluering av et pilotprosjekt med innfring av informasjon og kommunikasjonsteknologi i videregende skole. ITU-rapport 1998.
240
McFarlan, Angela (2004) ICT and Learning: Dust or Magic? ITU-konferansen Digital Agenda 2004 Simonsen, Pl Andr.A. og Valvik, Ragnhild. (2004) Elevenes fortellinger om IKT i skolen, PILOT Delrapport Agder 2, 2004 Wolcott, Harry F. (1992) Posturing in qualitative inquiry. Fra M. D. LeCompte, W. L. Millroy, & J. Preissle (Eds.), The handbook of qualitative research in education, Academic Press, s. 3-52
241
Empirical Software Engineering Education

Letizia Jaccheri, Thomas sterlie Department of Computer and Information Science Norwegian University of Science and Technology {letizia,thomasos}@idi.ntnu.no Abstract
We report about two iterations of an empirical software engineering course for PhD students. The two iterations share a common syllabus but are organized around different pedagogical strategies. We have compared the two course iterations by running a questionnaire based investigation.
1. Introduction
Empirical software engineering (ESE) is a sub field of software engineering aimed at applying empirical theories and methods for understanding and improving the software development process in organizations. We report on two iterations of a PhD level course in ESE; a seminar-based course held during the autumn of 2002, and a project-based course held during the spring of 2003. In this paper we evaluate which of the two approaches may be better suited for teaching ESE. To this end, we have empirically evaluated the two course iterations with the goal of finding the pedagogical strategy better suited for teaching ESE. ESE has relevance for both practitioners and researchers. For practitioners it is about evaluating tools and techniques for use in concrete cases [4]. Researchers run empirical investigations with the software development process as object of study. While it is equally important to teach ESE to each group of these two groups, focus of this paper is on teaching ESE to software engineering researchers, believing that our findings are equally applicable for teaching ESE to practitioners. There are some fundamental challenges for an ESE research project to succeed. First, researchers in general and PhD students in particular must be well acquainted with ESE methods. Second, ESE research is a cooperative activity within a research group. Third, the research needs to be relevant outside the research community. The research group must therefore have access, knowledge and familiarity with the software industry in order to study concrete and real situations, and to generate research questions relevant to the industry. Lastly, the research problems must also be relevant within the academic field of software engineering. Research problems have to be significant to both the international research community and the local industry. This means that research problems and questions must be shared and understood by the all members of the research team and by the industrial actors These are all challenges that need to be addressed by a PhD level course in ESE.
242
2. Empirical software engineering course

The course addressed in this paper is offered to PhD students in the field of software engineering. There have been two iterations of the course.
2.1 Common characteristics
The ESE course counts for 7.5 European Credit Transfer System (ECTS). It was run once during the autumn of 2002 at the software engineering group by the Norwegian University of Science and Technology (NTNU) for PhD students by the group, and once during the summer of 2003 at Simula Research Laboratory (http://www.simula.no) in Oslo for PhD students from both the University of Oslo and NTNU. Some of the attending students worked in ESE research projects, while some worked in other kinds of software engineering projects. All students had already had a general course on research methods in IT as part of their curriculum. The same syllabus was used for both course iterations. The syllabus was divided into three parts: motivation, method, and actual investigations [3]. Both iterations of the course ended with an oral exam with the teacher and an ESE expert outside of the university.
2.2 Differing characteristics
For the autumn 2002 iteration, students met two hours a week for 13 weeks. At each meeting, one of the students presented a paper from the syllabus. The teachers responsible for the course provided feedback and stimulated discussion around the paper. The second iteration of the course was organized as an event spanning over three days (21 hours). Here we list in chronological order the main tasks of the course. a) Students were informed about the syllabus and the course web page [3] one month before course start and were asked to read all syllabus before course start. b) Short introduction to the course content by the teacher. c) Introduction to the field of ESE by discussing examples of interaction with the Norwegian software industry by two Norwegian research managers. d) Group work with the goal of extracting the main issues from the syllabus. Each group had to find and list research hypotheses and data and their analysis. Moreover each group had to summarize one method to plan and perform investigations, and to summarize one investigation. e) Practical exercises coordinated by a performance and theater instructor which had the goal to introduce students to performance work and to stimulate cooperation among students. f) Production of a five minutes film advertising the ESE field. g) Choosing an actual study in the syllabus, characterize the investigation presented in the chosen actual study, according to the motivations, the method, the measurements, and the data analysis method. The goal of this exercise is simulate the planning and execution of an empirical investigation. h) Group work with the goal of obtaining a deep and critical understanding of the issues presented in the syllabus, also in light of the experience acquired during the mini
243
investigation. Each group will have to: choose one research hypothesis and its motivation (business, education, others); list and comment investigation planning and risks (what can go wrong); define (or reuse) guidelines for designing and running empirical investigations; choose one actual investigation. i) Essay writing. Each student had to write an individual essay to be handed in before the oral examination. The title of the essay assignment is Extract from [5] and from the other papers in the Methods part of the syllabus, the most important guidelines for designing and running empirical investigations. Characterize the investigation presented in the chosen actual study, according to these guidelines. This means that you have to comment about the motivations, the method, the measurements, and the data analysis method.
3. Course evaluation
In this section we evaluate the two course iterations. The planning and presentation is in a large part based on the guidelines provided by [8]. As these guidelines are specific to the experiment, we supplement the evaluation with techniques provided by [7]. The approach itself is inspired by [6]. Based on the two course iterations, we wanted to evaluate our course in order to plan its third iteration. Another goal of this evaluation was therefore to get general suggestions for improvement. We wanted to know what the better way of teaching ESE is: the seminar based or project based one? Our hypothesis was that: project based course are better suited for teaching ESE than seminar-based course with respect to the level of learning reached by students with respect to Bloom's taxonomy. Blooms taxonomy is reported in Table 1. The last column of the table (Sample behavior ESE) gives our interpretation of the taxonomy when applied to the ESE domain.
Table 1 Blooms taxonomy Level Knowledge Definition Student recalls or recognizes information, ideas, and principles in the approximate form in which they were learned. Sample verbs Write List Label Name State Define Sample behaviour The student will define the 6 levels of Blooms taxonomy of the cognitive domain. Sample behaviour ESE The students will be able to define the content of the different papers.
244
Comprehension
Application
Analysis
Synthesis
Student translates, comprehends, or interprets information based on prior learning. Student selects, transfers, and uses data and principles to complete a problem or task with a minimum of direction. Student distinguishes, classifies, and relates the assumptions, hypotheses, evidence, or structure of a statement or question. Student originates, integrates, and combines ideas into a product, plan or proposal that is new to him or her.
Explain Summarize Paraphrase Describe Illustrate
Use Compute Solve Demonstrate Apply Construct
The student will explain the purpose of Blooms taxonomy of the cognitive domain. The student will write an instructional objective for each level of Blooms taxonomy.
The students will explain the purpose of the give methods and investigations.
The student will be able to use one method for experimentation.
Analyze Categorize Compare Contrast Separate
The student will compare and contrast the cognitive and affective domains.
The students will compare and constrast different methods and investigations.
Create Design Hypothesize Invent Develop
Evaluation
Student appraises, assesses, or critiques on a basis of specific standards and criteria.
Judge Recommend Critique Justify
The student will design a classification scheme for writing educational objectives that combines the cognitive, affective, and psychomotor domains. The student will judge the effectiveness of writing objectives using Blooms taxonomy.
The students will design an investigation by choosing and perhaps combining different methods.
The students will judge the effectiveness of using ESE methods.
245
The goal of our investigation was to determine how to improve the way we teach ESE. The independent variable [8] of our investigation is therefore seminar or project based course. The dependent variable [8] is the level of learning reached by students with respect to the Bloom taxonomy.
3.1 Planning
The effect we wanted to measure was how much the students had learnt from attending the course. We decided to use Blooms taxonomy of learning as an ordinal scale to measure this. The columns Level, Definition, Sample verbs, and Sample behaviour, are taken directly from [1]. Column Sample behaviour ESE is our contribution. We decomposed the dependent variable in three categories, one for each part of the syllabus: motivation, method, and actual studies. We wanted to know how much each student had learnt from each part of the syllabus.
3.2 Design
We decided to implement our investigation as a questionnaire-based survey [5] that we circulated to all students attending an iteration of the course. The questionnaire had to be formulated in such a way that we could measure what the students had learnt and how well they have learnt. To this end we formulated five questions: When did you take the ESE course? (Autumn 2002 or Summer 2003?) Have you ever participated in an ESE investigation? If yes, prior or after attending the course? Our goal was to measure how much of the course contents the students had learnt from attending the ESE course, but we wanted to avoid the potential threats to validity connected with clearly defining what the teachers intended for the students to learn. To avoid these problems we, the researchers, would place the students within Blooms taxonomy ourselves. We formulated the following question in order to measure how much of the course contents the students had learnt: In your own words, what were the primary objectives of the ESE course? We did not ask directly what the students had learnt, but what they thought were the primary objectives of the course. We did this to make sure the students answered what they had learnt from the course contents, not what the teachers were supposed to tech them. To see how well the course succeeded in teaching the students the overall goal, to run empirical evaluations, we asked the following: On the basis of what you have learnt in this course, would you be able to plan and run an empirical investigation? Finally, we formulated the question: Do you have further comments on the course?
If we had asked the wrong questions, we hoped the students would tell us so in answering this question. Also, we were interested in feedback on how to improve the course, and hoped this question provided an opportunity for such feedback. 3.3 Data collection
246
Based on lists of the participants in the two ESE course iterations, we circulated the survey to all participants pr. e-mail. We included a introductory letter with the e-mail explaining why we were doing the survey, that we wanted to use the data for both improving the course and as material for a paper, that all responses would be treated confidentially, and a date within which we would like to receive the responses by. The responses to our questionnaire were on free form and are available at [3]. We e-mailed the survey to the seventeen students attending the two iterations. Two of the e-mail addresses we had were no longer in use. Of the fifteen persons receiving the message, we received response from twelve. A thirteenth attendant replied that the survey was received, but that he had not yet taken the exams for the course. We therefore decided only to use respondents that had taken the exams. Two additional respondents were disqualified. One was disqualified because he did not attend the course even though he was supposed to, and the second because he had not taken the exams yet. In the end, we had ten responses to analyze. The response was significant: twelve out of fifteen amounts to a respondent rate of 80%. The respondent rate for responses to analyze was also fairly good: 66,67%. Students from two organizations had attended the course iterations. The distribution between response from the two organizations was interesting. Out of twelve students from our department, eleven responded. Of five students from the other organization, two were not reached because their e-mail accounts were removed, one responded that he had not taken the exams, and one responded to the survey. There is therefore a significant misalignment in the data material. Nine respondents from our department (two were excluded as explained above) and only one from the external organization. The data from the survey and our analysis can be found at [3].
3.4 Data coding and analysis
The stories written by students provide a valuable piece of knowledge for those evaluating our course. To the question In your own words, what were the primary objectives of the ESE course?" students write answers like: Provide some insight into typical methods used in the field. or An overview of status, methods and theory related to ESE research, which seems to a large extent to directed at quantitative studies with statistical analysis. To the question On the basis of what you have learned in this course, would you be able to plan and run an empirical investigation? students write answers like: Yes. But more concrete skills about how to design questionnaire and how to analyze data should be learned. Not without more input, but I would have a better starting point for the planning phase. A lot of the work in the planning phase would consist of deciding how the investigation should be formed, and I think I have a better understanding about the basic concepts and where I could find out more about them. A detailed description of our coding process can be found at ???. Our findings are displayed in Figure 1.
247
Instance
S2003 A2002 S2003 A2002 S2003 A2002 0% 20 % 40 % 60 % 80 % 100 % No answer Knowledge Comprehension Application Analysis Synthesis Evaluation
Figure 1: Percentage distribution of coded??.
4. Conclusions
This work is in the intersection between ESE and software engineering education. Its main contribution is the description of a course in the field of ESE. First, the presented syllabus can be used as a basis for a dialog in the ESE community about which topics are of most importance when educating the future researchers in the field. Second, the two presented pedagogical strategies can be discussed further to find out which one or which combination of the two is most suited for which context. Another contribution is our customization of the Bloom taxonomy for software engineering field. It is relevant to the ESE community as several empirical studies are conducted with students as subjects[2]. Our work can serve as inspiration for those who want to evaluate the pedagogical effects of course where such experiments take place. There remain three open questions. First, that of comparing our course with other similar courses offered internationally. Secondly, a question which is not attacked by our work is: is it possible to teach PhD students how to cooperate with external actors? The final issue that we have not discussed is ESE education for practitioners and not only for researchers.
5. Acknowledgements
We thank all PhD students who participated to the course and who answered our questionnaire. We are in debt to Reidar Conradi and Monica Divitini who set up the first version of the syllabus and organized the first version of the course. We thank Tore Dyb for useful discussions and suggestions on the syllabus part. Special thanks go to Dag Sjberg and the Simula center group for giving us the possibilities to run the second version of our course and to make this course as a Norwegian National course.
Motivation
Method
248
6. References
[1] B. S. Bloom, Taxonomy of educational objectives: The classification of educational goals: Handbook I, cognitive domain. New York, Toronto: Longmans, Green. 1956. [2] J. Carver, L. Jaccheri, S. Morasca, and F. Shull, "Issues in Using Students in Empirical Studies in Software Engineering Education," presented at International Software Metrics Symposium, Sydney, Australia., 2003. [3] L. Jaccheri and T. sterlie, "Empirical software engineering Course," Trondheim - Oslo, Norway, http://www.idi.ntnu.no/emner/empse/, last accessed September 2004. [4] B. Kitchenham, "DESMET: A Method for Evaluating Software Engineering Methods and Tools," Keele University, Technical Report TR96-09, August 1996. [5] B. A. Kitchenham, S. L. Pfleeger, L. M. Pickard, P. W. Jones, D. C. Hoaglin, K. El Emam, and J. Rosenberg, " Preliminary guidelines for empirical research in software engineering," IEEE Transactions on Software Engineering, vol. 28, num. 8, pp. 721 -734, 2002. [6] M. Morisio, M. Ezran, and C. Tully, "Success and Failure Factors in Software Reuse," IEEE Transactions on Software Engineering, vol. 28, num. 4, pp. 340-357, 2002. [7] C. B. Seaman, "Qualitative Methods in Empirical Studies of Software Engineering," IEEE Transactions on Software Engineering, vol. 25, num. 4, pp. 557-572, 1999. [8] C. Wohlin, P. Runeson, M. Host, M. C. Ohlsson, B. Regnell, and A. Wesslen, Experimentation in Software Engineering: An Introduction: Kluwer Academic Publishers. 2000.
249

Nokobit 2004 Siste

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Nokobit 2004 Siste

Uploaded by

Copyright:

Available Formats

The 11th Norwegian Conference on Information Systems

Stavanger, 29. November 1.December, 2004

Program Chair and Editor Knut H. R. Rolland

Organizing Chair Bjarte Ravndal

Welcome to NOKOBIT 2004!

Welcome to NOKOBIT 2004!

Welcome to NOKOBIT 2004!

Knut H. Rolland, Program Chair

PAPER SESSION 1: Alternative Approaches to IS Development

PAPER SESSION 2: Information Systems Innovation and Adaptation

PAPER SESSION 3: IS Research and Issues of Collaboration

PAPER SESSION 4: IT Infrastructures and Standards

173 194 210

PAPER SESSION 5: IT/IS Learning and Education

Panelansvarlig er Arild Jansen, Avdeling for forvaltningsinformatikk, UiO.

Alternative Approaches to Information Systems Development

SESSION 1 Alternative Approaches to Information Systems Development

Papers are presented in the following order:

Alternative Approaches to Information Systems Development

A Workshop-Oriented Approach for Defining Electronic Process Guides A Case Study

Alternative Approaches to Information Systems Development

ELECTRONIC PROCESS GUIDES

Alternative Approaches to Information Systems Development

Alternative Approaches to Information Systems Development

Alternative Approaches to Information Systems Development

Defining Processes in a Medium-Size Software Company

A SATELLITE SOFTWARE COMPANY

Alternative Approaches to Information Systems Development

DEFINING REQUIREMENTS FOR AN EPG

DISCUSSING PROCESSES: THE PROCESS WORKSHOP

Alternative Approaches to Information Systems Development

Define input and output

Find related documents Delegate responsibility for implementation

Role-based reading of resulting process

Implement the process in EPG

Alternative Approaches to Information Systems Development

Alternative Approaches to Information Systems Development

Alternative Approaches to Information Systems Development

Alternative Approaches to Information Systems Development

Alternative Approaches to Information Systems Development

Alternative Approaches to Information Systems Development

Conclusion and Further Work

Alternative Approaches to Information Systems Development

Alternative Approaches to Information Systems Development

Alternative Approaches to Information Systems Development

Alternative Approaches to Information Systems Development

Skapende teater og utvikling av informasjonssystemer Likheter og forskjeller i prosess og produkt

Alternative Approaches to Information Systems Development

Alternative Approaches to Information Systems Development

Alternative Approaches to Information Systems Development

PROSESS FOR SKAPENDE TEATER

Alternative Approaches to Information Systems Development

PROSESS FOR SYSTEMUTVIKLING

Alternative Approaches to Information Systems Development

Alternative Approaches to Information Systems Development

Alternative Approaches to Information Systems Development

Alternative Approaches to Information Systems Development

Prveinstallasjon Testing av kunden Tilbakemelding om feil nske om forandringer

Alternative Approaches to Information Systems Development

Tabell 1. Skjematisk framstilling av utviklingsprosessene i skapende teater og systemutvikling

Alternative Approaches to Information Systems Development

Alternative Approaches to Information Systems Development

Alternative Approaches to Information Systems Development

Alternative Approaches to Information Systems Development