SECOND LIFE AS A PLATFORM FOR PHYSICS

SIMULATIONS AND MICROWORLDS: AN EVALUATION

Renato P. dos Santos

ABSTRACT
Often mistakenly seen as a game, the online 3D immersive virtual world Second Life (SL) is itself a huge and
sophisticated simulator of an entire Earthlike world. Differently from other metaverses where physical laws are
not seriously taken into account, objects created in SL are automatically controlled by a powerful physics engine
software. Despite that, it has been used mostly as a mere place for exploration and inquiry, with emphasis on
group interaction. This article reports on a study conducted to evaluate the SL environment as a platform for
physical simulations and microworlds. It begins by discussing a few relevant features of SL and a few
differences found between it and traditional simulators e.g. Modellus. Finally, the SL environment as a platform
for physical simulations and microworlds is evaluated. Some concrete examples of simulations in SL, including
two of our own authorship, will be presented briefly in order to clarify and enrich both discussion and analysis.
However, implementation of simulations in SL is not without drawbacks like the lack of experience many
teachers have with programming and the differences found between SL Physics and Newtonian Physics. Despite
of that, findings suggest it may possible for teachers to overcome these obstacles.

KEYWORDS
Second Life, Physics teaching, virtual worlds, physics microworlds, computer simulations.


INTRODUCTION

There are hundreds of game worlds, e.g. World of Warcraft, specifically created for entertainment, and
over 50 different Multi User Virtual Environments (MUVE) currently available (Taylor, 2007),
created to simulate real life in some sense. Among them, SL, followed by OpenSim and Active Worlds,
stands out as the platform that offer more services and tools for developing applications with quality
(R. Reis, Fonseca, & Escudeiro, 2011) even if it is not the one with the biggest user population
(Taylor, 2007). This is the reason why we have chosen it for our research purposes.

METHODOLOGY

Initially, specific features of the SL environment, relevant to its use as platform for physics
microworlds (Papert, 1980) and simulations, will be discussed in detail, as a support for subsequent
analysis.

Afterwards, the SL environment as a platform for physics microworlds and simulations will be
evaluated, using the criteria established by Reis & Andrade Neto (2002). We will focus on features
relevant to the teaching and learning than on others more relevant in terms of computing.
A few concrete examples of simulations in SL, including two of our own authorship, will be presented
briefly in order to clarify and enrich both discussion and analysis.

2


RELEVANT FEATURES OF THE SL ENVIRONMENT

On February 2012, the world of SL was made up of thousands of uniquely named Regions (Shepherd,
2012) - virtual parcels of land representing an area of 256 m x 256 m - linked together to make the
Maingrid, Private, Homestead and Openspace areas that spread over 1,998.26 km
2
(Shepherd, 2012)
of virtual land, not much smaller than the country of Luxembourg. Each Region is rendered by a
single process running on Linden Lab servers (Simulator, 2010) that simulates a full rigid-body
physical dynamics, including gravity, elasticity, and conservation of momentum (Physics engine,
2008), and an accurate, polygon-level collision detection of all the objects in the column-space above
its land.

Each server is edge-connected to four other machines as a grid of computers. This grid computes a
simplified solution of the Navier-Stokes equations to simulate the motion of winds and clouds that
time-evolve across the entire world (Ondrejka, 2004a; Rosedale & Ondrejka, 2003). It also runs users
scripts and streaming routines to send back all the data needed to view the world to anyones client
who is connected (Ondrejka, 2004a). As a result, the SL Sun usually rises and sets each 4 Earth
hours always directly opposite a full Moon (llGetSunDirection, 2009), objects fall under the effect
of gravity, trees and grass blow in the wind and clouds form and drift (Ondrejka, 2004a). Therefore,
SL is a big simulation of an entire Earthlike world.

Once logged through the client software, SL users can enjoy the 3D scenery, fly, drive cars, interact
with other avatars, play games, or create objects. In fact, well over 99% of the objects in SL are user
created (Ondrejka, 2004b), what has been characterized as a shift of culture, from a media consumer
culture to a participatory culture (Jenkins, Clinton, Purushotma, Robison, & Weigel, 2006).

In a rapid tour through many virtual spaces dedicated to Science, however, one will almost only find
real-world replica, mere institutional presences, with traditional classrooms, virtual boards exhibiting
2D PowerPoint presentations, and video seminars. We agree with Kapp (2007) that this is certainly
not the best use of Second Life or any other metaverse. As we will see below, the real advantage is
using it to do innovative things that could not otherwise be done in a classroom that reach into the
pupils imaginations (de Freitas & Griffiths, 2009).

Differently from other virtual environments, SL offers interactivity features to add to the objects
through its Linden Scripting Language (LSL) (LSL Portal, n.d.). Its structure is based on Java and C
languages and it provides nearly four hundred functions, among which several of interest to Physics
studies in SL used e.g. to access the position or the velocity of the object in the region or to apply a
force to it.

However, some important points should be taken into account when one wants to build a simulator or
a microworld in SL. See also (dos Santos, 2009) for a deeper discussion on these points.

Time can be affected by different kinds of lag and, therefore, SL will definitely not give response time
down to milliseconds consistently for quantitative simulations. There are no fluids in SL (dos Santos,
2009): outside the SL oceans, water is a mere texture applicable to an object. Besides, friction, air
and water resistance, and buoyancy were not fully implemented in SL. Remarkably enough, light
cannot be contained by the walls of a room and, as a consequence, light simply is there in SL, as an
state of illumination, without any physical mechanism involved in its production or propagation, as in
the most primitive conceptions of light (Lindberg, 1981). Finally, SL does not even try to simulate any
physics beyond Mechanics, and, therefore, a designer may have to use creativity to simulate other
interactions like electromagnetic or nuclear ones.

We have already demonstrated that SL does not simulate a real-world Physics virtualization or the
Galilean/Newtonian idealized Physics (dos Santos, 2009). It was also noticed that certain physical
quantities such as mass, acceleration, and energy have a very different meaning in SL as compared to
real-world Physics.
3



It was further noticed that SL is lacking of some basic functions to provide initial conditions to
objects, e.g. their initial velocity and position, as simulators usually have (dos Santos, 2009). We have
mentioned there that to circumvent this limitation, in certain circumstances, the designer can use the
llRezObject function that sets the position and velocity of the object when rezzed. However, physical
objects are, in principle, affected by SL wind and gravity and, therefore, they will not keep that
velocity constant and will deviate from its rectilinear trajectory. This difficulty can be seen in the
following three concrete examples.

Oddprofessor Snoodle, a Physics teacher, was recently creating a replica of an air track for her
Museum and Science Center (Figure 1). She found herself unable to have the glider to move along a
track at a given constant velocity even if using buoyancy to cut down on friction.



Figure 1 - Oddprofessor Snoodles air track replica at the Museum and Science Center
Source: photo taken by the author

As another example, Greis & Reategui (2010) have built a simulator themed as a pair of amusement
park non-wheeled bumper cars running on rail tracks and performing head-on collisions (Figure 2). In
a private communication to this author, Greis manifested the same kind of difficulties in controlling
the velocity in SL as Oddprofessor Snoodle faced.

4




Figure 2 - Bumper cars head-on collision simulator
Source: (Greis & Reategui, 2010)

To investigate the feasibility of building physics microworlds in SL, we have built a cannon
1
that can
simulate both Newtonian Mechanics and its predecessor, Buridans Impetus Theory as per users
choice. We have described in detail elsewhere (dos Santos, 2010a) our difficulties to counterbalance
gravity and to achieve a rectilinear trajectory simulation for the cannonballs and our workaround for
these difficulties.


1
The operation of this cannon can be watched on video (dos Santos, 2010b, 2010c). It is freely
accessible in our Second Life Physics Lab via the link http://www.fisica-interessante.com/virtual-
second-life-physics-laboratory.html#Buridanian_Cannon
5




Figure 3 - Projectiles falling down due to script delay to counterbalance gravity
Source: photo taken by the author

Now, we proceed to evaluate the SL environment as a platform for physical microworlds and
simulations.

EVALUATION OF SL AS A PLATFORM FOR PHYSICAL SIMULATIONS

Following Reis & Andrade Neto (2002), we will give more emphasis to features relevant to the
teaching and learning process than to criteria more relevant in terms of computing like stability,
aesthetics, compatibility, portability, and the like. We had, however, to generalize their criteria, as
their work was focused its analysis on simulations of mechanical collisions.

1 & 2. Representation of physical quantities in real time
Physical quantities are not automatically represented by the SL client. Nevertheless, LSL provides
functions to access directly some of them, e.g. llGetPos and llGetVel that return the position and
velocity of the object in the region, respectively. Therefore, these quantities can be represented
directly, in real time, during the entire duration of the phenomenon, as highlighted in Figure 1 and in
Figure 4. There are no functions in LSL to access the value of some other relevant quantities, such as
kinetic energy, momentum, and so forth that can be easily calculated by the script from those basic
quantities, however.
6





Figure 4 Information panel in our gamesim Lander Simulator
2
.
Source: photo taken by the author

3. Explicit relationship between conservation of physical quantities and type of collision
As we have seen above, the script either has access to the relevant quantities or can calculate them.
Therefore, it is not a problem for the programmer to make the script to display some kind of warning,
depending on the quantity is conserved or not. This could lead the student to pay attention to the way
the experiment happened and make the desired association.

4. Representation of physical quantities both in scalar and vector form
LSL functions return the value of physical quantities in the correct scalar or vector form. We have
already demonstrated elsewhere (dos Santos, 2012a) how the script can make use of the SL resources
to represent them in correct form.

5. Is it a simulation of an actual experiment?
Reis & Andrade Neto (2002) argue that difficulties in student understanding, arising from differences
between experiencing a phenomenon through an actual experiment or through a computer simulation,
could be minimized if it represented an actual, easily accessible, phenomenon in students everyday
situations or laboratory. If we compare the collision simulators of Figure 1 and Figure 2, we have to
agree that the former is more "real" in the sense that it duplicates a well-known laboratory equipment.
However, the interactivity, the sense of 'presence', the participant observation from inside the
experiment, of the second simulation, surely provides a richer, more intense, and meaningful
experience to the student (Greis & Reategui, 2010). Likewise, the exploration of interactive life-size

2
The operation of this gamesim can be watched on video (dos Santos, 2010d). It is freely accessible in
our Second Life Physics Lab via the link http://www.fisica-interessante.com/virtual-second-life-
physics-laboratory.html#Lander_Simulator
7


molecules (Lang & Bradley, 2009) and of fractals (Lee, 2007) do not represent 'real' experiments and
much less accessible ones, but are innovative and stimulating experiences that definitely cannot be
made in a classroom or in a laboratory. Therefore, we allow ourselves to disagree with Reis &
Andrade Neto (2002) on this criterion, when referring to simulations in immersive 3D virtual worlds
like SL.

6. Manipulation of initial parameters, e.g. velocities, masses, and positions
This feature is very important because it assists students in understanding their relationship with the
physical concepts and phenomena. It is feasible in SL, as the script can use the LSL llRezObject
function to set the initial position and velocity of a physical object. To define the objects mass, it can
use the llSetScale function, since in SL it depends on the size and not the 'material' of the object, as we
have already demonstrated elsewhere (dos Santos, 2012b).

7. Variety in visualization of physical quantities
LSL and SL resources to build and manipulate textures and objects allow the script author to represent
them in various forms, including bars, graphics, animations, icons.

8. Ease of use of the program, stability, and portability
The computational effort exerted by SL grid of computers leads to undesirable but understandable
system instability and occasional falls. On the other hand, as almost all the processing occurs on
servers, the client software is small and easily portable to a variety of operating systems.
Unfortunately, however, the simulation itself cannot be exported as Java applets and be run remotely
in a web browser as allowed by other platforms, e.g. AnyLogic (Rodrigues, 2008).

In conclusion, despite its rich resources, we cannot say that SL is an easy-to-use platform. Most users
agree that a high learning curve exists for new users (Sanchez, 2009), which means that any proposal
of using of SL for teaching has to set aside several hours, just to have the students become familiar
with basic tasks, e.g. walking, pass through doorways, go up stairs, manipulate objects, and so forth.

CONCLUSION

After his study on students issues in learning in SL, Sanchez (2009) concluded that designers can
create a user experience that will build on the strengths of the virtual world while overcoming the
obstacles. We agree with Sanchez, and affirm that despite the restrictions and differences between the
SL and a 'traditional' simulator, SL shows itself as a viable and flexible platform for microworlds and
simulations. It surely requires some creativity to overcome the difficulties of implementation seen on
the above examples that would be minimized if it had a friendlier interface for builders of simulations.
On the other hand, the rich immersive 3D massively multi-user experience SL provides is pleasant,
engages the user to explore the territory, and, therefore, offers a number of advantages over a 2D
simulator. How much innovative, however, was Modellus (Teodoro et al., 1996) by providing the user
with a very accessible interface, that eliminated the need to program graphic interfaces, at a time when
most of the personal computer worked in DOS. Our vision is that SL is equally innovative as a 3D
accessible platform.

REFERENCES

de Freitas, S., & Griffiths, M. (2009). Massively Multiplayer Online Role-Play Games for Learning. In
R. E. Ferdig (Ed.), Handbook of Research on Effective Electronic Gaming in Education (Vol. 1,
pp. 779-793). Hershey, PA: Information Science Reference.
dos Santos, R. P. (2009). Second Life Physics: Virtual, real, or surreal? Journal of Virtual Worlds
Research, 2(1), 1-21.
______. (2010a, July 30). Alternative Physics in Second Life [Web log post]. Retrieved from
http://secondlifephysics.blogspot.com/2010/07/alternative-physics-in-second-life.html
8


______. (2010b, September 9). Second Life Physics - Cannon simulation Part 1: Buridanian Impulse
Theory [Video file]. Retrieved from http://www.youtube.com/watch?v=OM7ToMjDdWI
______. (2010c, September 9). Second Life Physics - Cannon simulation Part 2: Newtonian Law
[Video file]. Retrieved from http://www.youtube.com/watch?v=reneVbLYDA0
______. (2010d, December 16). Second Life Physics - Lander Simulator [Video file]. Retrieved from
http://www.youtube.com/watch?v=aJgY1n_zQCQ
______. (2012a, March 14). Drawing vectors for physical quantities in Second Life! Second Life
Physics blog. Retrieved March 14, 2012, from
http://secondlifephysics.blogspot.com/2012/03/drawing-vectors-for-physical-quantities.html
______. (2012b, March 05). Mass is not conserved in Second Life! Second Life Physics blog.
Retrieved March 5, 2012, from http://secondlifephysics.blogspot.com/2012/03/mass-is-not-
conserved-in-second-life.html
Greis, L. K., & Reategui, E. (2010). Um Simulador Educacional para Disciplina de Fsica em Mundos
Virtuais. RENOTE: Revista Novas Tecnologias na Educao, 8(2).
Jenkins, H., Clinton, K., Purushotma, R., Robison, A. J., & Weigel, M. (2006). Confronting the
Challenges of Participatory Culture: Media Education for the 21st Century (p. 72). Chicago, IL:
The MacArthur Foundation. Retrieved from
http://digitallearning.macfound.org/atf/cf/%7B7E45C7E0-A3E0-4B89-AC9C-
E807E1B0AE4E%7D/JENKINS_WHITE_PAPER.PDF
Kapp, K. (2007, January 16). Gadgets, Games and Gizmos: MMORPG in ICT Education [Web
log post]. Retrieved from http://karlkapp.blogspot.com/2007/01/gadgets-games-and-gizmos-
mmorpg-in-ict.html
Lang, A. S. I. D., & Bradley, J.-C. (2009). Chemistry in second life. Chemistry Central Journal, 3(1),
14-33.
Lee, X. (2007). Math in Second Life. Retrieved from http://xahlee.org/sl/sl_math.html
Lindberg, D. C. (1981). Theories of Vision from Al-Kindi to Kepler. Chicago, IL: The University of
Chicago Press.
llGetSunDirection. (2009, January 5). Second Life Wikia. Retrieved from
http://lslwiki.net/lslwiki/wakka.php?wakka=llGetSunDirection
LSL Portal. (n.d.). Retrieved October 29, 2008, from http://wiki.secondlife.com/wiki/LSL_Portal
Ondrejka, C. R. (2004a). A Piece of Place: Modeling the Digital on the Real in Second Life. Social
Science Research Network Working Paper Series. Social Science Electronic Publishing, Inc.
Retrieved from http://ssrn.com/abstract=555883
______. (2004b). Escaping the Gilded Cage: User Created Content and Building the Metaverse. New
York Law School Law Review, 49(1), 81-101.
Papert, S. A. (1980). Microworlds: Incubators for Knowledge. In S. A. Papert (Ed.), Mindstorms -
Children, Computers and Powerful Ideas (pp. 120-134). New York: Basic Books.
Physics engine. (2008, September 13). Second Life Wiki. Retrieved from
http://wiki.secondlife.com/wiki/Physics_engine
9


Reis, M. A. F., & Andrade Neto, A. S. (2002). Simulao de Colises Dirigidas ao Ensino de Fsica.
Acta Scientiae, 4(2), 7-19.
Reis, R., Fonseca, B., & Escudeiro, P. (2011). Comparative analysis of virtual worlds. Proceedings of
the 6th Iberian Conference on Information Systems and Technologies (CISTI 2011), Chaves,
Portugal, 15-18 June 2011, 1-7.
Rodrigues, J. G. D. F. (2008). An analysis and evaluation of discrete production systems: a simulation
based approach (Doctoral thesis, Universidade do Porto, Porto, Portugal). Retrieved from
http://repositorio-aberto.up.pt/handle/10216/59259
Rosedale, P., & Ondrejka, C. R. (2003, September 18). Enabling Player-Created Online Worlds with
Grid Computing and Streaming. Retrieved from
http://www.gamasutra.com/resource_guide/20030916/rosedale_pfv.htm
Sanchez, J. (2009). Barriers to Student Learning in Second Life. Library Technology Reports, 45(2),
29-34.
Shepherd, T. (2012, February 17). Second Life Grid Survey - Region Database. Retrieved from
http://www.gridsurvey.com/
Simulator. (2010, February 11). Second Life Wiki. Retrieved from
http://wiki.secondlife.com/wiki/Simulator
Taylor, D. (2007, October 14). Second Life in perspective: A round-up of 50 virtual worlds [Web
log post]. Retrieved from http://www.fabricoffolly.com/2007/10/second-life-in-perspective-
round-up-of.html


Renato P. dos Santos
Associate Professor
PPGECIM (Doctoral Program in Science and Mathematics Education)
ULBRA - Lutheran University of Brazil
Av. Farroupilha 8001
92425-900 Canoas, RS
Brazil
Email: fisicaianteressante@gmail.com