''''A Quick Answer to 'What is Open Dialect?''''

Open Dialect is an Open Source Flex based Flash IDE.  Open Dialect is written in C# using the .NET 2.0 framework and has been ported via Mono and GTK+ to Linux and partially OSX.  The purpose of this software is to create Interactive Learning Environments, hence the name (Dynamic Inter-Active Learning Environment Creation Tool).  If you are interested in the goals and implications for the field of System Dynamics of Open Dialect continue reading, otherwise enjoy what amounts to (for the time being) an Open Source Flash IDE.


== ''The need for a System Dynamics based Open Source, Online Interactive Learning Environment Creation Tool.''' ==



'''Abstract:''' 


System Dynamics needs a tool, which allows modelers to create Interactive Learning
Environments that help modelers share the knowledge that they have learned by model building, with
users who only use their models. To do this they need tools that allow them to better represent the
relationships between model structure and model behavior. I have been creating a software tool called
Open Dialect, which allows modelers to create ILEs for the web that incorporate new model structure
visualization techniques. The purpose of this paper is to introduce you to Open Dialect, and to show
you that Open Dialect will allow modelers to create effective, intuitive, modern, aesthetically pleasing,
online single user ILEs that will most importantly help the close the gap in knowledge gained through
model building (the modelers) versus just model using (the users).

'''Introduction:'''

This paper outlines the construction of an ILE creation tool, which will allow for the creation of
visually appealing, simple to use, single-user and potentially multi-user web based ILEs that
incorporate new model structure visualization techniques. In this paper I will focus two main areas.
First, on the ability of this software to create ILEs that can reach the 1.3 billion people that are online
today (Miniwatts Marketing Group, 2008). Second, I will focus on how this software will help to
shrink the knowledge gap between modelers and model users by improving model and data
visualization techniques.

Current ILEs are generally lacking in many areas including: visual appearance and aesthetics,
web accessibility, operating system (OS) independence, and data visualization. The current method of
creating and using ILEs has served its purpose for the past 20 years, but it is time for ILE creation tools
to evolve and take full advantage of the power of the Internet and advances in computing in order to
better reveal model structure to model users. I believe that current ILEs are now reinforcing the
knowledge gap between modelers and model users, because generally current ILEs do not allow users
to easily access and learn about model structure, but are generally offered as the solution for model
based learning when group modeling is not feasible or possible.
As of the writing of this paper (May 2008) I have already created an Open Source tool called
Open Dialect, which allows models to be published on the web. Open Dialect can create ILEs that are
quite intuitive and aesthetically pleasing. All ILEs that are produced by Open Dialect are OS
independent and web accessible.

Sterman considers education and learning as two of the most important goals of system
dynamics (1994). In order to reach those goals the results of system dynamics research must be spread
to the general population. One of the best ways to involve the general population into modeling
oriented fields is through simulation or gaming (Alessi and Trollip, 1991). An example offered by
Alessi and Trollip are the Sim games (Sim City, Sim Earth etc). These games encourage learning
through experimenting with the model that exists underneath them through simulation. This allows the
user to learn about the system while gaming and simulating the model, rather then the traditional
approach of model construction.

It has been the hope of the System Dynamics field that ILEs would serve as the necessary push
to stimulate the large-scale use of models and system dynamics in society (Sawicka and Kopinasky,
2008). According to Sawicka and Rydzak the evidence to support the effectiveness of ILEs has not
been fully established yet (2007). Kopinasky and Sawicka identify group model building as the most
effective way to disseminate model based information, but they acknowledge the limits of this
approach when it comes to teaching en masse (2008). They suggest that ILEs ought to expand past
running simulations and experimentation to allow the user to explore model structure and function, in
order to learn what is causing behavior to happen rather then just what behavior is happening.

Davidesen suggests that ILEs should strive to include not only behavior, but structure because it
is the relationship between structure and behavior that is key to learning (2000). His main concern
seems to be model transparency, which is the idea that the ILE should not hide or abstract the
underlying model at all levels. He believes that models ought to be fully visible to users who desire to
learn about how the system is functioning. He thinks that the key to successful ILEs is it the ability to
accurately represent model structure and behavior relationships. It is his belief (and mine) that the true
power of system dynamics comes from its ability to relate system structure to system behavior.

Alessi takes a slightly different stance then Davidsen suggesting that always having the model
structure available does not always lead to optimal learning (2000). He suggests that in certain cases
having the model readily available in full detail is not necessary and can lead to slower learning. He
says that content should be matched to the target audience in order to ensure that confusion is avoided.
His major findings can be repeated most easily as “… conceptual simulations should have opaque
models, whereas procedural simulations should have both (opaque and transparent models).” Where
procedural simulations refer to for example flight simulators, where the model is not as important as
the results (Alessi, 2000).

A good example of model structure visualization helping to close the knowledge gap between
modelers and model users is Wheat’s MacroLab model (Wheat, 2007). He used Stella’s story-telling
feature, to teach basic macroeconomics using the feedback method and compared that to current
methods for teaching macroeconomics. He found that the students who were taught through the
feedback method knew more about the system then those who weren’t. This research is encouraging,
because it gives a concrete example of how model structure visualization can improve learning, and
close the knowledge gap between modelers and model users.

Much work has already been done on how to design ILEs in order to be effective under many
different situations. I will not go over this portion of the literature since I am not as much interested in
specific design techniques at this point, but rather I am interested in the utility of ILEs in general, the
tools used to create them and closing the knowledge gap between modelers and model users. As an
example of a future feature I would like to include in Open Dialect, I mention Stella’s story telling,
which allows modelers to let their audiences interact directly with models in a positive, controlled
manner (iSee Systems, 2008).

Alessi and Trollip have put forth that the web is the next logical step for the evolution of
computer based learning (2001). They state that there is much potential for the spread of ideas and for
learning via the web. In regards to data visualization, even Sterman in his text Business Dynamics
clearly states that data visualization is one of the next key steps in the development of system dynamics
(1994). Therefore, I consider my work to be important to the future of system dynamics as a field.
As of this date (May 2008) there are two other software suites, which have been released to the
general public that allows modelers to dynamically display the results of their models on the web. The
first is by a company called Forio Business Simulations, which in my opinion it offers a limited range
of possible ILEs with a significant cost (Forio Business Simulations). The second is Stella’s NetSim,
which as of this time I have no experience with, but I plan on gaining some right away (iSee Systems,
2008).

Generally the most important issue I have with current solutions is they do not go far enough.
They do not allow modelers to reach the largest audiences possible, they put unnecessary limits on how
modelers can show model results, and they prevent modelers from working outside the bounds
envisioned by the software designers. This I believe is one of the major reasons why ILEs that help
close the knowledge gap between modelers and model users have not yet been perfected. While there
are most certainly some exceptions to the above statement, I argue that in general it is true.

'''Problem Statement:'''

The major shortcoming in ILE development seems to be coming from the inability of current
software packages to allow modelers to teach the information they have learned by going through the
modeling process. I believe that one of the key causes of this shortcoming is the issue of data and
model structure visualization. Since learning and education are two of the most important goals of
system dynamics it is very important for those not involved in the modeling process to be able to
absorb and learn the knowledge created by going through the modeling process.

Most learning occurs during the model building process, not just by using the model. Arguably,
this is because model building requires an intimate knowledge of the connection between the system
structure and the system behavior. Model users typically are shown only half of this relationship, the
system behavior, and rarely are their attentions focused on the structure of the system. Current
software has addressed this problem but not sufficiently. Enabling users to better visualize structure
and its resulting behavior would go a long ways towards addressing the knowledge gap between
modelers and model users. By putting ILEs on the Internet, it allows this knowledge to be made
available to hundreds or if it's user-friendly, maybe thousands or more interested users.

'''Long Term Hypothesis:'''

I intend to show that Open Dialect allows modelers to easily construct single user web based
ILEs that incorporate novel visualization techniques, which help the ILE users better understand the
concepts the modeler was hoping to teach closing the knowledge gap between modelers and model
users. The key concept in this hypothesis is that users ought to have a better understanding of the
system that the ILE represents. To do this modelers will need to use novel visualization techniques
including: audio/visual interaction (movies and sounds), animation, or other more game like techniques
used in traditional computer based learning software. In order for this to happen modelers must be
easily able to use the tools offered by Open Dialect, and the user interface of Open Dialect should be
very familiar to modelers. It should draw on the user interface improvements and appearances in the
software packages they already use.

'''Results:'''

Many of the above stated goals of Open Dialect are well on their way to being achieved at this
time there exists a basic Open Dialect interactive development environment (IDE), which allows for
the creation of simple to moderately complex single or multi-user ILEs. A simple single user prototype
ILE can be viewed at www.T21NorthAmerica.com/PolicyResistance.html or seen in Figure 1.

Figure 1: A simple prototype ILE in the Safari Web-Browser.

Figure 2 shows the Open Dialect IDE; notice how ILEs are broken into pages. Each page also
has its own timeline complete with layers and frames so that advanced users can make complex
animations. By default this behavior is hidden in order to encourage novice users to use Open Dialect,
without being overwhelmed by superfluous information.

Figure 2: The Open Dialect IDE showing the prototype interface under design. You can see on the left
all of the components which can be incorporated into an ILE. Behind that is the Pages view which
shows a re-orderable list of all of the pages in an ILE. The bottom pane contains information on the
selected object, in this case the line graph object. The toolbars up top allow modelers to easily access
the most important parts of the program with ease.

Figure 3 shows the Event Wizard, which is a dialog that helps users write the necessary
Actionscript 3.0 code in order to handle events, and model interactions. In order for the ILE to know
how to respond to a click of a certain button the user must handle that buttons click event. The event
wizard allows users to code these events without knowing any syntax at all. All users need to know is
the logical sequence of actions to take place. As you can see the event wizard prompts users to enter
the necessary information so that the software can complete the action desired by the user. Also notice
how in the event wizard all of the attributes of all objects are potentially modifiable, this means that for
instance a certain model result could be used to move a rectangle across the screen, in other words it
opens up the possibility for new innovative visualizations that move away from simple line graphs,
sliders and buttons.

Figure 3: The Event Wizard showing code generated to Login to the Model during a click event. As
you can see this wizard offers the modeler many options on how to handle events, pay particular
attention to the fact that the program asks for the necessary information and then it constructs the
actual code seen on the right, which can be re-ordered by hand if necessary. Also notice the properties
pane, which allows users to access all of the attributes of any object in the entire document.
To compile and test an ILE all the user has to do is select the Publish SWF (or Test SWF) menu
item in the File menu which will build the currently loaded ILE and open it for testing in the local webbrowser.
The final step after testing is done, is to upload the resulting ILE (.swf) file to the appropriate
server running the ASP.NET Web Service, which runs the Vensim model.

'''Discussion:'''

I believe that current tools available to modelers do not allow, or severely hinder the
development of ILEs, which truly represent model structure, not just behavior. In Open Dialect it’s
possible to represent the state of the system using much more then graphs, charts and numbers. Open
Dialect allows the modeler to represent the state of the system through colors, shapes, charting, or any
combination of the above-mentioned techniques. In this way Open Dialect tries very hard to provide to
the modeler the necessary tools to reveal the structure of their model so that learning can be facilitated.
Right now the goals stated above have not been totally met, this software does not yet allow
people to easily enough create structure visualizations. Without this key component there is much
work to be done in order to automate the process and allow users to reveal the structure of their models
with the click of a button, or the drag of a mouse. Though, even without this ability, in its current state
Open Dialect does help modelers to create simple single user online ILEs, and by doing that I think it is
still encouraging learning.

Open Dialect can be and will become a very important tool for the modeler who wants to spread
the results of their work over the Internet. With Open Dialect ILEs that help to close the knowledge
gap between users and modelers are: relatively easy to create (and will become easier), OS
independent, visually appealing, up to date aesthetically, and dynamic. The goals of the Open Dialect
project are well within reach, especially within the next year.

Open Dialect is most certainly not done yet. Future features to Open Dialect may be the
addition of the ability to produce DHTML in addition to flash files allowing ILEs to be run on more
environments including smart phones that do not support Adobe Flash, particularly the iPhone and iPod
Touch. In addition, to adding more platforms Open Dialect may be customized to allow for the
production of desktop applications as well, using the open source Adobe AIR software development
kit. In addition, licensing issues permitting, Open Dialect will be able to interact with more modeling
software packages besides Vensim and OpenSim.

'''Conclusion:'''

Work on projects like Open Dialect is necessary in order to bring the ideas of systems thinking
to the general public. Since the Internet is such a major part of our society today, tools to publish
content to it are becoming more and more important if the audience for a particular project is the
general public. Open Dialect also has a lot to offer the System Dynamics community because it allows
user to create ILEs that are web accessible, modern, and reveal both the structure and behavior of a
model system with little effort.

Open Dialect will allow modelers to easily construct single user web based ILEs that
incorporate novel visualization techniques, which help the ILE users better understand the concepts the
modeler was hoping to teach. Hopefully this software can close the knowledge gap between model
users and modelers. Users will be the main beneficiaries of this work, because they will be able to
learn more about systems and systems thinking, capturing a larger part of the knowledge created by
modelers going through the modeling process. Modelers will also benefit from this work because they
will be better able to perform their roles as educators. Modelers will be able to easily use novel
structure visualization techniques in order to make the knowledge they have gained by going through
the modeling process more accessible to users in the form of an ILE.

'''References:'''
1. Alessi, S. (2000). Designing Educational Support in System-Dynamics-Based Interactive
Learning Environments. Simulation Gaming 31, 178.

2. Alessi, S. and Trollip, S. (2001) Multimedia for Learning. New York: Allyn and Bacon

3. Davidsen, P. (2000) Issues in the Design and Use of System-Dynamics-Based Interactive
Learning Environments. Simulation Gaming 31, 171-77.

4. Forio Business Simulations (2008). Forio Broadcast [Computer software]. San Franscio, CA.
Retrieved May 27, 2008. Available from http://www.forio.com/

5. iSee Systems. (2008). Stella (Version 9.0.1) [Computer software]. Lebanon, NH. Retrieved
May 22, 2008. Available from http://www.iseesystems.com/

6. Miniwatts Marketing Group. (2008) World Internet Users December 2007 Retrieved February
21, 2008, from www.internetstats.com

7. Sawicka, A. and Rydzak, F. (2007) Incorporating delays in the decision-making interface: An
experimental study. 25th International Conference of the System Dynamics Society, Boston,
Ma. 2007

8. Sawicka, A. and Kopinasky, B. (2008) Simulation enhanced descriptions of dynamic problems:
Initial Experimental Results. Eighth IEEE International Conference on Advanced Learning
Technologies (2008)

9. Wheat, D. (2007) The Feedback Method: A System Dynamics Approach to Teaching
Macroeconomics. The University of Bergen, Bergen Norway.