1.2 Overview of SVM

SVM simulates DCharts models in real time.

SVM takes the following steps to process a model description, specified on the command-line:

1. Read in the model description;

2. 1st parse: search for all the macros defined in the model and store them in the memory (macros with the same names as the parameters specified by the importing model or by the user on the command-line are ignored);

3. 2nd parse: construct all the necessary data structures in the memory according to the model description under all the other descriptors (for example, the state hierarchy is built by literally combining and understanding the description under all the STATECHART descriptors, and a transition is constructed for each TRANSITION descriptor);

4. initialize the model by executing its initializer (an arbitrary piece of Python code written under the INITIALIZER descriptor, which is empty by default);

5. simulate the model by running its interactor (an arbitrary piece of Python code running in a separate thread to accept user inputs and interact with the model, as discussed later);

6. wait until the interactor finishes, which means the user stops the simulation, or the model itself ends the simulation;

7. finalize the model by executing its finalizer ((an arbitrary piece of Python code written under the FINALIZER descriptor, which is empty by default).

More information and releases of SVM can be found at its homepage:

http://msdl.cs.mcgill.ca/people/tfeng/?research=svm