As the API callbacks for the visual debugging have been incorporated into the code, we need to generate and compile our sample watch models again. Depending on your target language, set the Generation target platform to either ‘C#: API’ or ‘Java: API’ in the property dialog for the ‘WatchModels’ graph, and run the ‘Autobuild’ generator. Once the code has been generated and compiled, the generation process will automatically start the MetaEdit+ API tool (shown in Figure 2-16) to provide the SOAP server to send the external requests to.

After that the familiar test environment will start again and you can choose which watch model you want to test. Activating and trying out the watch model will now visualize the execution by highlighting the active state in the diagram (as in Figure 2-17). Thus, you can now run watch applications and follow their execution through the models.

How does it work, then? The basic implementation idea of our visual debugger is very simple: at the end of each state transition, a SOAP message requesting highlighting of the next state in the model is sent to the SOAP server running in MetaEdit+. MetaEdit+ processes the message, highlights the requested element and sends an acknowledging SOAP reply back to the client application. For more information about the actual implementation of the visual debugger, see Section 4.1.1 ‘Visual Debugging with C#’ or Section 5.1.1 ‘Visual Debugging with Java’ depending on your target language.