For JHAVÉ Developers

JHAVÉ development can take a variety of forms:

• Developing new visualizations using the GAIGS support classes. This is most common way that others have contributed to the project. We provide extensive materials (below) on doing this from workshops we have conducted during summers at Grand Valley State and at SIGCSE Symposiums. Developing in his fashion does not require that you check out all the source code from our CVS repository.
• Extending the GAIGS structures by inheriting from the base StructureType class. How to do this is described in Chapter 4 of the GAIGS-JHAVÉ Visualization System Manual. It will require that you check out all the source code from our CVS repository. See this README file for a description of how to do this and then build the entire project.
• Writing a new Visualizer. This entails implementing the JHAVÉ Visualizer interface. In effect, this will involve designing your own scripting language (as Guido Rößling of the Darmstadt University of Technology and Ville Karavirta of the Helsinki University of Technology have done for the Animal and XAAL scripting languages respectively). How to do this is described in Chapter 7 of the GAIGS-JHAVÉ Visualization System Manual. It will require that you check out all the source code from our CVS repository. See this README file for a description of how to do this and then build the entire project.

Developing Visualizations Using the GAIGS support classes

The development of new visualizations with the GAIGS support classes is the most common way that developers contribute to the JHAVÉ project. In brief it involves using the GAIGS support structures to implement the algorithm you want to visualize. These support structures come with a writeSnapshot that you can use to annotate your algorithm's code with interesting events, that is, points in your algorithm's execution where you would like to see a JHAVÉ-rendered picture of the state of a particular data structure. Here are materials we have on using the GAIGS support classes to do this:

• Materials for short version of the developer's workshop
  • PowerPoint slides
  • Sort.java -- the starting point for exercises
• PowerPoints on GAIGS XML and JHAVÉ from the week-long summer workshops
  • Introduction to GAIGS Scripting
  • Generating GAIGS Scripts I
  • Generating GAIGS Scripts II
  • Advanced JHAVÉ Topics
• PDF slides on advanced topics covered in the JHAVÉ Developer's Forum (Summer 2007)
  • Using dynamically generated synchronized pseudocode
  • Adding audio explanations to your snapshots
  • Using/writing input generators, allowing learners to provide input data sets to your script-producing programs
• A jar file of the XML generation support classes. In case you don't work with eclipse at your home institution, you can still use this jar file in your favorite programming environment. Just be sure to set your classpath to include it. For example to compile and run from the command line:

  javac -classpath support.jar:. your-program.java

  java -classpath support.jar:. your-class-file

Then to view the visualization as you did at the workshop, download the jhave.jar file and run in "debug" mode:

 java -jar jhave.jar -debug

• JavaDoc for the XML generation support classes.
• The current version of the GAIGS How-To Manual.

David Furcy's JHAVÉ POP

For those of you who attended our "developer's forum" in June, you're aware of the great progress David Furcy has made in developing JHAVÉ visualizations of linked list and pointer operations. See the latest documentation on JHAVÉ POP.

Other Recent Changes

Some changes have been made in the support classes based on the suggestions we received during the JHAVÉ workshop held in June. These changes should all be backwardly compatible.

• Arrays can now have "empty" contents. If an array location is set to null, "null" appears as the contents of the array cell (as before). But if the contents are set to the empty string, the cell appears as empty. Doing this before threw an error.
• Bargraphs now show associated "row" labels centered below their respective bars. This allows developers to show indices (or what ever values they like). These labels can be multi-line. You can see an example of this at work if you use "myles_gaigs_ds" as the list when you start the client, and then choose the bubblesort example.
• The new GAIGStext structure supports positioning and font-sizing better than the old GAIGSlabel. See the online documentation.
• The new GAIGSlegend structure allows the easy creation of color-legend tables for the colors you use in your graphics. See the online documentation.
• Captions for snapshots and data structures and the new GAIGStext structure can now have multi-colored text. Use the usual hex color delimiters with the string to change color on-the-fly. For example,

  fo\#FF0000ba\#00FF00ar

  would display the string "foobar" with the first two characters in black, the next two in red, and the final two in green.

We still have some other suggestions to implement, but this is what we have for now.

This work is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 3.0 United States License.