We have achieved similar results in the VIATRA2 model transformation project, where we try to visualize the model space using a similar Zest-based visualization.

About usefulness: I hope, the tool is useful for software debugging – it is a new tool right next to variables display, the various break- and watchpoints, etc. etc. etc. For getting an overview of data structures, I find it sometimes useful – e.g. once I had a huge array, and an element right next to it, and wanted to know, which element is it in the array (by index). Instead of manually traversing the array, I opened it using the visualization, and then clicked the outstanding item in the variables view, and I got the item selected in the visualisation, together with its index.

About your questions (suggestions): can you please give some details, how to understand your question. The tool _is_ used to visualize instance variables (not class structure, etc.). And I’m not sure, how you would like to display array variables instead of the current one. The problem comes possibly from the fact, that I’m too into my project, and do not see others needs as good.

but it crashes Eclipse….oh well.

Your project seems to be getting there. There’s a similar project that does this too called Fujaba, but it’s a bit more complicated: http://www2.cs.uni-paderborn.de/cs/ag-schaefer/Lehre/PG/FUJABA/

Do you think your tool is a good way for students to debug their data structures with?

If not, well just ignore everything below (they’re just “suggestions” reworked as questions)
————

Right now I’m programming a heap and I’m using your tool to make sure that all the nodes are ordered correctly, so how successful have you been (so far) in displaying instance variables within each node? Also I’m implementing graphs using adjacency lists, these ones are a bit more difficult to view since the primitive arrays are displayed as nodes. Am I ever going to be able to view primitive arrays as arrays or am I gonna have to use ArrayLists in order for them to be properly displayed in the visualization?

http://code.google.com/p/debugvisualisation/

On the other hand it is harder to decide how to include the elements inside the node. An idea was to include all textual information as text information, but in some cases (typically EMF model instances) it was unusable: see the screenshots of the issue http://code.google.com/p/debugvisualisation/issues/detail?id=8

On the other hand we plan to address this issue somehow, but the concrete way is hard to see – as of now.

Thanks for the DDD link, I will look at it in more details.

I assume you’re using Zest? If you wanted to get more detailed in the boxes you might have to go with a more raw GEF though…