language parallax

I’ve been working hard on my demo, but unfortunately my blogging frequency is significantly diminished lately. My attitude towards the blog entries is a lot less casual lately — as if I were actually publishing ideas that I’m going to have to stand by in the future.

So I’m a bit reluctant to muse about various design tradeoffs. However — I would like to sketch out the next few months of work in a high level of detail:

One of the core objects in my system at the moment is the “CoevolutionaryEnvironment”. It’s where I’m experimenting with describing the interactions between the various metamodels/languages that make up the frameworks I’m using as examples. It’s primary function is to perform “structured co-evolution”, which requires interaction with the filesystem and with source code control (in this case, Subversion).

I hope to have a more detailed list of objects and methods by the end of January. What’s going on between now and then?

1. A week or two of work on CoevolutionaryEnvironment itself.

2. Modernizing the “platform”, including finding non-trivial ways of using more javascript.

3. Using BPEL as the DSL. BPEL is typically run on “engines”, but I still think it could serve as a motivating example for code generation, which means it can illustrate “structured co-evolution” just as well as anything.

4. Extending the set of transformations that participate in the co-evolution.

5. Basic views of all the data. Probably just a simple web app to view all the internals of the CoevolutionaryEnvironment, including history.

6. Potentially extending the viewer to be an editor as well. The initial goal is simply to created a canned set of transformations that demonstrate the theory, but adding some interactivity to the viewer is an obvious way to begin to explore how this stuff might look in a real IDE. This could make a lot of sense as an Eclipse plugin, so if I get this far I may rewrite the viewer so that the editor can leverage the Eclipse environment.

Not necessarily in that order.