language parallax

I’ve been concentrating on earning earning rent money lately, so I’m not as much into the research or related technologies at the moment. But I do want to try to keep writing at least one post a week until I can return to it… which should be around mid-September.

I’ve been thinking a lot about the title of my paper from the last post, as well as the few lines of equations contained therein.

As for the first phrase in the title — “Structured Co-Evolution” — I’m pretty happy with it. I noticed a very small number of papers out there talking about “co-evolution”, but it still seems to be virgin territory for computer scientists. I’ve thought about whether “structural” or “structured” makes more sense. I think the latter is the right choice. It places more emphasis on co-evolution itself as the core concept. The fact that it’s “structured” is important, but it’s a property of the core dynamic rather than a core concept itself.

As for the equations, I guess the biggest question I have is about my choice of function names: “evolve” and “coevolve”. Originally the first was called “coevolve”, but then I had a hard time coming up with a name for the platform-driven version. I don’t like evolve/coevolve. What I’d like to try to do is generalize this to the point where there is a single function or structure called “coevolve”, which can be used in a few different ways.

I expect a large portion of the dissertation will be about exploring how to generalize this function. Especially when I consider how the original “project” function relates to the co-evolution of the artifacts in the system. I think the project will have a natural set of milestones such as this. I’m essentially viewing it now as a software engineering problem. I need to find a normalized representation of the behaviors I’ve defined. Then document it. And of course base it on open standards as much as possible (to what extent can existing transformation languages apply?).