I was recently invited to help rework BasicCAS into a full-fledged Mathematica parser for the SymPy project, a challenge which I have accepted. The idea is that a Mathematica parser (along with a working backend, a much harder goal to meet, I’d say) would allow SymPy to benefit from the large body of free Mathematica code floating around out there. Part of the reason I’m taking this on is because I’d not like to see SymPy turn into a melange like Sage– the idea of tying together disparate CASes and mathematical libraries with a glue language is aesthetically unappealing to me, and pragmatically, what’s the point of dividing development effort between two Python-based metaCASes? A well-defined global syntax is the biggest preventative measure I can think of.

So now I have some fun goals for the summer, to round out my list of not-so-fun goals: figure out a grammar for Mathematica’s syntax (this may be, literally, impossible: I have heard it said that the only grammar Mathematica follows is the ex post facto grammar imposed by its parser’s behavior); write a parser for that grammar, test it; track down the inevitable incompatibilities, attempt to improve the grammar; recurse. It’s going to be a bitch, but fun, hopefully!

Possibly relevant posts:

• Mathematica-style parser (6/20/2005)
• PyPI package basicCAS (7/24/2005)
• Math parser update (6/22/2005)