language parallax

I was hoping to have as much synergies as possible with IBM’s SAFARI project. Part of that was to use the LPG (LALR Parser Generator — formerly the Jikes Parser Generator).

But I think I’m going to go with ANTLR for a few reasons:

1) It’s really close to its 3.0 release

2) John Mitchell (an old Stanford prof if mine from years ago) continues to be actively involved, so I have to think it has some credibility in academic circles.

3) It has a very active community who seem to contribute grammars all the time. The ones I need for my project are Java 1.5, XML, and SQL. It’s had all those available for a while (as well as new ones for version 3.0).

4) Terence Parr has a new book coming out specifically about creating domain-specific languages

5) It seems to come up a lot in the other java development I’ve done. For instance, there have been antlr plugins for Maven for a while.

6) I’m probably going to build my own ajax-powered web-based IDE, so I’m not sure how much I’ll be able to leverage from the SAFARI project anyways. My interface will probably be pretty bad for a while, but that’s not a high priority at the moment. In the meantime I can focus on the high-level architecture, algorithms, and papers. If the project has any momentum, I’ll make a better UI next year.

In fact the highest priority at the moment is creating a good demo. I don’t have to implement a complete solution — just enough to give people the idea. Not having to work with large frameworks will allow me the freedom I need at the moment.