Announcing Marpa 0.200000

Marpa is now at 0.200000. Following a standard rhetoric of version numbers, this indicates that it's an official release and a major step forward, but still alpha. Marpa is a general BNF parser generator -- it parses from any grammar that you can write in BNF. It's based on Earley's algorithm, but incorporates recent advances, so that it runs in linear time for all those grammars parseable by yacc or recursive descent.

The big news with Marpa 0.200000 is Marpa's 3rd generation evaluator. The previous version of Marpa had two evaluators -- one fast, but only good for producing a single parse result, the other capable of dealing with ambiguous grammars, but slower. The 3rd generation has a single evaluator which combines the best of both. Not the least advantage of this change is that it simplifies the documentation and the interface.

Marpa remains alpha and its interface may change. But I use Marpa in my own work -- it's the basis of some rather handly HTML utilities, if I do say so myself. But I would not at this point consider Marpa for anything mission-critical or "production".

An obsolete predecessor of Marpa, Parse::Marpa, is still on CPAN. I plan to remove Parse::Marpa from CPAN shortly. Anyone for whom that is a problem, please let me know. Remember, the heroes of old will feast forever in Valhalla, and Parse::Marpa will live eternally in backpan.

As for the future: The major barrier between Marpa and world conquest is speed. I am busily converting Marpa to XS.

Leave a comment

About Jeffrey Kegler

user-pic I blog about Marpa, my parsing algorithm, and other things of interest to techies.