[antlr-interest] Any plans of next ANTLR Release

[Terence Parr]

On Apr 30, 2010, at 11:21 PM, Ron Burk wrote:
> Of course, Thompson's algorithm (and Cox's analysis)
> is oriented towards building lots of NFA's on the fly
> for short-term use. Lexical analyzer generators have
> the luxury of enormous CPU time to spit out optimized
> (if not optimal -- though what one might be optimizing
> for can vary) DFAs.

I plan to just try out the NFA simulation (with some static analysis hints encoded into SPLIT instructions) to see how fast they are.  I must balance speed with simplicity etc.. in generated recognizers.  It's fun to play!

Ter