I think with Antlr
or JavaCC we will have to rewrite some of their code for our purposes. Antlr is
public domain, while JavaCC is now distributed under a BSD style license. We
can make whatever modifications we need for our purposes. (Not that I'm
suggesting either product is OK as a base. What software to use is still an open
question.)
We will need to determine our backtracing or
lookahead requirements as this is the main cause of speed degredation in grammar
based parser generators. A fairly up-to-date list is at:
Fred
Campbell wrote:
>Personally, I would advise against
ANTLR.
