incremental parsing

[Thomas Schilling]

Hi Roberto!

You referenced a quite interesting document on berkley
about an incremental parsing algorithm. (You should then 
also have a pdf document called "Practical Algorithms for
Incremental Software Development Environments" by the
same author Tim A. Wagner - if you haven't I think i can
give you the url.)

It sounds really powerful but what do you think are it's
great advantages towards little reparses in memory?
For code completion we would still need a class store
(i think) and it would be really much effort to implement
while i still don't see that great advantages.
(We need a kind of versioning system for every file in
the editor, an incremental lexer, a parse tree for every
open document (and maybe even referenced documents)
and, of cause, also the incremental parser, which, if based
on e.g. bison, also wouldn't be that fast since it operates
on files on disk, AFAIK.

But maybe you have other opinions or see other purposes
for such an ISDE. If so, please tell me.

Ciao Thomas