>>> "Paul" == Paul Eggert <[EMAIL PROTECTED]> writes:

 > Akim Demaille <[EMAIL PROTECTED]> writes:

 >> I still think dropping the scanner hacks and moving towards using
 >> the GLR parser is a better road.

 > Yes, that could be true.  But I suspect it's a bigger project, and in
 > particular it will require considerably more testing.

But more freedom in the end!

 >> But then, why -> instead of the more traditional `::='.  Both are OK.

 > Aack!  Too many colons!

Well, they are bound together, they won't make anything more complex :)

 > Besides, I thought that right-arrow was more traditional than "::=".
 > Didn't Chomsky use a right-arrow in 1956?  (::::-)

It depends what community you consider.  -> is definitely in the
spirit (err, letter) of Chomsky, but Backus introduced ::= to define
Algol.  Naur kept it, and therefore `::=' is really _the_ traditional
BNF symbol.  Not that it matters (nor that departing from -> was a
nice move).



Reply via email to