From: Philippe Verdy (verdy_p@wanadoo.fr)
Date: Sun Oct 14 2007 - 03:45:09 CDT
Hans Aberg wrote:
> So then it is not an extension of regular expressions, but a
> particular implementation, or some other, more specific computer
> language you are wroking on.
According to what is proposed in the latest UTC paper, that describes the 10
rewrite rules, there's nothing that what isdescribed there that I don't
support as well; it's just that I perform the rewrites in the reverse
direction, without changing the complete set of returned matches.
You may consider implementing the 10 rewrites the way it is described in the
recent UTC paper, but I'm sure that it is equivalent to performing the
reverse operations (using only 3 rewrite rules instead of 10, including one
that is common about cancelling double negations, which is trivial, so the
differences are in 2 rewrites against 8, and I suppress two operators using
only alternations and no intersections or differences).
What is implementation specific is this reversal, not the semantic; and also
the fact that I'm not limited to ONLY the longest matches (look into POSIX
regexps, there already exists operators to control this strategy for the
preferred match length, which is equivalent to saying that between several
possible matches, there's a priority and so an ordering is applied on top of
the result set).
This archive was generated by hypermail 2.1.5 : Sun Oct 14 2007 - 03:46:44 CDT