🚀 go-pugleaf

In article <9104252257.AA18628@galahad.cs.purdue.edu> svb@CS.PURDUE.EDU (Stephan Bechtolsheim) writes:
> has nothing to do with emacs. Check the manual: you can
> define what a word is etc., so there is no problem with that.
> 
> StvB

Umm... Correct me if I'm wrong, but doesn't emacs allow you to
redefine the *constituents* of words through the use of the syntax
table, but not exactly what the *definition* of a word is, i.e. I
could say that "from now on 'a'is a punctuation-constituent
character", but I cannot say something like "a word consists of a
non-empty string of word-constituent characters surrounded by
whitespace or punctuation-constituent characters".

I've thought about how to do this, since emacs' idea of words drive me
crazy too.  I've considered altering the word functions in vip.el to
use regexp-forward-search to look for patterns like P+W+\|P+W+, where
W is a regexp representing all word constituent characters (like in
the syntax table), P punctuation chars and W white chars (e.g.  (setq
W [0-9a-zA-Z])). I haven't gotten it working yet (no time blah blah).
Does anybody out there think this is good idea? That this is a lousy
idea? That I am making something simple very complicated and should
RTFM? In any case, I remain...
--
_______________________________________________________________________
						|  Bill Reynolds
	  				 	|  bill@inls1.ucsd.edu