greywolf@starwolf.starwolf.com (Segmentation Violation.  Core dumped.) writes:
> If my analytical logic is correct, functionally it's nothing more than:
> 
> ((cd /usr/share/dict; cat words web2); cat stop.list) > /tmp/scheck$$
>  fgrep -v -f /tmp/scheck$$ $file_list
> 
> Of course, I think fgrep is not up to the challenge, and that's only
> a functional outline, anywho.  Am I off too far?

Yes.  Spell was a lot smarter than that.  In particular, it handled
common suffixes, etc.  I seem to recall that there's a paper about it
that will tell you just how wrong you are, if you can find it.  8-)

You'll note that there are few plural forms in web2.  (I don't say "no
plural forms," because it's a damned big file and i'm not going to
read it.  But none of the words i checked were in there.  The ones I
checked were all of the plural nouns in this reply, plus a few
more...)  Your replacement ... won't work so well in pratice.

Also, note that /usr/share/dict/words is _NOT_ the traditional words
list, it's just a link to web2a.


cgd