> -----Mensaje original-----
> De: Gilles Detillieux [mailto:[EMAIL PROTECTED]]
> > Hmm. htdig doesn't index like that right now. Wouldn't it
> be easier to
> > *remove* stop-words from the query?
> > This is relatively easy to do with the new parser (I
> haven't checked the old
> > one :), and would give similar results.
>
> Maybe I'm confused here, but isn't this the bug that was fixed back
> in May by a simple patch to the parser (a missing return statement or
> something like that)? ...
Nope, IIRC that was a bug that would force a coredump with every query
containing *two* successive ignored (stop)words.
However, I checked the old parser code in 3.2.x and it works apparently OK
with stopwords in phrases, i.e. it ignores them.
So if your query is e.g. "word1 <a stopword> word2" it will find results
with either "word1 <any stopword> word2" or "word1 word2".
Isn't this good enough? Recall is 100%, precision not so much but...
--
Quim
_______________________________________________
htdig-general mailing list <[EMAIL PROTECTED]>
To unsubscribe, send a message to <[EMAIL PROTECTED]> with a
subject of unsubscribe
FAQ: http://htdig.sourceforge.net/FAQ.html
