At 11:33 AM +0200 9/5/00, Quim Sanmarti wrote:
> > Yes and no. It could be, but that's really slow. Since you know
> > *everything* is going to match, why not just grab DocURLs() or
> > DocIDs() up front? It's much faster than having to hit the word
> > database at all.
> >
>I'm not sure I'm understanding that. The query evaluation system works with
>word references. We won't be able to query things as 'title:*' or '* not
>foo' if we haven't word matches to filter, just doc ids. Isn't that the
>idea?
You have a very good point, alas. My hope was that in the general
case of a simple "return all matches" you wouldn't need to hit the
word database at all. I think we'll need to consult with Loic about
good ways to walk a list of words rather than generating the whole
list on-the-fly. I certainly don't have the memory for creating a
record for every word in my database. :-)
-Geoff
------------------------------------
To unsubscribe from the htdig3-dev mailing list, send a message to
[EMAIL PROTECTED]
You will receive a message to confirm this.
