Re: Real-time distributed Web search (Gnutella knockoff?)

Date view Thread view Subject view Author view

From: Eugene Leitl (eugene.leitl@lrz.uni-muenchen.de)
Date: Fri Jun 30 2000 - 15:35:06 PDT


Kragen Sitaker writes:

> So if you direct an entire query to a single web server, it won't know
> about more than a small fraction of the pages on the Web. So it will
> have a small probability of giving you any particular page of the ones
> you want.

Not if the query is redirected to the web server storing the page in
question (plus the direct net neighbours).

Since the local server knows how to translate queries into web
documents, there should be no reason why database entries should not
be also indexed, a major problem today.

Somebody please integrate webglimpse or htdig into Apache. With enough
time, every Apache out there will have a full-text search engine
integrated. Add a clean document submission mechanism (crontabbing a
total document tree reindex every night is not very efficient),
voila. The parallel query distribution part will be not cheap, though.


Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Fri Jun 30 2000 - 16:40:04 PDT