Re: Interview with David Wetherell of CMGI

From: Robert S. Thau (rst@ai.mit.edu)
Date: Thu Jan 25 2001 - 15:27:11 PST


kragen@pobox.com writes:

> Lexis-Nexis's work is not useful for prior art invalidating
> AltaVista's patents, though, as (a) they weren't spidering anything
> and (b) they didn't publish.

FWIW, some relevant patents (most culled from a post on /., but not
the earliest) are:

 US5864863: Method for parsing, indexing and searching world-wide-web
            pages
 US6021409: Method for parsing, indexing and searching world-wide-web
            pages

 US5970497: Method for indexing duplicate records of information of a
            database
 US6138113: Method for identifying near duplicate pages in a
            hyperlinked database
 US6128690: System for remote memory allocation in a computer having a
            verification table contains information identifying remote
            computers which are authorized to allocate memory in said
            computer

The earliest of these was applied for on Aug. 9, 1996. This is the
first one listed, which covers distribution of index information among
machines in a server farm. The second listed, which looks like it
claims any inverted index system, was applied for in 1998. But don't
take my descriptions as gospel; I'm not fluent in Patent.

So, any publication describing a spidering, full-text indexer before
August, 1996 would be interesting. Before August, 1995 might be even
more interesting; IIRC, prior art from more than a year before the
filing date is significantly harder for the patent-holder to counter.

NB that the first patent *itself* references an article on Open Text's
spider from Sept. 18, 1995...

rst



This archive was generated by hypermail 2b29 : Fri Apr 27 2001 - 23:19:04 PDT