W3C home > Mailing lists > Public > www-rdf-interest@w3.org > April 2002

RE: Documents, Cars, Hills, and Valleys

From: Joshua Allen <joshuaa@microsoft.com>
Date: Wed, 24 Apr 2002 08:43:18 -0700
Message-ID: <4F4182C71C1FDD4BA0937A7EB7B8B4C104F058F0@red-msg-08.redmond.corp.microsoft.com>
To: "Thomas B. Passin" <tpassin@comcast.net>, <www-rdf-interest@w3.org>
> My take on this is that we are so far from understanding how people to
all
> the disambiguating that they do, and how they can come to understand
the
> meanings of words from common usage and circular definitions, that it
will
> be a long time before we can get the computers to do it except for
simple
> cases - Cyc not withstanding.
> 
> So we better plan on making the most of simple cases.  Which is what
the
> original Web did in the first place.

Exactly.  Saying "we can just infer the meaning of an identifier from
the context by using AI techniques" is like saying "who needs URLs, we
can just link up all of our hypercard systems by inferring context of
the hyperlinks contained within".  Both are theoretically possible, but
then URIs have the "U" for a reason.  They are not intended to be local.
Received on Wednesday, 24 April 2002 12:10:02 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:53 GMT