W3C home > Mailing lists > Public > www-rdf-interest@w3.org > December 1999

A Killer-App ?

From: Seth Russell <seth@robustai.net>
Date: Wed, 22 Dec 1999 12:21:24 -0800
Message-ID: <386132C4.9384B071@robustai.net>
To: RDF SIG <www-rdf-interest@w3.org>
Tim Berners-Lee wrote:

> Comments on Harle & Fensel "Practical KR for the Web",
> http://www.cs.vu.nl/~frankh/postscript/IJCAI99-III.html

A passage in the paper above has excited my imagination:   "We provide a survey
and analysis of traditional, new, and arising Web standards and show how they
can be used to represent machine-processable semantics of Web sources having in
mind that this area may become one of the killer applications of AI."   So now
would some discussions of what a Killer-App of the Semantic Web look like be
appropriate on this list?

I think it is clear that the Killer-App will address the problem (quoting from
the paper above) of "finding information on the web and extracting it" .
Another thing is clear:  The WWW was not created by RFCXXXX but rather by the
NSF server and LYNX followed by MOSAIC.  So when the the question is asked:  How
can we expect RDF to gain a "wide acceptance on the Web"; we should be looking
for end user tools (servers and browsers) that will make web semantics work.
Another thing is clear:  the Killer-App will not be just some tool that allows
us to read and write RDF as RDF - this stuff is not an end in itself.  Rather it
will be some tool that allows us to retrieve and/or write write knowledge in a
language we already know and love - natural language.  It will help us publish
knowledge like throwing a needles in a haystack so someone else (even ourselves)
can find it later.

So what would a Killer-App with those (clear?) constraints look like?

Mind if I take a shot at it ...

For the reader:
*) It will be a browser plug in
*) Humans will communicate to it in natural language
*) It will allow the Semantic Web to be navigated in human understandable
terms.  You will not need to be a knowledge engineer to us it.
*) It will have a large local storage of knowledge that represents the user's
subjective view of his/her world.  This knowledge would be used to understand
(parse) the users requests and to organize the knowledge retrieved from the
semantic web in terms the user can navigate, and query.
*) It will employ user selectable filters (annotation services) of the masses of
information that are (will be) available.

For the writer (publisher):
*) Basically the same tool as for the reader with the added ability to:
*) Translate natural language into RDF and publish it on the Semantic Web.  This
could be for meta annotation of resources as well as for publishing of
knowledge in general.
*) It will employ tools and protocols that naturally work against the "Tower of
Babel Problem".
*) It will permit ~any~ publishers to ~earn~ a reputation as an annotation /
authentication service.

Well I can dream,  can't I?

Seth Russell
Received on Wednesday, 22 December 1999 14:48:28 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:07:28 UTC