- From: Sandro Hawke <sandro@w3.org>
- Date: Mon, 17 Nov 2003 21:13:30 -0500
- To: Stefan Decker <stefan@ISI.EDU>
- Cc: Jim Hendler <hendler@cs.umd.edu>, Benjamin Grosof <bgrosof@mit.edu>, adrianw@snet.net, www-rdf-rules@w3.org, phayes@ihmc.us
> Please apologize my ignorance - what is hard about doing closed world
> reasoning on a giving RDF graph?
It's not hard to do CW reasoning on a particular graph; what's hard is
to convey general knowledge (rules) involving CW assumptions. As
Benjamin said, formally the semantics of such statements are defined
with respect to some knowledge base (graph). Unfortuntely, we haven't
figured out how to handle that dependency in the open web environment.
(Well, cwm presents one solution, and there may be others. But
nothing's received much acclaim.) Some people don't see a problem
there, saying whoever/whatever is doing the reasoning is responsible
for only using CWA appropriately; others (including me) don't think
that scales well enough. I feel like I need to explain that better
(in response to an earlier message from Benjamin), but I'm travelling
now and haven't done it.
-- sandro
Received on Monday, 17 November 2003 21:12:57 UTC