- From: Bijan Parsia <bparsia@cs.man.ac.uk>
- Date: Fri, 23 Mar 2007 13:58:15 +0000
- To: Alan Ruttenberg <alanruttenberg@gmail.com>
- Cc: "Eric Neumann" <eneumann@teranode.com>, public-semweb-lifesci@w3.org
On Mar 22, 2007, at 8:53 PM, Alan Ruttenberg wrote: [snip] >> VK - Hasn't seen a use case for puning. Data type reasoning >> enhancements sound valuable. Rule chain inclusions would be very >> valuable. Qualified cardinality constraints is another >> enhancement, but >> can be done with OWL as is. Actually, not in any sensible way. Obviously, since SHOIN and SHOIQ (for example) have the same worst case complexity, there *is* polynominal interreduction, but it's definitely not *usable* (the current trick with subproperties is both not correct and considerably clutters up modeling). Consider that you can encode Many OWL DL ontologies in OWL Lite (e.g., including arbitrary disjunction and negation; jeremy carroll wrote a little tool to do that encoding). The results are unworkable for people (you definitely would never want to edit them) and reasoners (we experimented....horrible). [snip] Cheers, Bijan.
Received on Friday, 23 March 2007 13:59:02 UTC