- From: Jim Hendler <jhendler@darpa.mil>
- Date: Fri, 13 Oct 2000 18:06:28 -0400
- To: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>, heflin@cs.umd.edu
- Cc: phayes@ai.uwf.edu, www-rdf-logic@w3.org
I think a lot of the problem is that OIL does things in the language (as a DL system) that SHOE does with inference rules. For example, give simple rules defined classes aren't needed you simply say Definedclass(x,y):- Foo(x), bar(y), etc(x,y) as needed. however, rules on the web must be thought about carefully, not just because of complexity issues, but also the many problems that distribution, URI grounding, and other things bring up. SHOE also uses rules for the "other" kinds of equality - that is, if I want to say everything that is an X is also a Y, I can use a rule X(p) :- Y(p) but if I want to say what ontology O1 calls an X what ontology O2 calls a Y, then I may prefer equivalentto statements (See the paper Jeff referred to for more on how we use that). We believe that DAML-ONT should be completely rule-free for now. As we stabilize the language, we expect to release DAML-Rules (maybe not be that name) to extend the definitions to allow some kinds of rules (probably limited to something like Horn or less, for now). After that we work more and more towards extending, layering, and/or replacing (as the community decides) DAML-ONT w/DAML-LOGIC. that is the plan, and the processes by which this happen (still being worked out) will include a lot of open activities centered on this list and other places, so all will have a place to play. Please note that as much as possible, all seem to agree that a "layered" approach seems to make sense -- building languages like OIL on top of the nearest subset that works, building more complex markup systems that go beyond OIL or SHOE out of DAML-Logic, etc. Interestingly enough Shoe w/o rules is below OIL in expressivity, Shoe w/rules seems to be greater, KIF and ontolingua like languges are much more yet. Goal is to have all these things coexist and interoperate to the level possible -- my reigning principal in this (and I'm just one player) is that on the web there will never be a one size fits all solution, and our goal is to build as much and as exciting as possible so that the role of AI and logic on the web becomes clearer and anchored in the very languages and systems that run the web... -JH Prof. James Hendler Program Manager DARPA/ISO 703-696-2238 (phone) 3701 N. Fairfax Dr. 703-696-2201 (Fax) Arlington, VA 22203 jhendler@darpa.mil
Received on Friday, 13 October 2000 18:07:03 UTC