- From: Jim Hendler <hendler@cs.umd.edu>
- Date: Tue, 11 Feb 2003 13:43:30 -0500
- To: herman.ter.horst@philips.com
- Cc: Jeremy Carroll <jjc@hplb.hpl.hp.com>, www-webont-wg@w3.org
All - on phone w/Jeremy yesterday I realized a gap in communication - my use of "technical" and "editorial" were really inappropriate - these are all technical comments. What i'm really concerned with is issues that would significantly change the design and need to engender new test cases (like annotations), vs. those that could be fixed within the AS&S document itself without causing changes in the "design" -- the latter can continue to be worked on without needing significant amounts of WG time, the former, since they have non-local effects, need more. Apologies for my miscommunication - Herman is correct to take me to task below. as an example, take the proof of the correspondence theorem -- while that proof would be nice to have, it is really not crucial to the success or failure of our language - a proof sketch is enough and I wouldn't even object to deleting the section. If the proof were completed during LC, adding it would be exactly one of those changes I refer to above -- it doesn't effect other documents or change our design, just makes us more comfortable with our current design. If the theorem were disproved, it might be disquieting and cause redesign - I have trouble from your earlier comments telling if you think the theorem is wrong, or just that the proof needs completion - and if it is wrong, how it changes our design. Those of us who are not mathematicians can't alwasy tell from the technical discussion - all I'm really asking is for people to be clearrer about what needs change, what needs minor fixes, and what needs editing. btw, these comments are addressed to all reviewers and "issuette" raisers, not at Herman or Jeremy -- we're trying to reach closure, and folks like me who don't understand polysyllabic utterances need it in plain talk to know where we stand... -JH p.s. If proofs had to be completed before life could go on, none of us would be working on any of these devices because we would have no way to no for sure that P and NP are different, and therefore we'd assume all our algorithms are useless... proofs are wonderful when you have them, but not always necessary for progress... At 16:00 +0100 2/11/03, herman.ter.horst@philips.com wrote: > > >It is unclear how many of Herman's review comments on AS&S have been >> >addressed. >> >> Herman - same comment as to the above - most of your comments are >> editorial, and are just waiting for Peter to get around to them. >> Please pull out those you think effect the design of the language > >See also my preceding mail I just sent [1]. > >There are no comments with effect on the design of the language. >The main objection deals with the RDF characterization of >OWL DL and OWL Lite, for which Jeremy is now making an >alternative proposal. > >However, many other comments aim to help to make the mathematical >details of the semantics correct, and are not simply editorial comments. > >That WebOnt needs to do this kind of work can be illustrated with a >comment >from Pat listed in the log of the Bristol face-to-face meeting >of WebOnt: >> 09:17:32 [DanC] >PatH: yes, the correspondence between entailment on abstract syntax >and entailment on the triples is to be carried forward, and to >be checked carefully. > >If the design of the language would be all that matters, then WebOnt >could have gone to Last Call before dealing with mathematical >semantics at all. > [...] -- Professor James Hendler hendler@cs.umd.edu Director, Semantic Web and Agent Technologies 301-405-2696 Maryland Information and Network Dynamics Lab. 301-405-6707 (Fax) Univ of Maryland, College Park, MD 20742 240-731-3822 (Cell) http://www.cs.umd.edu/users/hendler
Received on Tuesday, 11 February 2003 13:47:19 UTC