September 04, 2001 Attendees: Mike Peter Frank Harold Richard Pat Ian Deb Stephan Sandro Hawke Jim Jeff Lynn Dan C. Announcements: None... Agenda: minutes from August 28 - approved W3C submission status (Peter) submitted to internal process and no one screamed. tomorrow will check if the process is complete, if so will submit to W3C Mike: suggest to submit it as multiple documents with the cover letter. Sounds like we have a month until the W3C accepts it to make editing changes. Possible for them to ask us to make changes, but that's unlikely. Comments about merging author and editor lists and make them alphabetically. Dan says "if I come up with something on the copyright stuff you are ok with putting it in the cover letter?" ... answer was yes Feedback from RDF Core about (Pat) most of the group are all for changing rdf:subclass to conform to the DAML suggestion. a small group of diehards are still opposed and are digging their heals in. Think it will go through. strawman DAML+OIL Query Proposal (Richard) task: to design a query language for DAML+OIL. from the proposal... 1. the amount of expressiveness! 2. the query being stated and the result being returned in DAML+OIL. RQL is relevant. an sql style language. a subset of the proposal. Includes inequalities on numbers and strings, non-monatonic operators (give me all of the known subclasses of a given class...), and allows disjunctions of rdf statements and negations of rdf statemnts. Question about non-monatonocity ... you say its non-mon if the answer is an inferrence ... you could also say that answer is not an inferrence at all but just a statement (??) Mike: interesting that the need in querries for having disjunction (find all californians or texans) points out one difference between querries and rule languages. Dan C. says no to Logically complete, guaranteed tractability. If you asked for subclasses you would expect to get implied as well as stated? Yes. Perhaps the answer that comes back needs to include one of three ideas. This is all there is (and I have proof), or here is all I can find and don't ask me for more (though there may be more ...), or here's one and I have more if you want them. Would suggest the option of here's all the answers and thats all of them should be in the semantics of the query language ... important to be sure that we recognize the database community, and that we don't reinvent lots of wheels that have been invented by this community. This is a good way to get database people back on the DAML wagon Richard: willing to take on either writing the strawman in SQL or describing it that way ... Paper from Ian that is relevant ... http://www.cs.man.ac.uk/~horrocks/Publications/download/2000/AAAI-2000.ps.gz possible updates to daml+oil-ex.daml Mike: has gotten half a dozen responses on the example ... shall we go ahead and look into making changes / updates on the example? Deb: perhaps add more text about why we didn't define the properties ahead of time? Perhaps an email of suggestions to everyone? Perhaps more text in the walkthrough rather than more examples in the ex.