Meeting notes: Dan: In order to constrain the shoe size property, you constrain the derived property. Shoe size property with a related value property with a range of decimal (or whatever). To relate properties with string values with properties with concrete values. Ian: Does this require mirroring the work done in RDF Schema but specifically in DAML+OIL Yes. Ian and Peter: Have been working to make the syntax look less restricting. Constrain the semantics so that they are only applicable when you are using the rules ... Still a few wrinkles. There is a new version of the proposal on the website where Ian and Peter's original property is. Jim: Much clearer that we dropped concrete and abstract. Suggest that we find a better set of words that people will understand better. Proposed names ... Pre-Defined and User-Defined. General consensus. We can have our cake and eat it too ... larger language you can use anything you want, but in the smaller sub language, you get more restrictive semantics which are computationally viable. Mike: Does anyone have any major problems with the proposal now? Dan?: Does it propose to include all of the XML Schema syntax? Possibly a separate problem. One concern ... Mike: One option is to actually pull in the XML syntax. Not a good idea. Jim: To Tim and Dan: Can we say that we are using this syntax, but if RDF chooses one, we will go with that. Answer, generally, yes. We will put something to that extent in the documentation. Committing to go with XML Schema as data type (surface syntax). Mike: Others with concerns? Tim?: There is no way to represnt an integer as a binary ... integers are represented in decimal. There are some decisions that were made when it was decided to go with XML Schema. Worst problems ... dont know how to order dates ... can only use decimal ... not hex or binary ... You can't write 1/3 ... It is not at all clear that we could do any better even if we wanted to and put the time into it Jim: The nice thing is that we can do both use the pre defined stuff, and that others can make their own stuff. Dates for instance, some will use the pre defined stuff, while some will want to try to capture the real logic of dates. Mike: Can you propose the separation on the restriction rather than the property? Ian: Concrete restriction and abstract restriction? Have to think about that. A little bit worried about it ... that you can use it with an abstract restriction in one place, and concrete restriction in another. This means they are not disjoint. Is it still traceable wether they have mixed their concrete and abstract? Ian: used the same url for the new one as the old one. Still has the old one. Mike: Break in and talk about the meeting next week. Planned for next week. Plenary session. Break out session for the language. To get feedback from the larger community. Birds of Feather sessions. Benjamin Grossof(?) - rules Lyn: Problem with the W3C semantic web activities talk during the planary. Jim: Suggest 15 minutes talk about that in general, also, happy to hear about DAML and the W3C. If Ralph is there ... he can present the semantic web Activity. Tim will be there Thursday ... that might change. What about people who should go to multiple break outs? Announcing Chairs? Mike: Hoping to announce chairs .. not done yet. Telecon? For break outs as well on 15th. Jim: Finalized agenda today. Tuesday: Presentations plus lang presentation Wednesday: Break outs and poster presentation Thursday: Report back from break outs Friday: Government science fair ... demos and presentations Possibly good contacts for more funding. Jim: Lots of buzz from his circle. Good job! Back to semantic talk. Looking at CT24 document. Do you have one of these axioms for every integer. It can be said for every integer, but you don't have to have it for every one. DAML Engine ... anything that takes in DAML and spits it out. Value and data type are not so different. ?? Ian: says XML Schema deals with this and they wrote a document about it Dan: says it is / can / will be done in DAML ?? Cardinality example is a red herring. Dan: XSI:type in the Ian and Peter proposal. How many different places is it allowed in the DAML syntax. Mike: We all agree that we dont want to put the XML types in the middle of RDF. They would be in a new file and be referrenced by URI. NOt doing this would break RDF stuff. Can we take something from Dan's proposal and extract it to fix Ian and Peter's Proposal? Possibly. Whether the behavior of the black box should be left unspecified (done somewhere else), or specify it in the DAML. Suggest that DAML be able to do everything ... but that means we are reinventing lots of stuff. Mike: If you have company x who wants a type for part numbers and they understand temp is always in farenheit. Their site would have a xsd file with this information. Dont recreate this in DAML, just read it in. Use current RDF tools Tim: Want to be able to do either. Recreate it so that we dont have to recreate the RDF Tools How are you going to get around reasoning in RDF? Q to Dan: not proposing to use and RDF schema. How do you get the string 10.0 to the number 10? Dan: Recognize the reserved word ... use a built in dispatch table to find code to read string to convert it to integer. Tim: First, the thing that converts some string representation and the integer rep and returns an object. Second .... Principal of Erasure: you can delete any of the triples and the statement has to still be true. Not everyone has bought in to this completely yet. Can't resolve the discussion without opening it.