SWSA Telecon Minutes of Nov 18, 2003 Michal Zaremba, Mike Huhns, Mark Burstein, Carole Goble Tim Finin and Massimo Paolucci joined mid-meeting. We began by discussing the B2B (RosettaNet) Use Case developed by Michal Zaremba. Mark: need to extend the "Scenario/Steps" sections with specific hypothetical examples that walk thru the particulars of (e.g.) a PO purchase example. Use of the block diagram that was pointed to by URL (http://www.rosettanet.org/PIP3A4) would be valuable. Extending that discussion to show the use of the related protocols mentioned in bubbles at the top of that figure would also help. The key point is to talk about the value added of semantics in those steps - for example, how would a broadly shared ontology of products/product components and their specifications help in "defining" each item in a product catalog (allow graceful extension to new items without re-issuing the catalog, for example). The example might include a query to the vendor for a product with some particular mimimum specs, which could then be included in a PO. An example of how the process model might change would be useful. (They talk about use of different variants of the PO process by different agents. If the fields only changed slightly, perhaps these differences could be addressed simply by annotation of the PO with some ontology version information (a URI) which pointed to a mapping of its parts onto the prior version, and to semantic definitions of the new field(s). Finally, for the status tracking/monitoring example, how might the status messages work if they related back to a published "external" process model that also specified how compensation would work if the order was canceled once the order reached a particular state? A worked example like that would be extremely valuable in revealing the requirements for an external published semantics of the process. The difficult issues here are things like - expected time to complete a stage of processing, vendor vs client cancelation or delay. ----------- Discussion of message from Frank McCabe Mark: My take on Frank's concerns is that we should be wary of developing too detailed an infrastructure because we cannot hope to keep up with industrial developments. I think our role is more to lay out the space and identify key core requirements to support semantic web service functionality than to develop and implement the infrastructure. Carole: I think he is saying we should make clearer the distinction between the things that are technology independent and the policies that make it work - difference between particular standards and requirements on standards. For example, the Rosettanet use case could be implemented in any suitable infrastructure. Mark: For example the RosettaNet case could use an HTTP transport layer and still have the same requirements for semantics. Carole: e.g. negotiation and contracting standard - does it match a use case - does it live up to the set of requirements we identify. Otherwise we will be very technology specific and will become outdated CG: Example: GGF made syntactic changes to WSDL to expose state - went to W3C and they said No - you will break all the tooling. So now they are trying to be more Web friendly. Next week: discuss Requirements document outline. Review set of use cases.