Minutes of SWSA Telecon of December 2, 2003 Michal Zaremba, Massimo Paolucci, Mike Dean, Mike Huhns, Mark Burstein, Tim Finin, Stuart Williams IRC Chat established on channel #swsa at irc.daml.org Tue Dec 2 11:00:59 2003: :mdean!-mdean@216.229.181.19 JOIN :#swsa Tue Dec 2 11:07:27 2003: :MarkB!-burstein@128.89.73.116 PRIVMSG #swsa :Hi Mike Tue Dec 2 11:17:23 2003: :Stuart!-skw@hplb.hpl.hp.com PRIVMSG #swsa :Hello... I've now made it onto the channel Tue Dec 2 11:19:29 2003: :mdean!-mdean@216.229.181.19 PRIVMSG #swsa :Tim Finin joined Tue Dec 2 11:21:51 2003: :massimo!+massimo@PERSEPHONE.CIMDS.RI.CMU.EDU JOIN :#swsa Tue Dec 2 11:22:21 2003: :Mike!-huhns@129.252.11.209 JOIN :#SWSA Tue Dec 2 11:28:13 2003: :timfinin!-finin@aibo.cs.umbc.edu JOIN :#swsa MB: In trying to develop an outline for the Reqs doc, I reviewed the WSA document at http://www.w3.org/TR/2003/WD-ws-arch-20030808/ They talk about the fact that XML is more fundamental than WSDL - that XML is the 'backplane' of WSA in that it promotes interoperability of protocol description and content communication. (http://www.w3.org/TR/2003/WD-ws-arch-20030808/#id2617100) In trying to define how we 'live on top' of WSA, I wonder whether we need to say specifically what we mean by semantics and its role in the various functions we claim will be made better by using SW. What architectural pieces do we need to add to the WSA picture? SW: There has been some traffic recently about discovery wrt UDDI - how to add semantics. We can claim that the ability to do subsumption inferences to support discovery is part of SWS not WS. We are interested in richer descriptions of interactions - use of extra levels of indirection between specific message content and purposes. None give an account of what is accomplished by the interface MassimoP: We want to build on top of the WSA, using XML. Last year IBM said DAML-S not BPEL, not use WSDL... so dismissed it. So saying explicitly that we build on top of WSA helps to fend off these critiques. MarkB: Yes, and also that we are building on Semantic Web, using XML-based semantic representations. And so we will use logical representations that are consistent with the semantic web. MB: Need a definitive Semantic Web definition to point to. SKW: http://www.w3.org/2001/sw/Activity says: " The Web can reach its full potential only if it becomes a place where data can be shared and processed by automated tools as well as by people. For the Web to scale, tomorrow's programs must be able to share and process data even when these programs have been designed totally independently. The Semantic Web is an initiative of the World Wide Web Consortium (W3C) designed to provide a leadership role in defining this Web. It develops open specific technologies that are ready for large scale deployment, and identifies, through open source advanced development, the infrastructure components that will be necessary to scale in the Web in the future. " Stuart: The only difference I can think of is that its description is richer - you can turn a non-semantic service into a semantic one by enriching its description. Two aspects - discovery and matchmaking, access (grounding) and interaction WS Policy is a way of making statements with conjunctive, disjunctive elements it is yet another description framework in the WS space not the SW space (RDF) has languages for policies (e.g. security policies) MD: did they explicitly decide not to use the SW? SW: I think they decided not to use RDF MP: ws policy is at xml.coverpages.org/ws-policyV11.pdf MD: So they used their own XML-based Lang? SW: Yep MP: it is ok if the soap msg has elements with groundings to owl, as long as there is a mappng from some ontology someplace to the information that goes on the wire then there is a bridge to known semantic web services SW:For WS-Policy and some commentary see http://xml.coverpages.org/ni2003-06-04-a.html MB: So conclusion of last bit of discussion is that we should be looking for requirements of the form: To support function x (one of a-m) semantic descriptions are requierd of the following bits Next week: Mark, Chris are in Rome at WISE. Stuart will lead discussion for whoever can make it. Mike Dean will initiate.