- From: Bijan Parsia <bparsia@isr.umd.edu>
- Date: Wed, 15 Mar 2006 17:20:30 -0500
- To: "Battle, Steven" <steve.battle@hp.com>
- Cc: "Carine Bournez" <carine@w3.org>, <public-sws-ig@w3.org>
On Mar 15, 2006, at 12:08 PM, Battle, Steven wrote: > Carine wrote >> Clarification about semantic annotation for wsdl: >> The *meaning* is actually in a separate document (or several >> ones). The annotation in the WSDL is supposed to *point* to >> *external information*. >> > > Has this been decided already? The charter says, "The Semantic > Annotations for WSDL Working Group is chartered to define one or more > properties of WSDL 2.0 components to point to additional semantics to > concepts represented by those components, e.g. interface, operation, > endpoint." > > Nothing in the charter restricts this additional semantics being > embedded in the WSDL document itself. Indeed if it is, the likelihood > is > that tools will simply ignore the additional semantic components. +1 And I'm generally IN FAVOR of an annotative approach (as with OWL-S). But of course, I can always import the wsdl into *another* wsdl, and add annotations there. But then why forbid doing it in one and the same document? Why would we decide this in advance? (There's lots of things that one might want to have external. An ontology that one is using. A schema. Another wsdl. But so? WSDL now can be multidocument. WSDL-S requires, afaik, putting pointers *in* the wsdl pointing *out*. OWL-S only requires pointers going *into* the WSDL. These are differences, but minor and more human factor than anything else. There are advantages and disadvantages of each. The fear of whatever supposed problem there is with the MEANING of a SERVICE, however, is not an issue :)) Cheers, Bijan.
Received on Wednesday, 15 March 2006 22:20:56 UTC