- From: Jacek Kopecky <jacek.kopecky@deri.org>
- Date: Mon, 02 Oct 2006 17:35:29 +0200
- To: SAWSDL public list <public-ws-semann@w3.org>
Hi all, as we just entered Last Call, we should start thinking about the next stage, i.e. Candidate Recommendation (CR), where we try to get proof of implementability and feedback from implementors. As part of entering CR, we will need to set ourselves implementation criteria. The usual criteria is "two interoperable implementations of all the specification features", but applying this to SAWSDL may be non-trivial. For instance in WSDL they have parsing and authoring APIs (testing that they interpret the same WSDL document as the same set of components), and they can also test that messages generated according to a WSDL description are interoperable. We are adding annotation hooks for an open set of non-normative use cases. We can easily test parsing and authoring APIs, but can we also realistically expect two implementations that will behave interoperably over the same SAWSDL file, if we don't actually specify what they should do? There are already tools that have touched SAWSDL, see [1] and [2], and we might want to try them out and see what they do. [1] http://lists.w3.org/Archives/Public/public-ws-semann/2006Jul/0008 [2] http://lists.w3.org/Archives/Member/member-ws-semann/2006Sep/0019 We will discuss this general issue in our upcoming telcons, but of course any email discussion, including implementation plans (what do we want to do with SAWSDL?) will be very welcome. 8-) Best regards, Jacek
Received on Monday, 2 October 2006 15:35:28 UTC