WS Architectural Loose Ends / Outstanding issues

The concluding section of the WSA document should probably [at least the
editors on the call today agreed!] be a listing and short description of the
architectural issues that a) are significant for the WS industry; b) we
cannot say anything at all definitive about; and c) cross WG /
organizational boundaries, so there's not an obvious group that can resolve
the issue on their own.
Not all outstanding problems are "architectural" -- for example, there
doesn't seem to be much disagreement on what a reliable messaging spec would
actually do from an architectural perspective, but for various
organizational/political/business reasons there's no obvious way to get
agreement on a standard here.  So, this isn't necessarily an "unsolved
problems in web services" list, it's an "unresolved ARCHITECTURAL questions
list."

My strawman list of these would include:

- Intermediaries: They're defined in SOAP but not WSDL, so there's no way
for a WSDL-driven tool to generate a sequence of intermediaries to implement
a service.

- WSDL properties/features: there's some discussion of not putting them in
WSDL 2.0, but the Choreography people plan to use them.

- RESTful SOAP features:  There's no support in the latest WSDL draft for
the SOAP 1.2 webmethod feature, and at least some fear that this will be
dropped from WSDL 2 for lack of time/interest. 

The idea here is to have a "please don't let these fall on the floor" list,
so we should be fairly inclusive.  That is, let's err on the side of
including an issue unless it is clearly accepted by somebody (at W3C, OASIS,
or WS-I, I suppose) as being a "must solve" question.  [If I have
misunderstood the status of WSDL 2.0, please forgive and correct me!]

So, are all of these really in this category?  What else is there?  Can
anyone propose (or point to) a clear, 1-paragraph or so description of the
issue and resolution options for any of these?

Received on Wednesday, 7 January 2004 21:17:35 UTC