Re: Multiple endpoints with the same interface

"David Orchard" <dorchard@bea.com> writes:
> > Are you suggesting that whether two wsdl:endpoints (nee
> > wsdl:ports) are semantically equivalent is orthogonal to
> > whether they are defined within the same wsdl:service?
> >
>
> Maybe I'm confused, but I thought that is the case.

+1.

> > The general issue was discussed at one of the face-to-face
> > meetings last year, perhaps as early as the Paris meeting. As
> > I recall, given two workarounds, the sentiment of the WG was
> > that the status quo was sufficient.
>
> hmm, any particular reason that you (or somebody else) would care to
> articulate?  Was there a lack of a compelling use case?  Or it became too
> complex?  I took a gander at the minutes of the paris f2f [1].  I saw the
> discussion that you mentioned about the ambiguity of whether same
interface
> and differing bindings is 1 service or 2, which I think is right on the
> mark.  I further followed into the Alexandria F2F and saw the service-type
> proposal - I still like having the address as a firstclass concept - and
saw
> the vote to accept service types.  But then I lost the trail :-(

We accepted serviceTypes, put it into the spec and then went down the
inheritance path. Having inheritance and serviceTypes both didn't
sort of make sense, so we dropped serviceTypes.

Have you looked carefully at how we support overloading? You should;
its rather amusing IMHO ;-).

Sanjiva.

Received on Thursday, 24 April 2003 06:05:35 UTC