Lists Home |
Date Index |
2/15/2002 12:02:44 PM, "Bullard, Claude L (Len)"
>The Jan 2001 MSDN provides a set of baseline specifications
>and the specs proposed for the Global Web Services. RDF is
>never mentioned. RDDL is never mentioned. Except for WSDL
>and UDDI, maligned on this list, none of the favorites of
>XML-Dev are mentioned. That tells me that beyond WSDL
>there is a very deep division of labor and perspective
>about what specifications are important to web services.
I'd say that XML-DEV collectively has a pretty good track record
in pointing out things that eventually succeeed (SAX being the
shining example) and giving early warning of things that are going
to be interoperability nightmares (Namespaces and W3C schema come
to mind). Even Paul Prescod just asserts that the REST approach
will probably underly the *next* generation of Web Services as the
limitations of RPC over HTTP become apparent.
>For the implementer, for the business and marketing
>manager, this pushed WSIO to the front of the organizations
>to pay attention to when adopting policy with regards to
>REST may be great guys. I'm all for it. But the specs
>the WSIO has before them don't mention it.
Sure. We as developers working for companies that must actually
sell something have to play the hand that the market has dealt us.
WS-I will probably play a valuable role in making sure that
everybody knows the rules of the game. That gives the current web
services paradigm the best chance it will have to solve some set
of real problems, even though that is probably a small subset of
the problems it is being touted to solve.
>The other side of that coin is that the
>WSIO has to produce, and given that this is a direct
>engagement between extreme competitors (read Ellison's
>comments of late), that is going to be a very hard production.
>If the standards bodies really want interop, they better
>get ready to move fast in the rapids. Anyone who shoots
>the rapids can tell you it requires intense focus.
I think Tim and Tim have it right -- the W3C has no business being
in the game of rubber stamping WSDL or trying to find the profile
of specs from 5 different organizations that actually can be made
to work together in the short run. This is not its core competence
any more (although arguably that's more or less what it did with
HTML 3.x, XML 1.0, and DOM 1.0). The *members* of the W3C need
exactly that, so they formed the WS-I.
Maybe I'm getting schizoid, but I have no problem believing that
the WS-I has an important role to play today in keeping the
SOAP/WSDL/UDDI/WS*/J2EE/.NET house of cards from collapsing
(recognizing that one does that with cardboard and glue).
Nevertheless, the W3C is "right" that a Web Services architecture
that exploits REST principles and is unencumbered by patent and
builds on a solid understanding of metadata will work better than
the mess that the WS-I is trying to sort out, although it may take
a few years for this to be obvious.
Or maybe the WS-I is sorting out the basic interoperability issues
so that J2EE and .NET can interoperate within firewalls and
between established business parters; the W3C is (or should be,
and XML-DEV should hold their feet to the fire) sorting out how
this can scale to the Internet, recognizing that some refactoring
of the WS-I output will be necessary to achieve this.