W3C home > Mailing lists > Public > www-ws-desc@w3.org > February 2002

Re: WSDL and W3C

From: Paul Prescod <paul@prescod.net>
Date: Fri, 15 Feb 2002 14:16:48 -0800
Message-ID: <3C6D88D0.9907F3CE@prescod.net>
To: w3c wsdl wg <www-ws-desc@w3.org>
Rich Salz wrote:
> 
> The mailing list is open (www-ws-desc@w3.org) to anyone.  You could
> join, and then post your proposal.  It will get rejected, because the
> fundamental charter of the group is "fix wsdl"
> 
> Perhaps another way would be to write up a set of requirements that your
> idea meets, but that wsdl doesn't meet.  Send them to me or the WG list.
>   At least they should get into the list, and you'd be owed an
> explanation of why various requirements are (or aren't) being met.
> 
> I assume when you say "forces RPC model," I assume you mean it's not REST.

REST is a big jumble of ideas. The Web works at the scale and
sophistication that it does because it is organized as a series of
resources with links between them. Web Services standards need to
acknowledge that. It must be possible to describe services that are
composed of multiple URIs and where the URIs are dynamically generated
at runtime.

The fundamental problem of dynamic systems is how to represent shared
persistent state, the Web has one solution, dynamically generated URIs
like this (particularly useless one):

http://slashdot.org/comments.pl?sid=28054&threshold=-1&commentsort=0&tid=99&mode=thread&pid=0

COM has a different solution. CORBA has a different solution. But you
can't get away without defining a solution and still have a useful
system. Imagine if COM restricted you to always using singleton objects.
It would be useless.

 Paul Prescod
Received on Friday, 15 February 2002 17:19:38 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:18 GMT