- From: Mark Baker <distobj@acm.org>
- Date: Wed, 24 Jul 2002 08:08:50 -0400
- To: Hao He <Hao.He@thomson.com.au>
- Cc: www-ws-arch@w3.org
On Wed, Jul 24, 2002 at 02:08:24PM +1000, Hao He wrote:
> hi, Mark,
>
> Yes, harvesting a specification does not give as much as we would like. The
> more interesting question about WSDL is perhaps where the overal role WSDL
> can play and should play in a WSA?
To start, I think we should document how people use it today. Then we
can examine the overall architecture per our requirements, and decide
exactly what role WSDL should play.
> Also, should WSDL include a description about state transitions for a
> RESTful WSA?
>
> Any thoughts?
What do you think? 8-) I'll leave the detail to later, but I personally
don't see much of role for large parts of WSDL (e.g. operations). WSDL
exists primarily to be an IDL-like thing, yet the big leap forward in
Web architecture versus previous distributed computing architectures,
was the use of a generic interface for all components. This means that
any component can communicate with any other because they share common
connector semantics ("actions"). With SOAP 1.1 + WSDL current practice,
connector semantics are shared only between components that have built
in knowledge of one another. Hence the much greater coordination costs
in getting them to communicate.
Yesterday I saw a presentation from an analyst from Zapthink who
suggested that by 2004/2005, Web services would be able to communicate
with each other without any compiled-in knowledge of particular types of
Web services. Wow, what a great idea, who would have thought of that?!
8-) Of course, the Web has been doing that since it was created.
But we don't need to get into that now.
MB
--
Mark Baker, CTO, Idokorro Mobile (formerly Planetfred)
Ottawa, Ontario, CANADA. distobj@acm.org
http://www.markbaker.ca http://www.idokorro.com
Received on Wednesday, 24 July 2002 07:56:41 UTC