RE: The UR Trout: Web Services, REST, SOAP

Jeff,
Please see my comments below.

Ugo

> -----Original Message-----
> From: Jeff Mischkinsky [mailto:jeff.mischkinsky@oracle.com]
> Sent: Thursday, July 10, 2003 12:58 PM
> To: David Orchard; 'Martin Chapman'; 'Sanjiva Weerawarana';
> www-ws-arch@w3.org
> Subject: RE: The UR Trout: Web Services, REST, SOAP
> 
> 
> 
> Personally I think the defining notion is WSDL, along with a mandatory 
> binding to SOAP/HTTP. The mandatory binding is to guarantee interop which 
> is what I thought is the raison d'etre of this whole effort.
> 

+1

> If the WSDL wonks decide that WSDL 1.2 can describe 'everything' that can 
> be done using a computer on a network, then I guess we'll have to burn that 
> bridge when we come to it.
> 

Well, so far WSDL 1.2 defines some specific bindings, but does not say that other bindings are excluded. So I would conclude that in principle WSDL 1.2 can bind to 'everything'. The issue, I think, is whether we want to say that those bindings are Web services if no alternative SOAP/HTTP binding is also provided.

> Remember, a definition to be useful, has to exclude (non-theoretical and 
> useful) things. Otherwise you've just got yet another definition for 'everything'.
> 

+1

> Since most of the web service advocates don't seem to think that web 
> services include CORBA, then certainly one litmus test for any definition 
> is whether it includes CORBA. (I'll even be nice and won't ask what happens 
> if I define an XML schema for GIOP messages, encode them as XML, and send 
> them along?)
> 

I don't fully understand your position here. Following your first statement, I would say that a service described by WSDL and offering two end points, one bound to SOAP/HTTP and another bound to CORBA, would qualify as a Web service.

> cheers,
>    jeff
>

Received on Thursday, 10 July 2003 17:21:44 UTC