Re: Resolution of issue 168

Hi Hugo,

On Thu, Jul 22, 2004 at 08:03:40AM +0200, Hugo Haas wrote:
> Hi Mark.
> 
> * Mark Baker <distobj@acm.org> [2004-07-20 12:37-0400]
> [..]
> > As an aside, I just wanted to note ...
> > 
> > On Tue, Jul 20, 2004 at 09:31:15AM -0400, David Booth wrote:
> > > 1. The parties that wish to interact using a Web service first agree on the 
> > > WSD and application semantics that will govern the interaction between 
> > > their agents.  How this is accomplished is outside the scope of WSDL 2.0.
> > 
> > that this step seems counter to the charter of the group, in that it
> > requires coordination prior to interaction.
> 
> The way I see it, the requester entity processing a WSDL 2.0 document
> and deciding to use it to start interacting is an agreement with
> the provider entity whose service is described. It is a scenario which
> didn't require coordination, just the requester entity entity making
> the decision to use the information described by the WSDL 2.0
> document.

But that information isn't everything that the requester needs to know
to start interacting; they also need to understand the semantics of the
operations.  Unless the architectural style itself constrains those
semantics - which SOA doesn't, but REST and other Internet scale
styles do - the style is unsuitable for ad-hoc integration due to the
burden it places on agents to establish agreement about operational
semantics prior to interaction.

But I think you knew that was my position. 8-)  Anyhow, I'm not trying
to "pick a fight", since I've previously failed to convince people of
this point on more than one occasion, and don't have any knew arguments
to offer.  Time, as they say, is the best teacher.

Cheers,

Mark.

Received on Thursday, 22 July 2004 13:39:20 UTC