RE: D-AG0005 - Simplicity Requirement

> -----Original Message-----
> From: Dilber, Ayse, ALASO [mailto:adilber@att.com]
> Sent: Monday, March 11, 2002 10:18 AM
> To: Champion, Mike; www-ws-arch@w3.org
> Subject: RE: D-AG0005 - Simplicity Requirement
> 
> 
> Mike's point is well taken.  I agree that it is difficult to 
> measure interoperability before the specs are released but I 
> guess one way of ensuring interoperability in the specs would 
> be that they are written in a way that there is not much room 
> left for different interpretations when implementing them.  
> The current wording of D-AG0005 requirement: "provides 
> simplicity and ease-of-use that does not impose high barriers 
> to entry for users of web services".  To me, one of the 
> implications of "simplicity and ease-of-use" is 
> "interoperability", i.e. no significant difference in web 
> services components, minimal architectural and technology 
> gaps, etc.  Any thoughts?  

I think there is consensus on that goal.  We're now trying to
figure out how to "measure" conformance with that goal while
the architecture is being drafted, i.e. to specify the
requirements in an unambiguous way that can actually guide
development of the architecture.

So, how could we define/measure/observe interoperability
in the reference architecture as it is being drafted?
Just thinking out loud here ...

- Independently developed formalisms of the reference architecture
could be mapped onto one another.  E.g., my drawing is isomorphic
to yours; they may look different, but if all my pieces map onto
yours and vice versa, then the architecture can be said to
produce interoperable specs. The same could apply to a UML
description, or some other representation of our words.

- The reference architecture can be seen as a "synthesis" 
(in a more or less Hegelian sense) of the existing chaos.
If the reference architecture shows how SOAP/WSDL/UDDI and
REST/Semantic Web are simply different instantiations of
the same abstract principles, we could say that the 
reference architecture is interoperable.  (We could also say
that we are geniuses; I think this is a pretty high standard
to set for ourselves!).

Anyway, I'm just trying to think of ways to define
interoperability that would be useful in defining an
abstraction such as the reference architecture.

-

Received on Monday, 11 March 2002 10:32:15 UTC