RE: Web Service Definition [Was "Some Thoughts ..."]

Comments embedded below.

> -----Original Message-----
> From: Joseph Hui [mailto:jhui@digisle.net]
> Sent: Thursday, February 28, 2002 6:10 PM
> To: Vinoski, Stephen
> Cc: www-ws-arch@w3.org
> Subject: RE: Web Service Definition [Was "Some Thoughts ..."]
> 
> 
> > -----Original Message-----
> > From: Vinoski, Stephen [mailto:steve.vinoski@iona.com]
> > Sent: Thursday, February 28, 2002 2:03 PM
> > To: Joseph Hui
> > Cc: www-ws-arch@w3.org
> > Subject: RE: Web Service Definition [Was "Some Thoughts ..."]
> > 
> > And as I explained on the teleconference: discovery isn't a 
> necessary
> > part of the definition, because I can write a system where 
> > applications
> > access web services via URIs and communicate with them via standard
> > protocols, all without the aid of any discovery service such 
> > as UDDI or
> > WS-Inspection. I have real-world proof of this: some of 
> > IONA's customers
> > have web services in production right now and are not using a 
> > discovery service in their systems.
> > As for description: the fact that XML-RPC has been around 
> for several
> > years now and has facilitated the creation of web services 
> without the
> > need for a description language ala WSDL should be proof enough that
> > description is not a necessary part of the definition, either.
> 
> The point of contention wasn't whether such web services exist,
> but rather should they be touted as web services built to standards.
> I'm uncomfortable that adding D&D to the definition may alienate the
> installed base you mentioned -- the "do-no-harm" principle, one might
> say; but I find it hard to ignore the benefit that D&D can bring
> in helping to promote the practice of *standardized* web services.

The definition does not disallow D&D -- rather, it explicitly does not
include them because they are not necessary. URIs, standard internet
protocols, and non-human-driven application-to-application interaction
are key, but D&D are not, as I and others have already explained.

> Now, I think we can work this out.
> Since your properties list does not use RFC-2119 terms like MUST
> SHOULD MAY, and it's not been said a WS must possess all properties
> listed.  That leaves the room for saying a WS must possess some of
> the properties listed.  In that case, then what problem will adding
> D&D to the list cause?  Would this work for you?

Not really, because RFC-2119 is about standards language. Our Web
Services definition need not use standards language, and in fact it
would be best if it didn't, to maximize broadness and generality. Let's
leave the standards language for our actual standards.

--steve

Received on Thursday, 28 February 2002 19:54:50 UTC