RE: Options we have with respect to the draft charters (i.e., RE: [fwd] Draft charters for work on Semantics for WS)

I agree that keeping semantic annotation out of the actual WSDL file may
be a good idea, particularly if you think about how one might make a
transition between using good-old-Web services and something new and
presumably better.
 
In fact, it has for some time really bothered me that the primary usage
of RDF in XML does not seem to have this character.  That is, it appears
to me that in most implementations data are either in XML or in RDF, in
the sense that if it is in RDF the actual data values appear in RDF
tags, not XML with some sort of semantic decoration that can be ignored
by ignorant processors.  It seems to me that this could cause big
problems in implementation since it makes it sort of an all-or-nothing
thing.  You either have to have your data in XML or RDF -- or you have
to duplicate it.  Backwards compatibility does not seem to be real easy
in this environment.
 
I know I didn't put that very clearly, but I'm not having an overly
coherent day.

________________________________

From: public-sws-ig-request@w3.org [mailto:public-sws-ig-request@w3.org]
On Behalf Of Rama Akkiraju
Sent: Tuesday, November 22, 2005 10:05 AM
To: Bijan Parsia
Cc: public-sws-ig; public-sws-ig-request@w3.org
Subject: Re: Options we have with respect to the draft charters (i.e.,
RE: [fwd] Draft charters for work on Semantics for WS)



Bijan, 

That's a good point. In our discussions with customers, we have noticed
that they tend to be apprehensive about anyone meddling with their WSDLs
even if it is to add semantic annotations and not change any service
interfaces. But I think this hestitation stems from the fact that
in-band annotation of WSDLs is not considered a standard. People would
be more willing to accept such an approach if it were a standard. Not
requiring modifications to WSDL is not necessarily preferable if it adds
to maintaining information about a service in multiple places. In any
case, I can see the point from both sides. I agree that the issue of
in-band & out-of-band annotations of WSDL should certainly be considered
in the WG. 

Regards
Rama Akkiraju

IBM T. J. Watson Research Center
Hawthorne, NY
e-mail: akkiraju@us.ibm.com




Bijan Parsia <bparsia@isr.umd.edu> 
Sent by: public-sws-ig-request@w3.org 

11/19/2005 12:53 PM 

To
David Martin <martin@AI.SRI.COM> 
cc
public-sws-ig <public-sws-ig@w3.org>, Amit Sheth <amit@cs.uga.edu> 
Subject
Re: Options we have with respect to the draft charters (i.e., RE:  [fwd]
Draft charters for work on Semantics for WS)

	





A small technical point which may have been brought up before. (I can't 
really speak to the politics since 1) I've been traveling for the past 
two weeks and haven't caught up and, alas, am unlikely to be able to 
and 2) as y'all know from the workshop I'm a bit resistant to moving to 
a working group in this area, though this one seems somewhat scoped.) 
One difference between the WSDL-S and OWL-S, as I understand it, 
approaches that might now be mooted by WSDL 2.0 but is perhaps worth 
raising is that  OWL-S descriptions are, in general practice, out of 
band annotations on the WSDL document. That is, the OWL-S document 
points to the WSDL document and doesn't *require* any modifications to 
the WSDL. This allows for multiple variant descriptions, perhaps from 
third parties.

Now some of this may be mooted by the RDF mapping...but probably not :) 
In any case, there are clear advantages to the "out of band" annotation 
approach, especially for adoption, so I would like to see such support 
explored by a working group.

(In other words, it's not just integrating with WSDL the spec via 
extension but integrating with deployed WSDL documents that is an 
issue.)

Cheers,
Bijan.