Re: A URI for the class of SOAP MEPs

Jacek Kopecky writes:

> The thing is that in W3C, one group should not make up URIs for things
> owned by another group, so we (WS-Desc) contacted XMLP. In order to
> possibly simplify the process, we suggested one URI in their namespace.
> I see now that the current thinking is against combining the resource
> with its description (same URI for both things), so the suggested URI
> turned out to be the wrong one. XMLP will consider this and give us a
> better URI. Not too big of a deal. 8-)
> 
> Personally, I agree that URIs are cheap, but if you want to have
> something at the end of the URI, it may suddenly become not quite cheap
> enough. 8-)

I think we're all set for now.  The chair of the XMLP workgroup has agreed 
[1] to put creation of an appropriate URI onto the agenda of an upcoming 
XMLP teleconference.  While there is some small risk that XMLP would for 
good or bad reason decide not to do something satisfactory, I think we 
should hope for the best, and suspend further discussion here.  Once XMLP 
comes up with a proposed direction, I'll send a pointer to www-tag@w3.org 
and semantic-web@w3.org, as well as to the WSDL WG, so you'll all have a 
chance to register concerns.  Should the XMLP WG decide to mint a URI, as 
we hope, I will remind them of the desirability that the URI resolve.  In 
the meantime, I'd like to believe we're all set.

FWIW:  while XMLP is being asked to mint a URI for the class of all 
possible Message Exchange Patterns (MEPs), the SOAP Rec. has since day 1 
had URIs for the two particular MEPs it does provide (specs at [2,3], the 
URIs themselves are [4,5]).  The good news is that the URIs do resolve, 
which sets a good precedent and suggests XMLP will agree to have the new 
one resolve too.  The possibly bad news is that I find the actual 
representations returned from those URIs to be a bit uninspiring.  From 
[4]:



This URI is used by the SOAP Version 1.2 Part 1 specification dated 07 May 
2003.
The latest version of the SOAP Version 1.2 Part 1 specification can be 
found at http://www.w3.org/TR/soap12/
For more information, please see the XML Protocol Working Group home page 
who produced the SOAP Version 1.2 specification.
Webmaster
$Id: index.html,v 1.2 2003/04/17 14:23:22 ylafon Exp $ 


All true as far as it goes, but not very specific.  Maybe or maybe not the 
TAG feels its worth the trouble to push the XMLP WG regarding the quality 
of these representations.   I'll in any case send a pointer to this 
response to distApp right now, so our XMLP members and chair can consider 
the question.  Thank you.

Noah

[1] http://lists.w3.org/Archives/Public/xml-dist-app/2006Jan/0117.html
[2] http://www.w3.org/TR/soap12-part2/#singlereqrespmep
[3] http://www.w3.org/TR/soap12-part2/#soapresmep
[4] http://www.w3.org/2003/05/soap/mep/request-response/ 
[5] http://www.w3.org/2003/05/soap/mep/soap-response/ 

--------------------------------------
Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------








Jacek Kopecky <jacek.kopecky@deri.org>
Sent by: www-tag-request@w3.org
01/20/2006 11:28 AM
 
        To:     Tim Berners-Lee <timbl@w3.org>
        cc:     noah_mendelsohn@us.ibm.com, W3C TAG <www-tag@w3.org>, 
semantic-web@w3.org
        Subject:        Re: A URI for the class of SOAP MEPs



Hi Tim,

On Fri, 2006-01-20 at 10:00 -0500, Tim Berners-Lee wrote:
> > the suggestion to use a section URI for the concept defined in the
> > section was first (to my knowledge) discussed in WS-Description WG,
> > as those WG members who aren't SemWeb-fluent didn't see the point of
> > inventing new URIs for the purpose of the RDF mapping.
> 
> RDF is made of URIs.
> Making up a URI should not be a big deal.
> Just make an RDF document and stick in what you know about the
> needed things.

The thing is that in W3C, one group should not make up URIs for things
owned by another group, so we (WS-Desc) contacted XMLP. In order to
possibly simplify the process, we suggested one URI in their namespace.
I see now that the current thinking is against combining the resource
with its description (same URI for both things), so the suggested URI
turned out to be the wrong one. XMLP will consider this and give us a
better URI. Not too big of a deal. 8-)

Personally, I agree that URIs are cheap, but if you want to have
something at the end of the URI, it may suddenly become not quite cheap
enough. 8-)

> If analogy is needed, by analogy with
> mapping non-XML format into XML,  it as though "some members
> of the group not so familiar with XML didn't see the need to make
> up tags when defining a mapping into XML"

Good analogy, except in our case XML people already know about URIs,
they only don't understand RDF; whereas in the analogy above the "some
members" would not know about tags so it would be easier to convince
them that they are cheap. 8-)

Thanks for the replies,

Jacek