RE: Web Services Definition and XML

I agree with Krishna on this. Lets define Web Services broadly here. We
don't need the word XML all through it.  If we want to
qualify "standards based description" to  "standards based XML description"
I can live with that.

I think that the architecture we define will have XML all over it as we
identify technologies to fill the various roles and aspects we define.
I don't want to say right now that if the bits on the wire aren't XML or
weren't derived from XML you don't get to be a web service.  What and
how things go on the wire is part of the architecture... not the
definition.

IMO, I think if we restrict it to XML on the wire we throw out LOTS of very
interesting real world, business use cases... creating divergence as
vendors go elsewhere for technology and architecture. We would be throwing
out our 'existing art' as well with SOAP attachments and MIME. Those
attachments
aren't XML.  We would be drastically restricting the applicability of Web
services to the enterprise integration problem if we restrict the wire to
XML. WSDL does not restrict the wire to XML. The industry is looking for
guidance on Web Services Architecture, lets not dismiss a bunch of them
right off the bat.

Small suggestion for progress...after we get through the defining the
requirements on Chris' schedule, we will get to define the architecture...
I think this is when we should nail this down.

Heather

"Krishna Sankar" <ksankar@cisco.com>@w3.org on 03/04/2002 10:50:41 PM

Sent by:    www-ws-arch-request@w3.org


To:    <www-ws-arch@w3.org>
cc:
Subject:    RE: Web Services Definition and XML



Dave,

 | The charter seems extremely clear that web services must be based upon
XML.
<KS>
 I reread the charter. To me the charter does not imply either that our
*definition* of web services must be based on XML not XML as the *sole*
implementation of web services.

 It only says that the set of technologies *identified* by this WG should
be
based on XML. It also says that the WG does not have to design the
technologies. As an extreme case, another WG or a later version can
identify
a very different stack of technologies, based on our definitions.

 In this sense, define as broadly as required and identify a set of XML
technologies to implement that definition, is our marching order. So if we
define web services as "using standard interfaces and using internet
protocols" we are covered. We need to identify WSDL and SOAP as the
description and message technologies.

 BTW, the frequency of the word XML is irrelevant here.
</KS>

cheers

 | -----Original Message-----
 | From: www-ws-arch-request@w3.org [mailto:www-ws-arch-request@w3.org]On
 | Behalf Of David Orchard
 | Sent: Monday, March 04, 2002 3:23 PM
 | To: www-ws-arch@w3.org
 | Subject: Web Services Definition and XML
 |
 |
 | I wanted to discuss a specific aspect of Web Services definition on a
 | separate thread, particularly the use of XML.
 |
 | If one takes a look at the charter of the Web Services Architecture
 group
 | [1], the word XML is the 4th word in the text.  The first 7 sentences
 | mention XML 7 times.  I'm counting as one the XML, XML
 | Namespaces, and XML
 | Schema fragment.
 |
 | Further, the 2nd goal is "The set of technologies identified
 | must be based
 | on XML. ".
 | The 6th bulleted goal is "The framework proposed must support the kind
 of
 | extensibility actually seen on the Web: disparity of document formats
 and
 | protocols used to communicate, mixing of XML vocabularies using XML
 | namespaces, development of solutions in a distributed
 | environment without a
 | central authority, etc. "...
 |
 | The charter seems extremely clear that web services must be
 | based upon XML.
 |
 | Now I'm a person that leans towards sometimes re-interpreting
 | charters, but
 | I draw the line in the sand on this one.  I believe that the Web
 Services
 | definition MUST make explicit reference to XML.  Perhaps the
 | actual bits on
 | the wire don't have to be XML - like using SSL or GZIP - but the
 | basis for
 | the inputs and outputs of the service sure have to be XML or a well
 | understood transformation.  I also include a packaging of XML
 | into something
 | like MIME or DIME as being XML based.
 |
 | Like I argued for URIs, I will also argue for XML in our
 | definition.  This
 | is a show-stopper.
 |
 | Cheers,
 | Dave
 | [1] http://www.w3.org/2002/01/ws-arch-charter
 |
 |

Received on Wednesday, 6 March 2002 11:03:24 UTC