W3C home > Mailing lists > Public > www-archive@w3.org > August 2004

WS-Addressing and Web architecture

From: Mark Baker <distobj@acm.org>
Date: Tue, 10 Aug 2004 17:10:33 -0400
To: hugo@w3.org
Cc: www-archive@w3.org
Message-ID: <20040810211033.GI30914@markbaker.ca>

Hi Hugo,

I wanted to convey my concern about the WS-Addressing submission from
a Web architecture perspective.

One of the principle objectives of this spec is to standardize on an
identification structure for message endpoint references.  The spec
claims that it needs to do this in a transport independent manner;

"Web Services Addressing (WS-Addressing) defines two interoperable
constructs that convey information that is typically provided by
transport protocols and messaging systems. These constructs normalize
this underlying information into a uniform format that can be processed
independently of transport or application."
 -- http://www.w3.org/Submission/2004/SUBM-ws-addressing-20040810/#_Toc77464314

Unfortunately, as with a lot of other efforts under the (IMO misguided)
"protocol independence" banner, Web architecture is contravened in
important ways.  In particular, what this means in this case is that
the URI to which a SOAP envelope is POSTed, no longer identifies the
resource which is to receive the message.  Or, looked at another way,
the resource which receives the message is explicitly not identifiable
by a URI.

IMO, URIs should be used to identify endpoint references, and moreover,
should be in the underlying protocol portion of the message envelope,
not in the SOAP portion of the envelope.

Unlike other specs such as SOAP or WSDL, I see no way in which the EPR
portion of this spec can be used in a manner even approaching
being consistent with Web architecture.

I'm not sure what action to take at this point.  I'd like to get your
feedback first though.

Thanks.

Mark.
Received on Tuesday, 10 August 2004 21:09:40 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:42:44 UTC