W3C home > Mailing lists > Public > www-tag@w3.org > December 2002

Re: Section 4; is REST's uniform interface constraint a "principle"?

From: Mike Champion <mike.champion@softwareag-usa.com>
Date: Fri, 13 Dec 2002 01:10:53 -0500
To: www-tag@w3.org
Message-id: <oprhbx4fhet3hq37@localhost>


Mark Baker writes:

> the Web Services Architecture WG
> have just decided[1] to wait to close the issue I raised[2] on this
> subject until the TAG takes a position on it.

Not exactly.  The WSA WG decided to take the time to consider your issue 
thoroughly, to develop a closely-reasoned reply, and to develop the WSA 
document to the point where it can be referenced authoritatively in that 
reply.  I did say that it was my personal opinion that IF the TAG decides 
that the uniform interface constraint is a principle of the Web as a whole 
THEN the WSA WG would  be obligated to address the issue in detail.   The 
point of taking the time to carefully consider your issue is to make sure 
that any eventual accounting to the TAG is thorough and consistent.

So, we will respond to your issue when we have completed the analysis, 
whether or not the TAG says anything about the "uniform interface 
constraint" in Web Architecture document drafts.  If they do promulgate it 
as a "principle", then we will reference our response to you in whatever 
feedback we supply the TAG as to its relevance to the Web services 
architecture.

Speaking ONLY FOR MYSELF, this is a thorny issue: Roy Fielding makes a 
compelling logical case for the principle, but many argue that it is not 
widely adhered to by many/most Web applications, human-oriented or machine- 
oriented. There is a LOT of procedural code out there that gets invoked by 
what amounts to function names and arguments in URIs, HTTP parameters, 
cookies, and godknowswhat.  Perhaps at some useful level of abstraction 
this is consistent with the "uniform interface constraint", I would not 
presume to judge.  But it is the clear consensus of the WSA WG that this is 
not one of the more pressing questions facing either the TAG or the WSA WG 
at this point in time.

I'd also like to reiterate another point I made in my earlier reply 
(without my co-chair hat on): SOAP 1.2 will recommend various changes that 
should make it easier to achieve the goal that many of us share of making 
Web services mesh more cleanly in with the Web as we know it, e.g., so that 
one can hyperlink to or bookmark a web service that simply retrieves 
information.  I would strongly suggest giving them time to get some 
traction in the toolsmith and user communities; if you are right about the 
power of ideas such as the uniform interface constraint, pragmatic 
developers won't need a theoretical principle to justify using it, it will 
"just work".  SOAP 1.2, WSDL 1.2 (IIRC), and the WSA will *allow* 
developers to use various principles promulgated by Dr. Fielding 
individually and the TAG collectively.  Let's see how these experiments 
turn out before making Recommendations or establishing Principles on the 
subject.
Received on Friday, 13 December 2002 01:11:34 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:14 GMT