Protocol independence and application protocols

[To www-ws]

Just mixing up the order a bit ...

On Mon, Apr 14, 2003 at 01:54:33PM -0400, Champion, Mike wrote:
>I also accept that a (hypothetical) SOAP binding that uses
> PUT to simply transfer messages to a queue identified by a URI rather than
> overwrite the queue [the issue discussed in the messages cited, IIRC] is
> probably not a good fit with HTTP's defined semantics.  But that's an issue
> of the details of a binding from SOAP to a "move bits around" [dodging the
> transport vs transfer issue] protocol, not a issue of whether application
> protocol neutrality is a good thing in principle. 

Fair enough.  I thought it was obvious that the prima facie disregard
of all defined semantics of all application protocols, was necessarily a
bad thing.  But I can accept that there could be at least one person who
doesn't believe this(! 8-).

But it's been my experience that practically all Web services proponents
believe that application protocols *are* transport protocols, not that
they are making a conscious decision to treat them that way despite
recognizing the differences (recall the early "GET is an application
semantic" threads).  If it were the latter, then I don't see why anybody
would have objected to my proposal about talking about the pros and cons
of this approach, rather than just treating it as an axiom which doesn't
need to be studied or defended.  Which brings me to ...

> > Not at all.  If you accept that a SOAP envelope saying "orderBook"
> > means the same thing over HTTP POST as SMTP DATA, then you must also
> > accept that it means the same thing over HTTP PUT.
> 
> I accept that.

Ok, great.  So was there a reason why you didn't support my proposal to
talk about the pros and cons of protocol independence[1] in the WSA
doc?

> This is probably a meta-issue better suited for www-ws, but it seems as
> though you don't believe that there's a middle ground between "REST is the
> alpha and omega of distrubuted computing"  and "REST is nonsense that should
> be ignored."  I believe that SOAP 1.2 has staked out a very useful middle
> ground where it is/

I do think there's a middle ground, and I continually argue for it
whenever I speak about how I believe SOAP should be used to achieve
maximum gain.  But it's very different than the middle ground you
appear to be talking about, and I do not believe that yours will see
much in the way of success for all the reasons I've given in the past;
it simply does not possess the properties necessary to succeed on the
Internet.

 [1] http://lists.w3.org/Archives/Public/www-ws-arch/2003Apr/0045.html

MB
-- 
Mark Baker.   Ottawa, Ontario, CANADA.        http://www.markbaker.ca
Web architecture consulting, technical reports, evaluation & analysis

Received on Tuesday, 15 April 2003 00:18:52 UTC