Re: Proposed issue; Visibility of Web services

I think about a year ago Paul Prescod was writing that the
purpose of SOAP was to foster the invention of myriad new
protocols.  I wonder if that's changed.

Abstractly, it's fine to say that SOAP intermediaries understand
SOAP messages, and so you have visibility.  What's not clear
is how that visibility extends into the application layer where a
custom application protocol lives.

This ain't rocket science.  If you just invented something I haven't
heard of yet, then my visibility into it is nil.  It's possible for me
to gain visibility, but that comes at a cost.

I suspect this discussion might do well to query the cost of
visibility in different concrete scenarios.  It seems like a no-brainer
that an application can be made visible to a set of intermediaries
by sync'ing up the latter to the custom protocol.  Whether that's
a good idea or not, though, depends on the cost, does it not?

My Two Cents,
Walden

----- Original Message ----- 
From: "Mark Baker" <distobj@acm.org>
To: "Anne Thomas Manes" <anne@manes.net>
Cc: <www-ws@w3.org>; <noah_mendelsohn@us.ibm.com>
Sent: Tuesday, May 27, 2003 2:11 PM
Subject: Re: Proposed issue; Visibility of Web services


> 
> On Tue, May 27, 2003 at 01:41:16PM -0400, Anne Thomas Manes wrote:
> > Mark Baker said:
> > > A generic processing model is not a generic application.
> > 
> > Maybe not, but that's not the point of this discussion. The point is
> > visibility.
> 
> ... and the role of a generic application in improving visibility.
> 
> > A SOAP intermediary has excellent visibility into SOAP messages.
> 
> I don't believe so.
> 
> One more time, from the top ... 8-)
> 
> A generic HTTP intermediary has better visibility into an HTTP
> transaction than a generic SOAP intermediary has into a SOAP
> transaction, because generic HTTP intermediaries are hardcoded to
> understand HTTP application methods, while generic SOAP intermediaries
> aren't hardcoded to know about any application methods.
> 
> Even if you believe that HTTP is just for humans and browsers, this
> should be self-evident, I believe.  But I understand, first hand, that
> it takes some reworking of mental models to grok.
> 
> MB
> -- 
> Mark Baker.   Ottawa, Ontario, CANADA.        http://www.markbaker.ca
> Web architecture consulting, technical reports, evaluation & analysis
>   Actively seeking contract work or employment
>

Received on Tuesday, 27 May 2003 14:31:09 UTC