W3C home > Mailing lists > Public > www-ws-arch@w3.org > January 2003

Re: Binding

From: Walden Mathews <waldenm@optonline.net>
Date: Wed, 08 Jan 2003 09:27:10 -0500
To: "Champion, Mike" <Mike.Champion@SoftwareAG-USA.com>, www-ws-arch@w3.org
Message-id: <003f01c2b722$07ec6860$1702a8c0@WorkGroup>

Mike,

I didn't see a response to these points by Mark, and I think there's
stuff worth getting into here.  I've replied to similar comments by
David Orchard in a different message.

>
> Well, sorta.  I see your point anyway.  The a priori information on how to
> access the specific service is encoded in the URI in REST,

It may look that way, but actually it's not.  It's encoded in two places:

  (1) in the web of server-side resources,
  (2) in the generic method interface

But I'd like to emphasize (1) right now.  If we're talking about queries,
the full context for asking a simple question is provided there.  Here's
an analogy that might hit home.  When you write a requirement specification,
if you spend enough time in the exposition of the problem domain, then
the expression of requirements themselves is simplified down to stating
simple conditions in the language of that domain.  By analogy, the web
of resources is that dense context, URI space is the language defined in
that context, and GET <uri> is the elegant invocation of the "language".

For contrast, if you attempt to write a requirement without having
defined the terms of the problem domain, then each requirement
statement is loaded with the burden of the context to make it make
sense.  That's the query with a zillion parameters.

Er, I think I may have been baffling, but I'm willing to try some more,
given any sign of interest at all.


[snip]

> URIs leverage the Web more fully...

Not the kind of URI you were talking about.  Just because Amazon has
URIs that won't fit on the side of a bus doesn't mean a good REST URI
has to look like that.  It's an appropriate place to expend some design
effort when you're building a RESTful web application, in my experience.

> I think this distinction between alternative ways of encoding and
exchanging
> the a priori information needed to invoke a web service, and the tradeoffs
> it engenders, is definitely worth noting in the WSA document.  Thanks, I
> think we're actually getting somewhere today.

May today be so fruitful...

Walden
Received on Wednesday, 8 January 2003 09:28:23 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:12 GMT