Re: FW: [xri-comment] Re: [xml-dev] OASIS TC Call For Participation:

[Mark Baker <distobj@acm.org>]

Hi,

I'm not subscribed to this list, so please keep me CCd.

Lindelsee, Mike wrote:
> I'm interested in knowing more about what you are proposing 
> here.  A little
> more detail might help me decide if I agree or disagree. ;)

It's hard to get the full motivation for the existence of this WG
from the announcement, but a couple of specific points jumped out to
me;

"Specifically, this service will reflect the simple
transactional nature of the WWW, i.e., it will use a small set of REST
(Representational State Transfer) or CRUD-like operators on an
infinitely extensible set of XRI-addressable resources."

and

"5) A REST/CRUD-like data exchange service and bindings for HTTP and SOAP"

What both of these have in common, is that they appear to assume that a
REST-like "exchange service" doesn't already exist.  HTTP is, in fact,
such a beast, as REST is the architectural style used by Roy Fielding to
craft its construction (as well as URIs).

If the announcement had given some indication as to why HTTP was felt to
be an inadequate solution for this problem, and a new "REST/CRUD-like"
facility designed, well, ok, I'd probably be responding to that instead.
But I saw no such indication, so I'm questioning the need for the
existence of this WG and its deliverables at all.

As a technical matter, the comparison between REST's uniform interface
and CRUD is superficial (a small set of general methods), so shouldn't
be taken too far.  Other systems, such as tuple spaces, share this
trait as well, but are also not following REST's uniform interface.  The
key differentiator about the uniform interface is that all methods are
meaningful to all things with identity, which is different than CRUD
(all methods operate on data sets), or tuple spaces (all methods operate
on containers/spaces).

Thanks.

BTW, a soon-to-be-submitted Internet-Draft authored by Dan Connolly of
the W3C and myself may be relevant here;

http://www.markbaker.ca/2002/09/draft-connolly-registries-00.txt

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