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

Re: REST; good for humans and machines

From: Mark Baker <distobj@acm.org>
Date: Mon, 6 Jan 2003 21:32:54 -0500
To: Peter Furniss <peter.furniss@choreology.com>
Cc: www-ws-arch@w3.org
Message-ID: <20030106213254.Z12258@www.markbaker.ca>

On Mon, Jan 06, 2003 at 10:34:53PM -0000, Peter Furniss wrote:
> Doesn't HTTP cover both layer 6 and layer 7 - and it's the layer 6 features
> that
> seemed initially desirable to "Web Services folk" - who then ignore(d)
> HTTP's
> layer 7 elements, treating it as just a typed-document exchange protocol.

That's correct.  Well said.

> (I've spent entirely too much time on OSI-related architectural discussion
> to
> want to go far down this path. And I especially don't want to think about
> whether
> HTTP has layer 5 features. Oh *!, I just have.)

Yes, let's not go there. 8-)

FWIW, I treat layers 5-7 as one big chunk.  I see little value in
having any commonality in layers 5 or 6, since interoperability is
principally a coordination problem, and layers 5 or 6 have nothing
to say about that.  This is commonly seen in practice; sub-layer 7
systems like IIOP and ONC have never been deployed at scale,
except when an app was bolted on top, like NFS was bolted on to
ONC.  But AFAIK, NFS doesn't have any important interop with any
other ONC based systems; for example, it doesn't magically
interoperate with DNS.

MB
-- 
Mark Baker.   Ottawa, Ontario, CANADA.        http://www.markbaker.ca
Web architecture consulting, technical reports, evaluation & analysis
Received on Monday, 6 January 2003 21:32:29 GMT

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