W3C home > Mailing lists > Public > www-ws-arch@w3.org > August 2002

Summary of what we already know

From: Paul Prescod <paul@prescod.net>
Date: Sat, 10 Aug 2002 05:23:44 -0400
Message-ID: <3D54DBA0.49B04629@prescod.net>
To: www-ws-arch@w3.org

I find myself repeating a couple of points so I would like to clarify:

 * Mark and I already understand that web services are about machine to
machine communications and that the existing web is dominated by machine
to human communications. Neverthless, if you read the specifications for
URIs, HTTP, XML and the rest of the Web, you will find essentially
nothing in them specific to problems where human beings directly
interpret representations of resources. XHTML is an obvious roadblock.
XML removes it. The communication of semantics is another tricky
problem. RDF, OWL, DAML-S, etc. are working on that problem. Mark and I
are trying to promote an architecture where we have a hope of using XML
and SemWeb technologies to address the machine-to-machine issues.

 * Mark and I understand that REST was originally designed for
hypermedia in the traditional sense. This is not evidence that it is
inapplicable elsewhere. XML was not designed for protocols and ebusiness
either. REST comes out of the same document processing and hypermedia
tradition that XML does. XML (especially while it was still SGML) faced
all of the same biases: "That's about human communication, this is about
machine-to-machine. That's about documents, this is about data. That's
too loose, we need something strict. That's too general, we need
something specific to our problem domain". Hypermedia is *one
application* of REST, as technical documentation is *one application* of
XML. Roy Fielding has always been explicit that REST is more general
(look at the abstract for the HTTP specification).

 * If I offer Expedia as an example of a REST web service it doesn't
mean that I think that Expedia is ready for machine-to-machine
communications today. It means that the differences do not lie in the
*protocols* but in the *data*. Machine-to-machine Expedia should of
course use XML and probably RDF and DAML or DAML-S. I believe it is
fundamentally counterproductive to fight against the established
dominance of a protocol that is already more general in design than
anything that preceded it or anything that has been proposed.

Now that this message is the archive, I'll point to it the next time
someone says: "You don't understand. We need ot do ebusiness and take
humans out of the loop."
-- 
XML, Web Services Architecture, REST Architectural Style
Consulting, training, programming: http://www.constantrevolution.com
Come discuss XML and REST web services at the Extreme Markup Conference
Received on Saturday, 10 August 2002 05:27:56 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:04 GMT