RE: D-AG0009; Semantic Web & Web architecture

What role, if any, RDF will play in Web Services?

Please point me to the right threads if this question has been asked before.

Thanks.

Hao

-----Original Message-----
From: Mark Baker [mailto:distobj@acm.org]
Sent: Tuesday, March 26, 2002 5:39 AM
To: Mike.Champion@SoftwareAG-USA.com
Cc: www-ws-arch@w3.org
Subject: Re: D-AG0009; Semantic Web & Web architecture


Mike,

> I was not on the teleconference last week, but I think this issue might
> be that some of us want to be convinced that explicitly requiring 
> that the web services architecture be consistent with the "semantic web"
> meets a concrete need of the member companies and our customers.  
> 
> The semantic web initiative may, over the course of this decade, 
> greatly facilitate the "understanding" of diverse business messages by
> software without human intervention.

That's true, but what is also true is that the Web has always been
"semantic" to an important extent.  What I mean by that is that messages
sent on the Web have always been expressions; compound assertions that
have be used to address issues of security, reliability,
routing/dispatch, etc.. since the Web was first devised.

I've given the example of the 410 response code, which is a way for a
resource to tell a client that it was purposefully removed.  This is
"semantic", because a client, even one without a human nearby, knows
what that means.  A "non-semantic" version would be a 200 response with
a body saying "oops, sorry, gone" - a machine would have no chance of
understanding that, though a human would.

What Tim means by "The Semantic Web", is to extend those kinds of
transfer-level semantics to the body of the message, where currently
they are mostly limited to the envelope.  i.e. the content/body of the
message was previously a black box to message processors.  So instead of
sending some HTML with words saying "this is a poem", it could include
an RDF assertion that would tell a machine that it was a poem.  But the
mechanics are basically identical to what happens today with HTTP
headers and response codes, though RDF generalizes the model
significantly.

>  That is obviously very relevant to
> the definition of "web services" that we came up with.  I would VERY
> STRONGLY
> urge the semantic web people to take a requirement (the next time they 
> are chartered, if they don't have it already!) to support many use cases
> from the web services world.

Amen to that. 8-)

FWIW, I think the RDF Primer includes a good example on "Intelligent
Routing" which is definitely very web-services-ish (and FWIW, it's
what my company already does 8-).

http://www.w3.org/TR/2002/WD-rdf-primer-20020319/#routing

>  I think we clearly need a considerable degree
> of liason with the semantic web people to make sure that the web services 
> architecture reflects the concrete progress that the SW can demonstrate in
> meeting
> web services use cases.  I'm a bit leery, however, of any requirement that
> is
> more specific than that we be "aligned" with the SW in some way.  This is
an
> independent activity; we should be "aligned" to benefit from the SW's
> successes, but
> not held back by its failures.

I think "aligned with" is still a fairly strong statement, and I'd
be happy with that.

MB
-- 
Mark Baker, Chief Science Officer, Planetfred, Inc.
Ottawa, Ontario, CANADA.      mbaker@planetfred.com
http://www.markbaker.ca   http://www.planetfred.com