W3C home > Mailing lists > Public > www-ws-desc@w3.org > September 2005

Re: LC 323 thoughts

From: Kendall Clark <kendall@monkeyfist.com>
Date: Thu, 15 Sep 2005 10:51:31 -0400
To: Jacek Kopecky <jacek.kopecky@deri.org>
Cc: David Orchard <dorchard@bea.com>, WS-Description WG <www-ws-desc@w3.org>
Message-ID: <20050915145131.GN32626@monkeyfist.com>

On Thu, Sep 15, 2005 at 04:33:28PM +0200, Jacek Kopecky wrote:
> 
> Hi, 
> 
> I think this is a very good summary of our options. My view is that
> SPARQL in fact does have multiple operations (most probably the four as
> described in point two below) so we should tell DAWG to model their
> interface that way. This will allow semantic description of the
> operations as well later.

If you tell DAWG that, and we were to listen (-wink-), we'd end up having
*20* operations instead of the one we have right now. That strikes me as
completely, well, bad. (How 20? Well: SELECT-xml-get, SELECT-xml-post,
ASK-xml-get, ASK-xml-post, CONSTRUCT-rdfxml-get, CONSTRUCT-turtle-get,
CONSTRUCT-n3-get, CONSTRUCT-ntriples-get, CONSTRUCT-rdfxml-post,
CONSTRUCT-turtle-post, CONSTRUCT-n3-post, CONSTRUCT-ntriples-post,
DESCRIBE-n3-get, DESCRIBE-ntriples-get, DESCRIBE-rdfxml-get,
DESCRIBE-turtle-get, DESCRIBE-rdfxml-post, DESCRIBE-turtle-post,
DESCRIBE-n3-post, DESCRIBE-ntriples-post). Wow, that's a lot of operations
when we thought we had one, query.

(Okay, I may be exaggerating a bit, for effect, here by separating Turtle,
N3, and N-Triples, but *if* it is an exaggeration, it's not much of one.
There are users and use cases for all of those distinct serialization types
of an RDF graph.)

> But we still may consider allowing multiple different output content
> types if we find a use case where we'd agree a single operation can have
> multiple different output content types. How about HTTP GET? Yes, plain
> HTTP GET is what I mean. 8-)

Or how about SparqlQuery.query? It's a service that takes a SPARQL query as
input and returns XML or RDF. Seems straightforward to me. And it means that
all of SPARQL protocol's use cases are suddenly valid, desirable WSDL 2 use
cases.

> But as long as such a use case is not logged as an LC issue (and don't
> log mine, I'm not pushing it in any way), I feel we shouldn't spend much
> time on it.

Well, in my capacity as a DAWG member, I did send such comments earlier
today, and I'd be surprised if they don't get logged as LC comments.

Cheers,
Kendall
Received on Thursday, 15 September 2005 14:52:43 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:37 GMT