Re: UDDI and semantics

Paul,

I think we can abstract this discussion to an abtract "REGISTRY" whether it
is UDDI or ebXML or any other.

As someone suggested that there may be no need for a registry at all if the
metadata about web services is published as resources on the web and search
engines harvest, index and categorise that information. Finding a web
service will be no different from finding a news article then. :-)

Cheers,
Max


----- Original Message ----- 
From: "Paul Denning" <pauld@mitre.org>
To: "Max Voskob" <max.voskob@paradise.net.nz>; "www-sws-ig"
<public-sws-ig@w3.org>
Sent: Thursday, November 20, 2003 7:08 AM
Subject: Re: UDDI and semantics


> At 09:54 PM 2003-11-18, Max Voskob wrote:
>
> >Your comments and ideas are very welcome.
>
> Max,
> I will speculate that you will meet resistance to the rdfBag idea as it
> will make UDDI servers more complex to implement.
>
> Lacking the rdfBag, you can now use a categoryBag to point to an RDF or
OWL
> file.  You would need a taxonomy tModel, unchecked, where the keyValue is
> the URI of the RDF or OWL info.
>
> This seems more in keeping with the way things seem to work with UDDI
where
> it points to things outside of UDDI for details.
>
> This would imply that some engine would look though UDDI for all such use
> of this "taxonomy" in a categoryBag, follow the keyValue to retrieve and
> index the RDF/OWL, and provide a query and subscription/notification
service.
>
> Paul
>
>
>

Received on Wednesday, 19 November 2003 17:22:11 UTC