Re: KGC-1: capability discovery

On Thu, 2004-03-18 at 11:01, Kendall Clark wrote:
> -- Name --
> 
> Generically discover/publish server capabilities
> 
> -- Description --
> 
> Abelard often needs to query an arbitrary list of RDF storage servers
> for assertions about a set of URIs he cares about.

Abelard... great name... now... why does he care about these URIs?
What life persuit/business is he in? What's his role?
Is he an IT application designer?
A web site designer? A financial analyst? A plumber?

Can we find something less technical than "capability discovery"
to label this use case with? I think you've done a good job
of motivating a requirement around preserving an investment in extant
data access mechanisms. But I'd like the story to start with
something more real-world than "an arbitrary list of RDF
storage servers".

>  Abelard's client
> software includes support for three different query
> languages. Abelard's client software connects to each RDF storage
> server and determines whether it supports one of the three query
> languages it knows about. Abelard's client software chooses, based on
> priorities set by Abelard, to send different queries to different
> servers.
> 
> Heloise makes RDF available over the Web using an RDF storage
> server. Heloise's server, implementing DAWG, supports the DAWG query
> language; but it also supports the Versa[1] RDF path query
> language. Heloise's server makes its query language capabilities
> available in a machine readable form. It negotiates with clients in
> order to choose the most appropriate query language that they have in
> common.
> 
> -- Value --
> 
> Abelard can formulate queries in a variety of query languages, which
> preserves his existing investment. Abelard's software vendor and
> Heloise's software vendor can develop and sell extensible, relatively
> generic systems.
> 
> Abelard and Heloise are able to exchange data in an interoperable
> fashion without requiring out-of-band, human negotiation about
> capabilities.
> 
> -- Other --
> 
> Rather than writing several parallel use cases just like this one, I
> invite the reader to substite the following for "query languages" in
> this use case:
> 
>      * RDF serialization types
>      * context support
>      * publicly available models with read access
>      * publicly available models with write access
> 
> Footnotes: 
> [1]  http://uche.ogbuji.net/tech/rdf/versa
-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
see you at the WWW2004 in NY 17-22 May?

Received on Thursday, 18 March 2004 13:01:30 UTC