Re: protocol draft available

On Thu, Nov 11, 2004 at 08:27:15PM +0100, jos.deroo@agfa.com wrote:
> Kendall,
> 
> While reading your draft (which I more than enjoy!) I was wondering
> how it would look for the cases we use daily..
> (although I could do this myself, I thought your help would be better :))

> http://josd.he.agfa.be:8081/2004/04rm/refGuidelines/pregnancyProtectionFetus.n3
>   http://www.agfa.com/w3c/euler/rdfs-rules.n3
> and modi
>   --nope i.e. "no proof explanation"
>   --think i.e. "find all proofs"
> return the answers for
>   --query 
>  
> http://josd.he.agfa.be:8081/2004/04rm/refGuidelines/pregnancyProtectionFetusQ.n3

Jos,

First, can't do this exactly because my draft doesn't offer any way to
apply arbitrary inference rules to some graph. But let's assume that
yr SPARQL server knows to do RDFS inference on this graph. Your
scenario, as I understand it, is a single RDF query, held in a URI,
against a single RDF graph, held in another URI. 

To prevent wrapping, assume that

U1=http://josd.he.agfa.be:8081/2004/04rm/refGuidelines/pregnancyProtectionFetus.n3
U2=http://josd.he.agfa.be:8081/2004/04rm/refGuidelines/pregnancyProtectionFetusQ.n3

So, 

GET /U1?[QUERY_ID]=U2 HTTP/1.1
Host: josd.he.agfa.be:8081
Accept: text/n3 <!-- you want n3 back, so you con-neg for it -->
User-Agent: agfa-sparql-client/0.0

and the server would respond:

200 OK HTTP/1.1
Server: agfa-sparql-server/0.0
Content-Type: text/n3

:An :mustHave :confirmationOfNonPregnancy.

Since our QL is silent on the issue of inference, I thought it best
that the protocol be silent, too. Otherwise, I think that's the
equivalent protocol request and response.

Best,
Kendall Clark

Received on Friday, 12 November 2004 00:27:00 UTC