Re: SPARQL Protocol for RDF

    From: Giovanni Tummarello <giovanni@wup.it
    <mailto:giovanni@wup.it?Subject=Re:%20SPARQL%20Protocol%20for%20RDF&In-Reply-To=%3C429EFDFE.90603@wup.it%3E&References=%3C429EFDFE.90603@wup.it%3E>>

    Date: Thu, 02 Jun 2005 14:39:26 +0200
    Message-ID: <429EFDFE.90603@wup.it>
    To: "Seaborne, Andy" <andy.seaborne@hp.com
    <mailto:andy.seaborne@hp.com?Subject=Re:%20SPARQL%20Protocol%20for%20RDF&In-Reply-To=%3C429EFDFE.90603@wup.it%3E&References=%3C429EFDFE.90603@wup.it%3E>>,
    semantic-web@w3.org
    <mailto:semantic-web@w3.org?Subject=Re:%20SPARQL%20Protocol%20for%20RDF&In-Reply-To=%3C429EFDFE.90603@wup.it%3E&References=%3C429EFDFE.90603@wup.it%3E>



Seaborne, Andy wrote:

>> to your
>>knowledge would the current protocol allow asking something like a
>>remote CBD as in uriqa ?      
>>    
>>
>
>DESCRIBE, with a system that gives a CBD as the description, will do what you want.
>
>  
>
hurray! :-)

>Theer are many choices for a "description" and they depend on the nature of the data (don't do a blank node closure on a large collection of FOAF - it might not be that helpful).  
>

  

I agree with Andy on this point, and support the decision that DESCRIBE 
not have a single
fixed form of response. Though I still strongly feel that *some* 
generally usable form
of description, as a default, unless otherwise warranted, would make 
facilitate a far
higher degree of interoperability and consistency of behavior.

    Noticed that ;-) stopping when bnodes have IFP would probably be more
    useful.. We need to formalize an MSG definition considering IFPs in it
    wondering how and if the properties change that much.

Note that the latest CBD spec (presumably soon, even today, to be
updated on the W3C site) discusses in length, with examples, the
challenges relating to providing descriptions for graphs with heavy
use of blank nodes and IFPs.

(until the W3C version is updated, c.f. http://sw.nokia.com/uriqa/CBD.html)

Cheers,

Patrick

Thanks again,
Giovanni

Received on Sunday, 5 June 2005 05:49:38 UTC