Re: [GRAPHS] Follow your graph nose.

Le 02/03/2011 14:39, Fabien Gandon a écrit :
> Antoines,
>
> As I said at the end of my message I can imagine this third option but I didn't see it mentioned so far.

I should have paid more attention to your email, I think I overread the 
last sentence :P

>
> However I think it also suffers from some of the problems mentioned for each option.
>
> For instance one problem raised at the workshop and equivalent to what was done within the POWDER WG is that a graph may be huge and you might want to know that before retrieving it.
> In some senarios (e.g. mobile web) it is good to be able to query only the metadata about the resource before deciding to retrieve the resource or in our case part of the resource (typical use case of POWDER).
>
> Thus the second option has the advantage of providing you with metadata about the graph that you can use to decide whether or not to retrieve the full graph.


Ok. From what you say, I have the impression that choosing between 
option 1 or 2 is more a matter of punctual choice that each data 
provider has to make for their own data.

I am not really aware of what the POWDER WG is doing in details, but I 
was reading the draft for voiD [1] which recommends RDF descriptions for 
the metadata you're talking about. However, it distinguishes Datasets as 
separate entities, and RDF graphs. Maybe the metadata you're mentioning 
should be attach to a dataset description, and let the graph dereference 
to the data. Especially, you're example in one of your other emails with 
a SPARQL endpoint, would better fit in a voiD description, since 
endpoints are attached to datasets rather than to a single graph.

Just my thoughts, but I'm open wrt this issue.


[1] Vocabulary of Interlinked Datasets. Editor's draft. 
http://www.w3.org/2001/sw/interest/ED-void-20110301/


Regards,
AZ

> Cheers,
> --
> fabien, inria, @fabien_gandon, http://fabien.info
>


-- 
Antoine Zimmermann
Researcher at:
Laboratoire d'InfoRmatique en Image et Systèmes d'information
Database Group
7 Avenue Jean Capelle
69621 Villeurbanne Cedex
France
Lecturer at:
Institut National des Sciences Appliquées de Lyon
20 Avenue Albert Einstein
69621 Villeurbanne Cedex
France
antoine.zimmermann@insa-lyon.fr
http://zimmer.aprilfoolsreview.com/

Received on Wednesday, 2 March 2011 14:19:44 UTC