Re: Reflections on Update


Le 8 mai 09 à 17:58, Seaborne, Andy a écrit :

> This is to suggest that we start with a smaller set of capabilities  
> for update.  Sort of split update into "phase 1" and "phase 2".

> Looking back on the discussion, the timescale we have and the  
> overall set of features, it seems to me that making sure some core  
> part of update made visible to the community early would be a good  
> course of action.
> The core might be changes to some graph, without specifying which  
> graph as part of the update language.  This would also help shake  
> out the alternatives of a non-language based approach although what  
> I took away is that ideas for a non-language approach do not  
> immediately extend to collections of graphs without particular  
> models of relating name to graph (so they aren't a panacea).
> One aspect of a language-based approach I do like is that an update  
> script can be used for local changes, without HTTP.
> A possible core:
> INSERT DATA { triples }
> DELETE DATA { triples }
> MODIFY DELETE { template } INSERT {template } WHERE { pattern }
> And later:
> Graph management
> Graph store
> Multi graph operations

IMHO, LOAD should be also part of the core (unless you considered it  
can be, via HTTP and not within the language ?)
This is clearly needed for stores that act as repositories of remote  
RDF graphs (e.g. personal data management, enterprise middleware, etc.)
Hence, ability to add / remove RDF graphs in the store is imho  
something that should be included in SPARQL/Update - core.



> 	Andy
> --------------------------------------------
>  Hewlett-Packard Limited
>  Registered Office: Cain Road, Bracknell, Berks RG12 1HN
>  Registered No: 690597 England

Alexandre Passant
Digital Enterprise Research Institute
National University of Ireland, Galway
:me owl:sameAs <> .

Received on Wednesday, 13 May 2009 15:26:52 UTC