Re: Multiple resource representations during LDPC POST (5.4.1)

John,

DESCRIBE output is an RDF graph as well as "resource representation"
is. And you're exactly right -- no problem if LDP POST takes RDF graph
as input -- but why call it "representation" as 5.4.1 does, and not
"graph"?

I'm arguing that one cannot PUT or POST a resource representation (in
full). One can only GET it, possibly as product of multiple graphs, as
with the DESCRIBE result.

I think "description of resource" from SPARQL is actually the true
meaning of LDP "resource representation", and defined better:
http://www.w3.org/TR/sparql11-query/#descriptionsOfResources

One (political) problem for the LDP specification, if it moves from
using "representations" to "graphs", is that it becomes even more like
Graph Store Protocol, and therefore less relevant. It is not a
practical problem for developers though -- I have implemented an
LDP-like system without all the complexities of this specification,
using SPARQL 1.1 queries and Graph Store Protocol only.

Martynas
graphityhq.com

On Tue, Jan 7, 2014 at 4:33 PM, John Arwe <johnarwe@us.ibm.com> wrote:
> Martynas, this feels like violent agreement ;-)
>
>> How can the client submit the complete representation in a single RDF
>> graph (which here is the entity body), if a resource representation is
>> potentially constructed from multiple graphs, as my DESCRIBE example
>> shows? ...
>
> SPARQL DESCRIBE's output is an RDF graph, according to [1].  So "what's the
> problem?" if LDP's create-via-POST takes as input an RDF graph?
> If you're looking to preserve provenance, I think you're beyond Turtle and
> into other syntaxes like Trig.  LDP requires Turtle but certainly allows
> others, so Trig and friends can play.
>
>> ... Or is it somehow partial representation that is meant here? Or
>
>> does the LDP prohibit multi-graph representations?
>
> <chorus> LDP requires Turtle but certainly allows others, so Trig and
> friends can play. </chorus> ;-)
> I don't know that we have an explicit MAY on others, but I suspect we do.
> Since the default reasoning is "in the absence of a MUST NOT, you can do
> anything" silence implicitly allows it; people sometimes object to explicit
> MAYs that convey the equivalent, but how far you take each is somewhat a
> matter of editorial taste.
>
>
>
> [1] http://www.w3.org/TR/sparql11-query/#QueryForms
>
> Best Regards, John
>
> Voice US 845-435-9470  BluePages
> Tivoli OSLC Lead - Show me the Scenario
>

Received on Saturday, 11 January 2014 12:26:40 UTC