Re: Links and graphs

Hi John,

I had prepared an email to that effect, but didn't take time to finish it.

On 11/11/2013 08:46 AM, John Arwe wrote:
>  > yes, but it would probably be pretty cool if we could align things
>  > so that LDPCs could be SPARQL endpoints with their ldp:created LDPRs
>  > being the named graphs.
>  >
>  > That would allow us to make a lot of friends on the SPARQL side.
>
> Could you enumerate the gaps and make a concrete proposal or 3 to close
> them?

LDP should define the following invariants:

1. a Resource is a Named Graph [1] (now defined in RDF 1.1)

2. A GET on a Resource URL only returns the corresponding RDF Graph

3. during a REST interaction, the RDF Graph corresponding to the
Resource URL behaves as a Default Graph

4. ldp:created is managed by the Container for all created/deleted
resources

I believe that these invariants match our intuition about Linked Data,
are easy to explain and greatly improve the modularity.


It's now _easy_ to extend an LDPC with SPARQL capabilities: the
Default Graph is defined by 3. and ldp:created would tell us where the
contained LDPRs are (ie. in the linked Named Graphs).

Most of querying features from the current draft are now subsumed by
the following query:

[[
$ curl -X POST \
    --data-binary @query.sparql \
    -H "Content-Type: application/sparql-update; utf-8" \
    http://example.org/netWorth/nw1/assetContainer/
]]

where `query.sparql` is something like

[[
.... WHERE {
  ?ldpcS ?ldpcP ?ldpcO .
  ?ldpcS ldp:created ?ldpr .
  GRAPH ?ldpr { ?ldprS ?ldprP ?ldprO } .
} ORDER BY ?ldprP
]]

Corrolary: no need for "5.1.3 Ordering" either.

Best,
Alexandre.

[1] http://www.w3.org/TR/2013/CR-rdf11-concepts-20131105/#dfn-named-graph

>
> Best Regards, John
>
> Voice US 845-435-9470 BluePages
> <http://w3.ibm.com/jct03019wt/bluepages/simpleSearch.wss?searchBy=Internet+address&location=All+locations&searchFor=johnarwe>
>
> Tivoli OSLC Lead - Show me the Scenario
>

Received on Monday, 11 November 2013 14:59:54 UTC