Re: Service or graph store naming.

On Mon, Feb 7, 2011 at 3:22 PM, Gregory Williams <greg@evilfunhouse.com> wrote:
> ..snip..
> Oh dear. The SD document seems to have slipped into some confusion here. My recollection is that the modeling:
> [] a sd:Service ; sd:url </sparql>
> was done because there's not always a one to one relationship between service and endpoint url. Doing the modeling as
>
> </sparql> a sd:Service
>
> means that to support situations where a service is accessible through several URLs means the SD must either duplicate a lot of the triples or support sameAs reasoning.

Ok, I see.

>> In the way you think about the service and the graph store, what is
>> the relationship between them? Is it one-to-one? i.e.:
>>
>> sd:defaultDatasetDescription a owl:FunctionalProperty
>
> I agree with this. I can add this to the SD vocab if there's agreement.

Ok, but I see in Andy's previous email, he doesn't seem to agree with
this.  I will follow-up there.

> ..snip..
>> @prefix tag: <http://w3.org/2001/tag/awwsw/http#>.
>> sd:manages
>>  a owl:FunctionalProperty;
>>  rdfs:subPropertyOf sd:defaultDatasetDescription;
>>  rdfs:domain sd:RESTDatasetService;
>>  rdfs:range GraphStore
>
> Are you suggesting adding this to the SD vocab?

This snippet was just meant to make sure the way I'm thinking about a
dataset HTTP protocol instance and a graph store matches how you and
Andy think about it.  Once there is consensus we probably won't need
all those constraints and new terms.

>> GraphStore owl:intersectionOf ( sd:DataSet tag:Get200WildRdfSubject
>> MutableWebResource )
>
> I don't know what tag:Get200WildRdfSubject is, nor how MutableWebResource might be represented.

tag:Get200WildRdfSubject is defined as

[[[
This class consists of a consensus union of classes of the
200-responders that occur in RDF assertions "in the wild". What
belongs here is pretty much an empirical question that should be
answered by looking at RDF that has been applied to URIs for which (a)
200 responses are delivered for GET requests
]]]

So, it's basically all the RDF resources in Linked Data.  I was just
trying to think about how to formally define a graph store as we have
been thinking about it: an RDF dataset that is mutable and 'web
accessible'.  The idea of a mutable web resource is not defined in any
RDF vocabulary I know, so I was using MutableWebResource.

-- Chime

Received on Tuesday, 8 February 2011 21:47:28 UTC