Re: ISSUE-5 Definition of Resource

I went back through Cool URIs and httpRange-14 discussions, and some other 
stuff to check up on this.

Which terms in the message are being proposed to be put up for working group 
consensus, by the way?

On 12/18/2014 12:55 PM, Arthur Ryman wrote:
> I'd like to summarize the discussion from the WG today and ask that we
> arrive at a consensus on the meaning of terms. Here are definitions that
> align with W3C specs.
>
> From a web point of view, a resource is any identifiable thing. We
> identify them using URIs.

A resource is anything.  Not all resources are identifiable.  Not all 
identifiable resources are identifiable via URIs.  Not all resources that are 
identifiable via URIs are identifiable via HTTP URIs.

> From an HTTP point of view, there are two kinds of resource, namely
> information resources and real-world objects. The term "real-world object"
> denotes any resource that is not an information resource. This implies
> that fictional characters are real-world objects.

That's the terminology in Cool URIs for the Semantic Web 
http://www.w3.org/TR/cooluris/.  Older terminology talked about information 
resources and resources. Even though it seems strange to say that information 
resources are less real than fictional characters there appears to be no 
reason to not use the Cool URIs terminology.

> An HTTP server should return a 3XX response code when a real-world object
> URI that it hosts is requested via HTTP GET. The response should redirect
> to an information resource URI that has information about the real-world
> object.
>
> An HTTP server should return a 2XX response code when an information
> resource URI that it hosts is requested via HTTP GET. The response should
> contain a representation of the information resource in some content type,
> ideally one of the content types given by the Accept header.

In general, yes.  Various error conditions might produce different responses.

Hash URIs are handled using a different method, but the URI access part works 
this way.

> For the purposes of the wg, we are interested in RDF content types.

For the RDF graph that is being validated, yes.

> An RDF representation consists of a set of triples which can be thought of
> as forming a graph, technically a directed, labelled graph.

A set of RDF triples can be thought of as a kind of graph, but not exactly 
this kind of graph.  I'm not sure what "RDF representation" is supposed to be 
here, though.

> The nodes in an RDF graph are labelled by RDF terms, i.e. URI, Literal,
> and Blank Node. The arcs are labelled by URIs. There are other constraints
> defined in the RDF specs.

An RDF graph is a formal construct.  Nodes in an RDF graph are either IRIs, 
literals, or blank nodes.  RDF graphs don't have arcs or edges, instead they 
have triples whose subject is either an IRI or a blank node, whose predicate 
is an IRI, and whose object is either an IRI, a blank node, or a literal.

When one thinks of an RDF graph as a regular kind of graph, one has to be very 
careful to be faithful to RDF graphs.  In particular, the graphs that 
correspond to RDF graphs are multi-graphs.

> Since we can visualize graphs as geometric objects, the term "shape" has
> been adopted to describe sets of graphs that share certain
> characteristics, e.g. those required by some application. A shape
> describes the expected contents of a graph. This includes expected arc
> labels, occurrence constraints, etc.

This bit was the subject of a previous email.

> The term "resource shape" is an abbreviation for the "shape of the graph
> of the RDF representation of an information resource".

This last bit was the subject of a previous email.

peter

>
> _________________________________________________________
> Arthur Ryman
> Chief Data Officer
> SWG | Rational
> 905.413.3077 (phone) | 416.939.5063 (cell)
> IBM InterConnect 2015
>
>

Received on Friday, 19 December 2014 19:33:30 UTC