Re: RDF-ISSUE-105 (datasets-webarch): Graphs, datasets, authoritative representations, and content negotiation [RDF Concepts]

Sorry about the previous mail; I hit 'send' too soon, obviously...

On Wed, Nov 7, 2012 at 4:11 PM, RDF Working Group Issue Tracker <
sysbot+tracker@w3.org> wrote:

> RDF-ISSUE-105 (datasets-webarch): Graphs, datasets, authoritative
> representations, and content negotiation [RDF Concepts]
>
> http://www.w3.org/2011/rdf-wg/track/issues/105
>
> Raised by: Richard Cyganiak
> On product: RDF Concepts
>
> According to AWWW, a URI owner may supply *authoritative representations*
> of the resource identified by that URI, and doing so is a benefit to the
> community [1].
>
> Also, if a URI has multiple representations associated (via content
> negotiation), then fragment identifiers should be used consistently between
> these representations [2].
>
> AWWW also states: “By design, a URI identifies one resource. Using the
> same URI to directly identify different resources produces a URI collision.
> Collision often imposes a cost in communication due to the effort required
> to resolve ambiguities.” [3] Given that RDF graphs and RDF datasets have
> disjoint definitions in RDF Concepts, this raises the question whether
> content negotiation can be used to negotiate between a graph-bearing format
> and a dataset-bearing format.
>
>
My assumption is that, as dataset have no formal semantics, they can
represent the identified resource in various ways, but as we did not
specify one of them to be the "correct" one, we can not really answer the
questions below...

  pa



> This raises a number of questions that this WG should be able to answer:
>
>
> a) Is a Turtle file published at <xxx> containing the triple ":a :b :c"
> equivalent to this TriG file?
>
>    { :a :b :c }
>
> b) Is a Turtle file published at <xxx> containing the triple ":a :b :c"
> equivalent to this TriG file?
>
>    <> { :a :b :c }
>
> c) If a Turtle file containing the triple ":a :b :c" is published at
> <xxx>, and the publisher also wants to provide a TriG file via content
> negotiation (containing only a single graph), what would that equivalent
> TriG file be? Or would this be a URI collision?
>
> d) Given that publishers should use fragment identifiers with consistent
> semantics between content-negotiated representations, what restrictions
> does this TriG file, published at <xxx>, place on the use of the yyy
> fragment in other representations of <xxx>? What about Turtle
> representations? What about HTML representations?
>
>    <#yyy> { :a :b :c }
>
> e) Does the following TriG file published at <xxx> establish a normative
> representation for <yyy>?
>
>    <yyy> { :a :b :c }
>
> f) Does the following TriG file published at <xxx> establish a normative
> representation for <xxx#yyy>?
>
>    <#yyy> { :a :b :c }
>
>
> [1] http://www.w3.org/TR/webarch/#representation-management
> [2] http://www.w3.org/TR/webarch/#frag-coneg
> [3] http://www.w3.org/TR/webarch/#URI-collision
>
>
>
>