Re: Web Semantics for Datasets

On 7 October 2011 04:04, Sandro Hawke <sandro@w3.org> wrote:
[...]
> Specifically: Consider a "graph naming" to be the association of a
> graph name N with a graph G.  For the graph naming to hold, every
> successful dereference of N yielding an RDF graph must yield G.  For a
> dataset D to hold, its default graph must hold (as normal in RDF) and
> every graph naming pair in D must hold.
>
> Example 1:  This dataset
>
>   <http://example.org> { <s> <p> <o>. }
>
> means that if anyone is able to dereference "http://example.org"
> and obtain an RDF graph serialization, the serialized graph will
> consist of the single triple, <s> <p> <o>.  Failure to dereference
> does not make the graph naming untrue, but a successful dereference
> yielding a different graph does.
>
> Example 2:  This dataset can never be true:
>
>   <http://example.org> { <s> <p> 1. }
>   <HTTP://example.org> { <s> <p> 2. }
>
> ... since one cannot get different results dereferencing URIs that differ only in the case of the scheme component (as per RFC 3986).

(a minor point, but...)

You absolutely can get different results dereferencing URIs that
differ only in the case of the scheme component. This follows from the
observable fact that you can get different results from dereferencing
the selfsame URI at different times (a scenario you bring up later).

For example:
http://www.elsewhere.org/pomo/ and hTTp://www.elsewhere.org/pomo/ are
URIs that differ only in the case of the scheme component. Fetching
them with RFC 3986 compliant software will result in different graphs
/ documents.

Perhaps you mean something like:

* you should not get different results dereferencing URIs that differ
only in the case of the scheme component; if this happens, something
is violating a spec.
* if you get different results dereferencing URIs that differ only in
the case of the scheme component, it is not possible for this to have
been caused by a RFC 3986-approved activity.
* it is impossible to get different results dereferencing URIs that
differ only in the case of the scheme component; if you think you have
achieved this, you are mistaken, and whatever you did does not count
as 'dereferencing'.
* it is possible to get different results dereferencing URIs that
differ only in the case of the scheme component; it is impossible for
the use of different schemes to have causally affected the behaviour
of the responding HTTP server.
* it is possible to get different results dereferencing URIs that
differ only in the case of the scheme component; it is possible but
non-standard for the use of different schemes to have causally
affected the behaviour of the responding HTTP server (perhaps the
request was communicated through some case-preserving non-HTTP-1.1
protocol (e.g. something like http://en.wikipedia.org/wiki/SPDY )

There are a few more variations, but you get the idea.

I suspect a bigger problem is the amount of personalised,
cookie-mediated or session-based content out there, but I need to
think through the rest of your post more carefully.

BTW maybe it's come up already in these discussions, but
http://www.w3.org/TR/HTTP-in-RDF10/ might be useful when describing
the complexities of transactions with the Web, rather than treating
de-referencing as a single pure un-analyzed action. But even the level
of detail catalogued in http://www.w3.org/TR/HTTP-in-RDF10/#example is
never sufficient to guarantee a particular response (as your
time-varying comments acknowledge).

cheers,

Dan

Received on Friday, 7 October 2011 09:54:30 UTC