Re: Yet another problems with URIs and reference

First, if there is an approval process then I would start by modeling
that. A particular assessment and/or approval is an event worth
representing - you've just said so.

Then the question is how to link the approval to the intended
referent. It wasn't precisely clear to me what that is - is it the URI
itself?

To refer to the URI and then connect the URI to the resource you would need to
a) name the URI, "http://xx/Berlin"^^xsd:anyURI
b) connect the URI to the resource

Roughly expressing  (in OBI/IAO terms)

<approval_process1> has_specified_output <recording_of_approval_1>
<recording_of_approval_1> is_about <http://xx/Berlin>
<http://xx/Berlin> named_with "http://xx/Berlin"^^xsd:anyURI

named_with isn't in IAO but conceptually it is a subproperty of
inverse(is_about)
I would add a bit if I were doing this in OBI because I want to
maintain OWL2, and could send you the specific RDF if you were
interested.
But that's the idea.

I leave to you the definitions of approval_process and
recording_of_approval. The former would be an OBO:planned process, the
latter an IAO:data item. Where I live, anyways ;-)

-Alan



On Thu, Nov 10, 2011 at 9:29 AM, Andrea Splendiani
<andrea.splendiani@bbsrc.ac.uk> wrote:
> Hi,
>
> I have a curiosity/problem/request for suggestions which is loosely related
> to the Referee/Referred identification.
>
> In my case, I need to predicate on a URI, intended as an identifier.
>
> So (keeping canonical example), I can have:
>
> http://xx/Berlin                This is the URI of the thing per se.
> http://xx/web/Berlin    Points to a web representation
> http://xx/data/Berlin   Data representation
>
> Now I can predicate in all these documents on http://xx/Berlin referring to
> the actual city.
>
> But I need to predicate on "http://xx/Berlin" intended as an identifier: in
> my case, this goes through cycles of approvals and states. In practice I
> would need to tell whether my URI is approved, reserved, invalid! and so
> on...
> These are clearly not properties of the city of Berlin!
>
> How is this case usually approached ? In principle, it should be quite
> common.
>
> What I was considering (in order of preference):
>
> 1) perhaps some as hoc extension of OWL (annotations). At the end this is an
> overgrown versioning mechanism
> 2) creating a dichotomy between "identifiers objects" and "identifiers
> codes". This could make some sense in my case as identifiers come as codes
> that inherit some 'unconvential' semantics.
> 3) ignoring the distinction and putting it relegating to the semantics of
> relations (a bit like 1, in the end, but dirtier)
> 4) ignore informations on URIs on the web (delegating it to some internal
> engine). This may be a last resort solution.
>
> Any idea ?
>
> ciao,
> Andrea
>
>
>
>
> Andrea Splendiani
> Senior Bioinformatics Scientist
> Centre for Mathematical and Computational Biology
> +44(0)1582 763133 ext 2004
> andrea.splendiani@bbsrc.ac.uk
>
>
>
>
>
>

Received on Monday, 14 November 2011 22:27:44 UTC