Re: nanopublications are they ok with RDF graphs?

On Sun, Sep 22, 2013 at 11:53 AM, Paul Groth <p.t.groth@vu.nl> wrote:

> Hi All,
>
> I was reviewing the Trig spec. We use trig to express nanopublications -
> light weight forms of provenance attached to graphs. Each nanopublication
> has three graphs associated with it:
>
> - an assertion graph
> - a publication info graph
> - a provenance graph
>
> The provenance graph points to the the assertion graph.
>
> GRAPH :assert { ... }
> GRAPH :provenance { :assert prov:wasDerivedFrom :xyz . }
>

> Is this ok with what's coming out?
>

Paul,

In terms of TriG and other concrete syntaxes for datasets (JSON-LD,
N-Quads) the above
dataset can be safely represented. Using the same PName (or URI or Blank
Node label) in
a graph statement, and as a graph label works perfectly. However, the
interpretation of
`:assert` in the graph labeled `:provenance` referring to the graph
`:assert` is application
specific logic. From RDF 1.1 Semantics:

"When a graph name is used inside RDF triples in a dataset it may or may
not refer to the
graph it names. The semantics does not require, nor should RDF engines
presume,
without some external reason to do so, that graph names used in RDF triples
refer
to the graph they name."

In your case applications dealing with nanopublications would simply need
to be written with the understanding that graph names (IRIs, BNodes)
used in graph in statements should refer to that graph in the dataset.

Could you please respond to public-rdf-comments@w3.org as to whether this
response
resolves your question.

Yours sincerely,
Gavin Carothers


> I ask because we have a ton of post translational modifications and
> protein isoforms modeled like this [1]/
>
> Thanks!
> Paul
>
> [1] http://www.semantic-web-journal.net/system/files/swj461.pdf
>
>
> --
>
> -----------------------------------------------------------------------------------
> Dr. Paul Groth (p.t.groth@vu.nl)
> http://www.few.vu.nl/~pgroth/
> Assistant Professor
> - Web & Media Group | Department of Computer Science
> - The Network Institute
> VU University Amsterdam
>