Semantics

InterOpMay2008 Sessions

Trieste, May 19-23

Session 1 (monday 19, 16:30-18:00, Vulcania 1)

Vocabularies

• Overview of vocabulary document (APM), plus open issues (NG, AG, ...) (ppt)
  • Towards an IVOA Vocabulary (APM, NG, SD, ...)
  • Mapping between Vocabulary terms (Alasdair Gray ppt, pdf)
  • Publishing and maintaining vocabularies (NG, pdf)
  • Discussion to resolve the WD open issues. Conclusions summarised by APM in slides.

We will have a brief introduction (5-10mins) of the themes in the WD, then some talks to highlight the open issues (10-15 mins each) and then a discussion at that point on how to resolve these.

Issue 1: format of master file

Resolved that no format mandated for the master source

Issue 2: format of distributd files

Matthew suggested mandating RDF/XML only, Alasdair agreed. So Resolved that distributed vocabularies MUST include an RDF/XML form, and SHOULD include a Turtle form. Sébastien suggested including warnings about pitfalls when serialising as RDF/XML.

Issue 3: vocabulary versions

Resolved that vocabulary namespaces will contain no version information, but resolve to the most current version of the RDF description of the vocabulary. Vocabulary maintainers SHOULD keep old RDF files available long-term.

Issue 4: vocabulary maintainers

Resolved that vocabularies will not be maintained by a process set up by this recommendation. Vocabularies will be maintained by whichever WG or external authority 'owns' them.

Rob: encourage the various interest groups to get involved and adopt the vocabularies into their processes.

Issue 5: which vocabularies will be distributed

Resolved to include the 5 uncontroversial ones, and refer to, but don't include, the IVOA-T. Sébastien will investigate SKOSifying the Vizier vocabulary, which will act as a further implementation. Someone (can't remember who!) offered to compare the standard IAU constellation list at CDS with the simple vocabulary included in the document.

Issue 6: mappings

Alasdair talked about mappings, recommending that vocabularies declare no inter-vocab mappings; such mappings are published separately; mappings declared in one direction. Thus resolved.

What metadata is required in a mapping file? DC terms? This should be clarified in the document. Some discussion of how vocabularies and mappings should be developed and maintained. There are both technical and social issues to be resolved here, but they don't have to be resolved by this document's process. Norman will investigate wiki-like tools which can support a process like this.

Let this go to Rec, referring to the SKOS WD, with a note at the beginning of the document explaining the situation, and that we'll produce another version when the SKOS WD goes to Rec.

Is everyone happy with those?

Session 2 (tuesday 20, 9:00-10:30, Vulcania 1) (swap with Theory IG) (ppt)

UCDs

• New RFM?

Ontologies

• The ontology of object types and vocabularies (APM, SD, AR)
• Use cases (SD)
• Linking Open Data - arXiv, ADS and more (Norman Gray, pdf)
• ...

Session 3: Joint Semantics/VOEvents/ReR/Theory

(thursday 22, 16:30-18:00, Vulcania 1)

• Using vocabularies (APM, ppt)
  • VOEvent
  • Vocabularies and the registry (SD)
  • Discussion on how to exploit the vocabularies in a VO, e.g. by VOEvent, the registry, etc.
• A draft vocabulary for simulations (F.Le Petit, pdf)