Semantics Calls 7

The seventh edition of our telecon series will took place 2022-09-26 at 15:00 UTC.

Minutes are below.

VEPs (standing item)"> (1) Current VEPs (standing item)

VEP-009: We are still waiting for proper use cases; it seems we will see some soon-ish from high energy/astroparticle.

There was a question about adding/editing dataproduct-type concepts; since this vocabulary is still a draft, edits can still be made directly until Datalink 1.1 becomes PR.

(2) UCD List 1.5 progress

Iteration on the last typos, last touches to document and build process, make test. Probably ready for TCG review. Mireille will upload and PEN.

(3) UCDs coming up for 1.6

Proposal circulated on semantics mailing list: https://seafile.unistra.fr/f/380b1a54599444d1967b/

Goal: Mapping between ESO observation parameters and UCDs – and these parameters should have UCDs since they will appear in table columns at ESO.

Right now, there is a plan to add a new branch obs.*. But then: Is there really a point to tell apart, say, obs.*.wind from a phys.wind (that perhaps works for other planets, too)?

Markus remarked he is a bit skeptical about the *.cn2 parameters, as they seem a bit over-specific. It would seem, however, that this is a rather general concept in adaptive optics. And there is no simple way to precisely describe it with another combination. Similarly specialised terms that already are UCDs include src.orbital.TissJ (Tisserand parameter with respect to Jupiter).

(Sébastien) For observation parameters, we tried to be very careful not creating new words in both obs;* and phys.* ! In my opinion, the position from the root is not a major issue. It would be more confusing to have similar terms in two branches, e.g. phys.wind AND obs.atmos.wind

(4) UCD trouble in SSAP and SLAP: status reports

pos.outline;instr.fov is fixed in SSAP. There were some doubts because: what about obscore s_region seems rather similar but has a different UCD. On a closer consideration, we believe the two things are distinct enough to justify the distinct UCDs

Also for SSAP, meta.ref.url;meta.curation is fixed in UCD List 1.5.

SLAP: dealing with meta.id;em.line iso em.line (which is secondary). There is an Erratum to fix this; Mireille will push it through the TCG.

(5) FAIRsFAIR requirements and us

See: https://doi.org/10.5281/zenodo.4314321, http://mail.ivoa.net/pipermail/semantics/2021-June/002797.html

This is an RDA output on having FAIR vocabularies. So, how do we think this is important for the IVOA? Would we like to conform to this document?

On the requirement that each term should be resolvable, we're certainly good, as the URIs do indeed resolve, but then the vocabularies as such are not really findable as such from the outside. Perhaps we ought to try to register them through fairsharing.org? But what kind of functionality would be enabled by that? Do they have other vocabularies? Carlo will do a bit more research on whether fairsharing is really the right place.

Marco: They already harvest the IVOA document repo. Perhaps they can harvest the semantics repo, too? Also: perhaps the EOSC has a specific repo for vocabularies. Marco will have a look.

A couple of examples: https://fairsharing.org/search?q=vocabulary https://fairsharing.org/search?q=ivoa

Should we have a cross-vocabulary search engine for our concepts? Fairsfair recommends that, but it seems they are thinking of a global search engine. Has someone set this up? Since our vocabularies are available in RDF form, once such a thing exists, feeding it with our concepts would be easy.

Do we have Provenance on terms? Can we link the VEPs? But how? Fairsfair apparently doesn't say how to do that link. Let's see if someone does and then do like they do. Nobody has volunteered to actually follow up here, so let's all keep our eyes open.

And sustainability? Well: Markus suspects the IVOA will always be a shoestring operation, and so there's very little we can formally do about the danger that it'll all go belly-up. But as long as the stuff is in actual use, Markus suggests there'll be people who will take the modest load of maintaining our doc/schema/vocabulary repos.

(5) organisation, observatory, facility_name, instrument: status report

Postponed to next meeting

(6) arith.diff and arith.ratio in UCDs: Why are they S?

Also postponed.

(7) AOB

There's two mailing lists on iova, UCD and UCD-tech. Does anyone remember what this was for? Marco proposes to remove the users and just keep the archives. Mireille will look into these and then tell Marco.

Baptiste and Mireille are in the process of cleaning out stale UCD pages on the Wiki.