VOTable and VOUnits

[Markus Demleitner]

Dear Colleagues,

On Fri, May 19, 2023 at 12:09:41PM +0200, Molinaro, Marco wrote:
> Il giorno ven 19 mag 2023 alle ore 11:54 Mark Taylor <
> m.b.taylor at bristol.ac.uk> ha scritto:
>
> > I believe we could write MUST without breaking the rules;
> > it's not true that previously valid VOTables would become invalid,
> > since only VOTables declaring version='1.5' would need to be
> > bound by new requirements introduced in VOTable 1.5.

Nice trick.  I like it.

> > But, my opinion is that we ought to stick with SHOULD at least for
> > now, and perhaps indefinitely.  It's nice if units can be
> > machine-readable, so VOUnit should be encouraged, but my guess is
> > that the overwhelming usage of the units attribute in VOTable is
> > being read by humans, and for that purpose non-VOUnit-compliant
> > units are much better than no units at all.

If you're talking about all the mad units introduced for technical
reasons (or, umm, linguistic economy) like the almost proverbial
"millicrab": VOUnits provides for that.  You'd be writing m'Crab'.
Or, if you absolutely have to, 'dex'.

If you're talking about large amounts of metadata in different
syntaxes:  Well, that *is* a point. But then... VizieR has cleaned up
their unit strings recently, and if a MUST here can entice other data
centres to do the same thing, that'd certainly be wonderful.

By the way, pyVO (if I understand matters right -- Brigitta, perhaps
you can chime in) is transitioning to making QTables from stuff
coming from the VO by default, in which case (I think) units are
parsed by a machine all the time whenever pyVO is in the game.  This
may shake the "overwhelming usage" argument a bit.

Still:

> > The danger of writing MUST here is that data providers who have a
> > load of human-readable-but-non-VOUnit-compliant units will provide
> > no units rather than rewriting them, to avoid compliance errors.

I think none of us is confident enough for a MUST at this point, and
probably a SHOULD will already work nicely enough to nudge people
towards using good, machine-readable unit strings, in particular if
that way their users can then write

  table["radial_velocity"]/(0.01*u.pc/u.Gyr)

or something like that.

> specifications (e.g. from talks in last week's interop),
> we might put an issue on both VODataService (which is
> not in github, AFAIK) and TAP to clarify the texts there.
> Similarly to what's up here in VOtable.
>
> Would that help? Is it too much?

Well, I'd say it's a classic for our -Next wiki pages.  And I'm
definitely surprised I've not introduced the VOUnits "should" into the
last VODataService update (where it would have nicely fit).

But yeah, VODataService was probably the last standard published out
of Volute (its cycle started in 2018 when I still had hoped we could
escape github).  I suppose we'll have to touch it anyway (if only for
product-type support), so pulling it over to github would be a nice
touch.  I'll do it some time July-ish, but I'd love to be beaten to
it.

          -- Markus