VOUnit for solar density or metallicity?

[Markus Demleitner]

Hi,

On Mon, Jun 06, 2022 at 09:21:29AM +0100, Mark Taylor wrote:
> On Wed, 1 Jun 2022, Norman Gray wrote:
> > I don't have a strong position on the question of using eg SFUs
> > in preference to '10**-22W.m**-2.Hz-1', but I do recall that,
> > during the initial discussion of the VOUnits spec, the point was
> > made (in that case in the context of jupiterMass) that 'SFU' has
> > at least some value as documentation, which the equivalent in SI
> > base units doesn't.  The VOUnits strings are intended to be

Hm... I'm deeply unconvinced that SFU has an extra documentation
value, as, for all I know, it is defined strictly in terms of SI
units -- and that's a difference against Jupiter or solar mass, ideas
of which (and, on the long run, even the values of which) change.

Me, I'd try to keep abbreviations purely made for convenience (which
of course includes Angstrom, but that ship has sailed) out of the
mandatory core of VOUnits, where our "upstream" standards don't force
us to accept them; and since optional features are always a pain,
that would mean keeping them out entirely.

> I agree with this in principle.  In practice much the most common use 
> of VOUnits (strings appearing in places where VOUnits are required 
> or recommended) will be as human-readable text rather than 
> machine-parseable tokens, so my feeling is that VOUnits should err 
> on the side of inclusivity and attempt to accommodate unit 
> representations that people actually want to use which are well 
> defined and not obviously nasty - so yes to SFU, no to Sun.

Hm -- with many VOTables being consumed by astropy, which, I think,
tries to build quanitites with the units, I'm not so sure the
assessment of "most common use" is still right.  And whether or not
I'm wrong: The reason we're writing a standard is to make it *easy* for
machines; for humans we probably wouldn't need any standard at all.
SFU alone wouldn't really hurt.  Having dozens of such convenience
translations at some point will.

       -- Markus