VOUnit for solar density or metallicity?

[Norman Gray]

Markus, Marco and all, hello.

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 machine-readable, but they're not intended to be _only_ machine readable.  I think we say something to that effect in the standard, but the point could be made more strongly, that the units string should or could be usable as a label on a graph, where something like 'SFU' would be preferable to the expansion of it.

That argues in favour of finding some way of supporting discipline-specific units, which is more formal than simply classing them all as 'unknown'.  Something like that has occurred to me.  Markus notes...

On 1 Jun 2022, at 8:48, Markus Demleitner wrote:

> So... I'll need quite a bit more convincing before I could see myself
>  liking the idea of such a shadow unit list; I also note that there
>  already is http://qudt.org/2.1/vocab/unit

The current Unity set of 'known units' (which is precisely aligned with the VOUnits standard set) is supported by QUDT v1: our set is a subset of the QUDT1 units, plus a set of additional ones, including the Jansky, defined using the same framework, and then turned into Java and C code.

QUDT 2 changes the framework significantly, but added a mechanism for collecting units into 'systems of units', including (see <http://qudt.org/2.1/vocab/sou> for gory details):

   "Astronomic System Of Units"
   "CGS System of Units"
   "CGS System of Units - EMU"
   "CGS System of Units ESU"
   "CGS System of Units - Gaussian"
   "Imperial System of Units"
   "Planck System of Units"
   "International System of Units"
   "US Customary Unit System"

I've been tentatively looking at what would be involved in switching Unity to this version of QUDT, and in particular what would be involved in using the same framework to collect the 'VOUnits' set of units.  That would be almost entirely for the library's internal purposes, but I mention it to note that this sort of thing is otherwise supported.

But also (and here I finally get to the point), I think it would be technically feasible to support multiple 'pages' of known units, so that it would be possible to validate a units string with the 'core+solar' list or 'core+radio', and so on.

That wouldn't satisfy Markus's desire to purge 'SFU' from his table metadata, but it would mean that 'SFU' could be included in a specification in a principled way.

> for the furlong/fortnight
>  enthusiasts [oh: fortnight isn't even in QUDT].

Finally -- quell that panicked thought -- be reassured that I have no current plans to support the Imperial or US systems in Unity. *cough*  This is in the air locally, though: In an increasingly desperate attempt to deflect attention from domestic political implosions, the UK government has floated the idea of 'bringing back imperial units', to general hilarity.

Best wishes,

Norman


-- 
Norman Gray  :  https://nxg.me.uk
SUPA School of Physics and Astronomy, University of Glasgow, UK