[fitsbits] Question on angstrom and erg units in FITS standard

[Erik Bray]

On 11/04/2014 03:23 PM, Demitri Muna wrote:
>> 2. Angstrom and erg are /deprecated in the IAU list/
>
> Deprecated means (to me) that you should not be using feature X, but if you do
> (and you can) /don't be surprised if in the future if your file that uses
> feature X no longer works (which in this case could mean the file can be read
> but will not be written out the same way as above).
>
> That said, I don't see why an updated FITS format can't attach any unit (e.g.
> anything defined in astropy.units) to any column (barring ambiguity of course).
>

I agree with Demitri on this point though.  There's no reason, say, data in 
Angstroms should be forbidden from being written a FITS file.  There's no 
technical reason it can't.  And for any (almost, modulo ambiguities) any unit it 
should be usable as long as there's some syntax that allows a machine to 
recognize those units.  VOUnit gives a pretty good one.

I think the FITS standard ought to be clarified in such a way that it suggests 
certain units for quantities of certain dimensions, but does not arbitrarily bar 
the use of other recognizable units.