New RFC (Was: VOUnits: _another_ version, based on implementation feedback)

[Norman Gray]

Dear all,

On 2013 Nov 6, at 11:25, Norman Gray <norman at astro.gla.ac.uk> wrote:

> The DocStd document <http://www.ivoa.net/documents/DocStd/> says (Sect. 2.2) that "Such revisions must be minor in nature, or else the document must return to Working Draft status."  Perhaps the formally correct thing to do, therefore, is reissue this as a WD and put it immediately into RFC.  I hope that most people who would comment on that would have now commented in the post-TCG phase, so I imagine that this RFC would be quiet, and the document could then be immediately and automatically promoted to PR and thus TCG review.

I have (finally!) updated the document, and the associated grammars, to match the discussion on-list (or at least I think I have -- please check).

Mireille created a new RFC page (thanks, Mireille!), which is pointed to in the text below, just above my signature.

I've uploaded the document to the IVOA document registry but, given the date, I doubt this can, or indeed should, be attended to before the New Year.

I've suggested an end-date of 2014 January 31 for the RFC.

Thanks, everyone, for your patience.  We're on the home stretch now.

Very best wishes,

Norman
[shutdown -h +10]

----

A few comments on the changes I've made:

I've unilaterally added 'au' as a permissible alternative to AU, because it
seems perverse to prefer atto-atomic-units to astronomical units.

I've deleted the reference to \unit{Sun}, because there's been no discussion of
how to actually use this, and it's not obvious what the syntax should
be; it's not supported in the Unity library.

The Bessellian and Tropical years ('Ba' and 'ta') remain as FITS
units, but they remain 'unknown units' within the VOUnits spec.

The text now states more prominently that `unknown' (lower-case) is
the single indication that a column has unknown units, and that an
empty string positively indicates that a column is dimensionless.
Both must be checked for before unit-parsing starts, since both
produce invalid unit-strings.  My reservations about these two points are:

  * I don't like case insensitivity (agreeing with Markus), but there
    seems a case for it here, since I can imagine someone trying to
    remember whether the magic string was "unknown" or "UNKNOWN",
    remembering the wrong way, and ending up with a correctly
    parseable (though unknown) unit called the 'UNKNOWN'.

  * The problem with an empty 'dimensionless' string is that it's not
    clear that someone has deliberately stated "this quantity is
    dimensionless", as opposed to "I forgot to put in the units", or
    "I couldn't be bothered", or "I don't know what the units are and
    didn't know I was supposed to write 'unknown'", all of which are
    plausible interpretations of "", if one were to find it in a file.



----
This is part of a discussion of "Units in the VO"
IVOA Proposed Recommendation 1.0-20131224
Editor(s): Sébastien Derrière and Norman Gray

PDF: http://wiki.ivoa.net/internal/IVOA/Dec2013VOUnitsRFC/PR-VOUnits-1.0-20131224.pdf
changes: https://code.google.com/p/volute/source/browse/trunk/projects/std-vounits/VOUnits.tex
changes: https://bitbucket.org/nxg/unity/commits/all
RFC: http://wiki.ivoa.net/twiki/bin/view/IVOA/Dec2013VOUnitsRFC

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