[Units]

[Douglas Tody]

Hi -

It appears we are talking about two different things here: standard
unit representations vs restrictions on the units permitted by an
individual protocol.

Certainly how units are represented should be an IVOA-wide standard
so far as possible; we should also try to be compatible with FITS
usage, perhaps as a superset.  The different protocols should all use
the same units representations.  The main purpose of an IVOA Units
standard should be to define how we specify units.

However, the actual set of units permitted by a protocol ("restriction"
of units) should be left to the individual protocol to define.
While some consistency is desirable and possible, in general the most
natural choice of units is very problem-specific, and may even differ
upon input and output.

 	- Doug


On Thu, 28 May 2009, Christophe Arviset wrote:

> Dear all
>
> I agree with the two points as stated by Bob. That would mean that the UNITS 
> doc will probably need more updates than other IVOA standards, but that is 
> natural.
>
> Cheers
>
> Christophe
>
> Robert Hanisch wrote:
>> ---- Original message ----
>> 
>>> Date: Thu, 28 May 2009 13:13:10 +0200
>>> From: Francois Ochsenbein <francois at vizier.u-strasbg.fr>  Subject: Re: 
>>> [Units]   To: dal at ivoa.net, dm at ivoa.net
>>> 
>>> 
>>> 
>>> Hi Mireille,
>>> 
>>> About your questions, in reverse order:
>>> 
>>> 2. The document on Units shoud describe the generic units, meaning that
>>>   something called 'unit' supplied by any of the VO protocols would
>>>   have to be compatible with the document on units;
>>> 
>> I agree with this.  The units document should reflect common usage in the 
>> community, as end users will encounter data from many sources, most often 
>> in FITS format with FITS-convention units.
>>
>> 
>>> 1. Individual protocols may restrict the units they accept to a subset
>>>   of the generic units -- and then logically the list of acceptable
>>>   units for a given protocol (say simDAP) should be specified in a
>>>   section of the protocol document (simDAP).
>>> 
>> I would suggest just the opposite approach, i.e, to define a core, 
>> restricted set of unit representations that the various VO protocols are 
>> expected to incorporate.  I think this approach maximizes interoperability. 
>> Should a protocol require units definitions not included in the core, that 
>> protocol could define an ancillary set (expressed, I would hope, following 
>> the same rules as for the core set).  If these additional units are of 
>> common value they could be added to the units document.  If they are 
>> peculiar to the protocol, their definitions could stay there.
>> 
>> Bob
>> 
>> - - - - -
>>
>> 
>>> It becomes a be a problem if one of the protocols requires a unit
>>> which is incompatible with the document on Units ; if it turns out
>>> that very specific units have to be created for some services (I'm 
>>> thinking e.g. to simulation data) the document on Units
>>> should propose a way of creating and referring to new units
>>> which do not collide with existing ones (what I called
>>> 'user-defined units')
>>> 
>>> Francois
>>>
>>> 
>>>> Hi all,
>>>> In the DM1 session, we had a discussion (too short) about Units,
>>>> following the suggestion made by A. Micol on these same lists.
>>>> The questions are :
>>>> 1)should the IVOA protocols allow only a subset of the IVOA supported
>>>> Units expressions ?
>>>> If yes , who would volunteer to list up the wished units for each 
>>>> protocol?
>>>> SIAv1,
>>>> SIAv2
>>>> SSA
>>>> SimDAP
>>>> SLAP
>>>> others?
>>>> 
>>>> 2) should these supported units sets published in protocols documents
>>>> or in the Units draft?
>>>> 
>>>> thanks for your comments / suggestions .
>>>> Mireille, DM chair
>>>> 
>>> =======================================================================
>>> Francois Ochsenbein    ------   Observatoire Astronomique de Strasbourg
>>>   11, rue de l'Universite 67000 STRASBOURG  Phone: +33-(0)390 24 24 29
>>> Email: francois at astro.u-strasbg.fr (France)    Fax: +33-(0)390 24 24 17
>>> =======================================================================
>>> 
>
> -- 
> Thanks in advance
>
> Cheers
>
> Christophe
>
>
> -------------------------------------------------------------------
> Christophe ARVISET                       Christophe.Arviset at esa.int
>
> European Space Agency (ESA)
> European Space Astronomy Centre (ESAC)
> Science Operations Department
> Science Archives and Computer Support Engineering Unit
>
> P.O. Box 78
> 28691 Villanueva de la Canada                 Tel: +34 91 813 12 78
> Madrid - SPAIN                                Fax: +34 91 813 13 08
> -------------------------------------------------------------------