[DALI] timestamp format

[Rob Seaman]

Alternately for this feature we could ditch ISO-8601 and go with CCSDS 301.0-B-4:

	http://public.ccsds.org/publications/archive/301x0b4.pdf

which assumes UTC for similar reasons as in astronomy and for which the Z is optional.  In this case I believe the hyphen and colon are not optional, even if otherwise omitting fields, e.g.: -11-06T11:10: (or something like that).

The point being the IVOA recommendations have as much validity in our community as standards in other communities.  Backwards compatibility with FITS and STC seems sufficient to not reopen this issue.  Perhaps just make a "Z" optional and move on?
--

On Nov 7, 2013, at 10:56 AM, Rob Seaman <seaman at noao.edu> wrote:

> The 1988 standard is available here:
> 
> 	http://metric1.org/8601.pdf
> 
> If a statement "local in astronomy means UTC" isn't sufficient, somebody may want to buy the current standard (I have 1991 somewhere) to make sure we're arguing from an up-to-date knowledge of the normative standard.  Implicit in Steve's and Arnold's comments is a lot of recent evolution of civil timekeeping, in particular of timezone usage.
> 
> Rob
> --
> 
> On Nov 7, 2013, at 10:29 AM, Patrick Dowler <patrick.dowler at nrc-cnrc.gc.ca> wrote:
> 
>> On 11/06/2013 12:09 PM, Steve Allen wrote:
>>> If I recall correctly the discussions leading to the FITS Y2K
>>> agreement explicitly wanted to prevent the complexities of
>>> timezone interpretation from entering the FITS standard.
>>> The first choice is the one that was made, specifically noting
>>> that it is only a subset of ISO8601.
>> 
>> Not sure I follow. I agree with the goal of removing tz complexity from the situation I do not understand* why one would not just include the Z (always) and thus actually comply to the ISO8601 spec. To not include the Z, one cannot say this is a subset of ISO861, where that spec says subset (profile) means "one of the allowed forms, with the same meaning". What we seem to have is one of the allowed forms, with a different meaning. At least that's how I read the meaning/intent of ISO8601...
>> 
>> 
>> * aside from backwards compatibility with FITS usage
>> 
>> -- 
>> 
>> Patrick Dowler
>> Canadian Astronomy Data Centre
>> National Research Council Canada
>> 5071 West Saanich Road
>> Victoria, BC V9A 2L9
>> 
>> 250-363-0044 (office) 250-363-0045 (fax)
>