DATE-OBS='31/12/99'
Pierre Maxted
pflm at star.maps.susx.ac.uk
Fri Jun 28 14:54:48 EDT 1996
-- Preben Grosbol <pgrosbol at eso.org> wrote:
: : :
> It is very important not to fool old FITS readers. Thus, the first
>8 characters of data value strings for 'DATExxxx' keywords must be
>maintained as 'dd/mm/yy'. I agree with Rob that the double keyword
>solutions e.g. 'CENTxxxx' have a number of disadvantages. Also a whole
>set of new e.g. ISO8601 formated data keywords may be too much.
: : :
If our first priority is indeed to "not to fool old FITS readers" then
there is _no_ way to cope with DATE-OBS='31/12/99' because a two-figure
year is inherently ambiguous once somebody starts writing data after
31/12/99 now that pre 1901 data has been archived as FITS. In this case
the best you can do is to make sure no FITS reader can ever read in a year
and then interpret it incorrectly. A solution I'm sure many people won't
like is to strictly reserve the two-figure year format for the years 1900
to 1999, any other date being written 'DD/YY/XX' e.g. '01/01/XX'. You
might call this the "fail-safe" solution since it provokes an error
whenever there is some ambiguity over the date. I leave the argument over
the required new keyword containing a four-figure year to others.
- Pierre
_-_-_ _-_-_
| Dr Pierre Maxted (pflm at star.maps.susx.ac.uk) |
| | Astronomy Centre, University of Sussex | |
| Falmer, Brighton, BN1 9QH |
-_-_- Procrastinate now!!! -_-_-
More information about the fitsbits
mailing list