[fitsbits] 64-bit integers

[Michal Szymanski]

On 10 Sep 2001 17:07:09 -0400, William Pence <pence at tetra.gsfc.nasa.gov> wrote:
>I would like to reopen the discussion about whether 64-bit integer images
>and 64-bit integer binary table columns should be supported in FITS files. 
>...
>The case for supporting 64-bit integers in FITS has gotten even stronger in
>the 3 years since that previous discussion.  64-bit CPUs and operating
>systems are now standard in the computer industry.  Also, as we enter the
>"Virtual Observatory" era, it is important that FITS be able to support all
>the types of data files that people (both observers and theoreticians) will
>want to store and access.

I fully support the idea. 64-bit integers have been in use for a long
time now, even before 64-bit CPUs had entered the market ("long long int"
extension present in 32-bit versions of GCC).

Taking into account that the 64-bit CPUs are already the reality and
that it will take quite a while before the proposal enters the standard
and another while before is gets adopted into existing software
packages, I think it is HIGH TIME to do it.

regards, Michal.

PS. I know it has also been discussed but maybe it would be worth
thinking about unsigned integers again? :-))
Most (if not all) CCD-s produce data naturally in unsigned format and
converting it to "signed + BZERO" has always been a pain.

-- 
  Michal Szymanski (msz at astrouw.edu.pl)
  Warsaw University Observatory, Warszawa, POLAND