[fitsbits] 64-bit integers

[William Pence]

Arnold Rots wrote:
> 
> If we are considering tinkering with the data types, at last, should
> we consider including old unfinished business and add the unsigned
> integers as well? 

No, let's not confuse the current poll with more choices.  The arguments for
supporting 64-bit integers are quite different from the case for supporting
unsigned 16-bit and 32-bit integers.  FITS already supports unsigned
integers by casting them into signed integers with an offset, so adding a
new unsigned integer datatype does not provide any new basic capability to
FITS.  It simply provides an alternate (duplicate) way of representing the
same information.  Adding 64-bit integers, on the other hand, adds new
capabilities to FITS.

So the choices in the poll remain:

 1  YES - I think 64-bit integers should be added to FITS
 2  NO  - I do not think 64-bit integers should be added to FITS, at
          least not at this time.
 3  UNDECIDED - I have not yet made up my mind
 4  NO OPINION - I don't care one way or the other

Remember to send your vote to pence at tetra.gsfc.nasa.gov, not to this
newsgroup.

Rob Seaman wrote:
> 
> It also isn't clear (to me, at least) precisely what proposals are being
> discussed.  It's hard to have an opinion without at least a straw proposal.

The current proposal would allow 64-bit twos-complement signed binary
integers, contained in 8 bytes, to be represented in FITS images and binary
table columns.  The images would have BITPIX = 64, and the table columns
would have TFORMn = 'K'.

-Bill Pence
-- 
____________________________________________________________________
Dr. William Pence                          pence at tetra.gsfc.nasa.gov
NASA/GSFC Code 662         HEASARC         +1-301-286-4599 (voice)     
Greenbelt MD 20771                         +1-301-286-1684 (fax)