[fitsbits] 64-bit integers

[Preben Grosbol]

On Monday 02 May 2005 22:14, William Pence wrote:
> Given the past controversies about adding 64-bit integer support to FITS,
> it is somewhat remarkable that there has been so little discussion of this
> now. Should this silence be interpreted to mean that everyone here is in
> favor of starting the official approval process on this proposal?  If so,
> then the formal 30-day public comment period could start soon.
I do recognize that 64-bit integer will be needed for pointers and special
data (e.g. time stamps).  My main concern was the overhead it may place
on existing data processing systems which in principle would have to upgrade
their code to accommodate the new data type.  For a system like MIDAS,
one would probably convert I*8 image to R*8 (with a truncation warning)
and never write them.  For table, some work would have to be done on
table data structures and general computation with table column. Also most 
old applications would not be able to handle I*8 columns.

Thus, I'm not hot for 64-bit integer support but it will be needed some
time in the future.  It will probably take significant time before standard
tasks would be able to access such data as only rather special
applications would need it.

Preben Grosbol