On Jan 7, 2009, at 8:52 AM, Tim Jenness wrote:
>> OK to extend starlink.version like this, or will that upset CADC
>> (seem to remember the svnversion tag was added for them).
>
> That should be fine. I don't think it was ever stated that the
> software version we write into survey data products has to be a
> number (if we switched to git it would be a md5 checksum).
The ACSIS FITS header spreadsheet we use to coordinate these things
with CADC says that the pipeline and algorithm engine versions are to
be integers. I'll bug Russell and find out if this can be easily
changed to a string. Shouldn't be too hard given we've never properly
used the algorithm engine version header, as far as I know.
Brad.
|