WCS question

[Greg Hennessy]

I am working on producing a set of images on CDROM that will have
astrometry according to the Greisen and Calabretta A&A 1996 paper, but
the four allowed values for RADECSYS are two for FK4, one FK5, and one
geocentric apparent place. The images I will be working on are
reduced against the ACT catalog. Are there any conventions to what
RADECSYS should be for this? I would assume 'HIP' but would like some
advice.