On Mon 2008/06/16 07:37:27 MST, Rob Seaman wrote
in a message to: Eric Greisen , Mark Calabretta mcalabre@at
nf.csiro.au
and copied to: FITSbits , William Pence William.Pence@nasa.
gov

Perhaps simply excise "(e.g. as is done for random groups data)"? The
less said about random groups, the better :-)

As much as you optical types would like to get rid of them, random
groups are still valid FITS constructs and still widely used. But
if you can think of a better contrast to the regular structure of
an image then I'd be happy to consider it.

Regards, Mark