I finally got version 3.2. Playing with the new features, I tried
gap> g:=ThreeGroup(243,1);;
gap> t:=CharTablePGroup(g);;

It took just about an hour to construct the character table. This is a
lot more than I am used to for groups of this size (roughly) from
version 3.1. However, I was appalled when I looked more closely and
discovered this group is the cyclic group of order 243. An hour to find
the character table of a cyclic group? (This is actual running time,
according to ps.)

Is there some flaw in the method being used here? I am a bit worried
that there might be similar performance problems if a group happens to
have a large cyclic subgroup or quotient group in the wrong place.

Note this is on a clunky old Sparcstation 1+, no speed demon, using the
binary I got from wuarchive.wustl.edu. I haven't compiled GAP 3.2
locally yet, but I intend to.

Character tables of other groups of order 243 seem to be constructed
much faster (minute or two). So far the cyclic one is the only one
like this I have found.

David Sibley NT3O
sibley@math.psu.edu