The purpose of this note is to warn you of a serious bug in the Size
function which has been introduced in GAP 3.3.

Because of a possible overflow in an uncontrolled variable of some C
routine in the GAP kernel the Size function may return a wrong value
for a finitely presented group (without any warning or evidence).
Eamonn O'Brien from Canberra has sent us an example.

We have no reason to believe that the Size function is faulty for any
other kind of group.

I have unfortunately introduced this bug when we tried to speed up the
Size function for finitely presented groups and to extend the range
of its applicability in version 3 release 3 of GAP. To the best of our
knowledge, the Size function has been correct also for finitely presented
groups for the releases of GAP preceding release 3 of version 3.

The bug will be fixed in the next patch or release. For the time being
you can use the the function

Index( G, TrivialSubgroup( G ) );

which will practically do what the Size function for finitely presented
groups did in version 3.3 (and hence be as efficient as that function
was).

Volkmar Felsch, Aachen