>
> > No relators are stored in the record. GAP knows how to compute with
> > the group g, which is a Klein four-group as expected, so
> >
> > gap> Comm(g.1, g.2);
> > IdAgWord
> >
> > works, but why are relators not written in the record describing g? I
> > know I can put the relators in explicitly by doing
>
> Each ag word carries a pointer to a data structure <D> created by
> 'AgGroupFpGroup'. The right hand sides of all power-commutator
> relations are stored in this data structure <D>, but in such a way that
> they need less space for non-trivial entries than the relators in
> abstract generators and no (extra) space for trivial entries.
> Creating the presentation in abstract generators would require a vast
> amount of space, 28 bytes for each trivial relator.
>
> best wishes
> Frank
>
> PS: To be precise: power-conjugates relations for single and tuple
> collectors and power-commutator relations for combinatorial collectors
> are stored in <D>.
>
>
But how can I see the relations of an ag group G that I read in from
e.g. the library of solvable groups or the library of 2-groups?
I want to find out how the generators G.1, ... correspond to the ones
in a presentation of G I have in another list of solvable groups,
or to correspond them to another description of G as e.g.
a semidirect product, or some other extension.

Ralf Dentzer

P.S. Is there a function to test for isomorphism between two
ag groups? This would maybe solve my problem, but I
did not find such a function. Moreover this functions
should also provide an isomorphism mapping generators
of one group to as short as possible expressions in
the generators of the second group.

P.P.S What I am doing now is guessing the correspondance and
verifying the relations I know, but this is rather
tedious.