> < ^ Date: Mon, 15 Nov 1993 14:32:01 +0100
> < ^ From: Martin Schoenert <martin.schoenert@math.rwth-aachen.de >
< ^ Subject: Re: Size(g)
Dima Pasechnik writes in his e-mail message of 1993/11/15

To the best of my knowledge, the following by-product of the use of ^C
while computing strong generators for a permutation group should not
occur.

Let G be a permutation group defined by a list of generators, and let the
first function called after the definition of G be Size(G). I press ^C^D
before it finishes, then call Size(G) again. It returns an answer
instantly, since G.size is already defined. Moreover, sometimes G.size is
wrong in this case. I beleive that G.size should actually be undefined.
At least I don't know of any place in the GAP manual mentioning such
effects.

This is a know bug (calling is an unwanted by-product is very nice ;-).

The problem is that as soon as GAP starts to compute a stabilizer chain
it enters this in the group record. When the computation of the
stabilizer chain is interrupted the group record thus contains an
unfinished and sometimes incomplete stabilizer chain. This is a bug.
The stabilizer chain should not be entered in the group record until it
is complete.

'Size' only sees that the group records has a stabilizer chain and
returns the product of the indices. This is correct but only under the
assumption that the stabilizer chain is complete. Other functions for
permutation groups will also fail with such an incomplete stabilizer
chain (and their failure may be more difficult to detect).

Fixing this problem is planned for the next release of GAP. I am sorry
but I cannot offer an immediate solution or workaround. Just be careful
and keep a healthy distrust for results computed by GAP.

Martin.

-- .- .-. - .. -.  .-.. --- ...- . ...  .- -. -. .. -.- .-
Martin Sch"onert,   Martin.Schoenert@Math.RWTH-Aachen.DE,   +49 241 804551
Lehrstuhl D f"ur Mathematik, Templergraben 64, RWTH, 52056 Aachen, Germany

> < [top]