Can I suggest a minor improvement to Normalizer, especially for permutation
groups. I recently did

Normalizer(g,h)

where h was, in fact a subgroup of g, although actually represented as a
parent group in its own right. This ended up in GroupOps.Normalizer and showed
no sign of returning (g was S10, h was D10 in regular representation).

on the other hand doing

h := AsSubgroup(g,h);
Normalizer(g,h);

produced the required result reasonably quickly.

Perhaps this situation could be checked for somewhere and the quicker algorithm
used in that case. Even better, h & g could be embedded in the group they both
generate.

Steve