[GAP Forum] The stability of group libraries

[Alexander Hulpke]

Dear Peter, Dear Forum,

> Just a brief addendum to Alexander's comment.
> 
> Ákos Seress, Primitive groups with no regular orbits on the set of
> subsets, Bull. London Math. Soc. 29 (1997), 697-704, identifies primitive
> groups by their (old) GAP numbers. It is now a bit of work to match them
> with the new numbers.
> 
> In cases like this where the numbering changes, for whatever reason, I
> think it is important to have a mechanism to match old numbers to new.

I could not sympathize more with this. The problem with the primitive groups library before 4.2 was that in some cases (basically if the socle is not simple), the arrangement of the primitive groups depended on a subgroup lattice computation (the intention had been to save storage), which in turn might have depended on the setting of the random number generator. That is, in two different sessions of the same release, `PrimitiveGroup' might have given different (nonisomorphic)  groups for the same number.

This problem was the reason for changing the library. While in many cases the group numbering never changed (the intention was to make the numbering stable, not to relabel for the sake of it), it is hard to specify what did not change or to give a translation list.

The only range of degrees, for which I dare to make a promise about what changed or give a translation list, is  the groups of degree up to 50. These always agreed with the lists by Sims, as well as the published version by Buekenhout/Leemans. (This incidentally includes all groups that occur in Ákos' paper you refer to.)

All the best,

  Alexander