[Nauty] how many nauty-users ?

[Brendan McKay]

* Sterten at aol.com <Sterten at aol.com> [021121 20:16]:
> 
> how many people on earth have installed nauty ?
> how many subscribers are there for this mailing list ?
> 
> my vague guess is 10000 and 20

The mailing list was only set up recently so there are not
many members yet.  Probably it will grow to a hundred or
two eventually.  The total user base is impossible to
determine, but I think that a few thousands is plausible.

I forget the exact reason I chose "dreadnaut".  It is about
20 years since the first version.

> One additional request would be to get the whole automorphismgroup.

The program nautyex3.c demonstrates how to do that.  Groups can
be exceedingly large, so this is rarely useful.

> Some canonical labeling of the group-isomorphism class
> would be excellent.
> I assume, we can construct a graph G(H) for each group H,
> such that H1 and H2 are isomorphic, iff G(H1) and G(H2)
> are isomorphic. And then use labelg on G(H).

This is possible in principle, but it isn't really the right
way to go about group isomorphism.  In the case of isomorphism
of permutation groups, there are programs which test conjugacy
that are the fastest.  Magma has such a program, I believe.

> does everything what you did with graphs generalize to digraphs
> or relations and could be done in principle if you only found
> the time to implement it ?

It is only a matter of free time.

Brendan.