[Nauty] New release of nauty: 2.2b5

[Brendan McKay]

* Sterten at aol.com <Sterten at aol.com> [030511 00:02]:
> Brendan wrote:
> 
>  >nauty22b5.tar.gz, available at http://cs.anu.edu.au/~bdm/nauty,
> 
> I like this directg. 
> To build it with GCC and all my old .o files from the
> last version still present, I had to copy the new naugroup.c,directg.c

That's ok in this case, but in general I can't guarantee correctness if
you use a mixture of old and new files.

>  >Read undirected graphs and orient their edges in all possible ways.
>  >Edges can be oriented in either or both directions.
> 
> you mean:
> all 4 possibilities between 2 different vertices are included,
> so all  irreflexive  relations or binary adjacency matrices with 0 in the 
> main diagonal are allowed.

No, only three possibilities:  x->y, x<-y, x<->y

> For a possible further restriction we can use the -o switch.

In which cases there are two possibilities:  x<-y, x->y
 
>  >    -e#:#  specify a value or range of the number of directed edges
> 
> total number of edges, not edges between two vertices

Right, I changed the wording.
 
>  >    -o     orient each egde in only one direction, never both
> 
> 0 or 1 directed edge between two vertices

No, exactly one directed edge.

>  >    -T  use a simple text output format (nv ne edges) instead of digraph6
> 
> I'd like the adjacency matrix as additional option (just 0s and 1s)
> in one line with an optional blank when the source-vertex changes

The eventual plan is to have a standard digraph format that can be
recognised and converted to many other forms by listg and showg.
 
>  >Tournaments of order 7:
>  >% geng -cq 7 21 | directg -u -o
>  >>Z 1 graphs read from stdin; 456 digraphs generated; 3.10 sec
>  >(This is the worst case for the program and is certainly NOT a good way to
>  >generate tournaments.  If you want some small tournaments you are better
>  >off fetching them from http://cs.anu.edu.au/~bdm/data/digraphs.html.)
> 
> "good" means "fast" here ? 

Yes.

> But it's still useful to have this directg.exe method at hand.
> 
>  >There is currently no method provided for generating strongly-connected
>  >digraphs, but I'm thinking about it.  Nor is there a way to include loops.
> 
> ...yet ? or never will be ?

Depends on demand and time.  Similarly multiple edges.  Both can be
implemented using much the same method.

Brendan.