Dear all,

I'm sick, so just a short update and reply.

Update: I modified most of species.as.nw to deal with "generalized"
isomorphismtypes as announced and maybe better explained by Ralf (I don't have
the energy to read his mail really, though.

I have algorithms for Compose, Derive, Point for labelled, unlabelled and
everything inbetween.

I don't have a clue how to do Cartesian products and Functorial Composition,
except for the labelled case or in special cases. (for example, pointing is a
special case of the cartesian product)

> But still, the problem is finite and thus solvable in principle by applying
> all permutations from S_n and building cosets.

I don't think this is entirely true. What we need for a "stupid" algorithm to
do the unlabelled case - not the cases "inbetween" - is a function in, say,
CartesianProduct

coerce: % -> IsomorphismType

that yields the isomorphism Type of a structure. Please show me how to do that
for the Cartesian Product! (This would be the equivalent to having a normal
form)

An alternative possibility is to generate the structures, and for each
generated structure check whether it is isomorphic to one of the structures
already generated.

But really, I think the best would be to find a way to cast the problem into
the problem of generating all nonisomorphic graphs or designs and use nauty to
generate them. However, it is very unlikely that this is possible at all.

I guess, generating the isomorphismtypes of an arbitrary cartesian product is a
real challenge.

Martin


-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Aldor-combinat-devel mailing list
Aldor-combinat-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/aldor-combinat-devel

Reply via email to