[Aldor-l] exporting a domain in aldor - update

[Martin Rubey]

Dear Aldor experts,

Ralf Hemmecke <ralf at hemmecke.de> writes:

> On 01/01/2007 08:21 PM, Martin Rubey wrote:

> > #include "aldor"
> > define Cat: Category == with {
> >     Dom: OutputType;
> >     f: () -> Dom;
> > }
> > DomA: Cat == add {
> >     Dom: OutputType == Integer;
> >     f(): Dom == 1783; }
> > DomB: Cat == add {
> >     Dom: OutputType == String;
> >     f(): Dom == "abcd"; }

I just checked: replacing the definitition of f() with

    f(): Dom == 1783 pretend Dom; 

does work in compiled code. I wonder why it segfaults in the
interpreter. Furthermore, I really need another solution, to make the code
available in axiom - as you might know, in axiom types are not first class
objects...

> Of course, the above code does not represent what you have in mind.

Why not?

> What about explaining your actual problem instead of translating your problem
> (wrongly) into Aldor and claiming it were the problem? What do you want to
> achieve?

The full problem is difficult to explain without going into the details of
aldor-combinat. I'll try -- simplifying some things:

We have a category

define CombinatorialSpecies(L: LabelType): Category == with {
    ...
}

that exports an operation (simplifying here a little)

structures: Set L -> Generator %;

One example would be the domain SetPartition, for which structures produces all
set partitions of the given list.

Now I want to provide a second operation

isomorphismTypes: MultiSet L -> Generator ???;

that produces all multipartitions, i.e., multiset partitions of the given
multiset. For example, for m = [1,1,2] the multipartitions are

[[1,1,2]],[[1,1],[2]],[[1,2],[1]],[[1],[1],[2]].

Of course, these are not of type SetPartition. They are, in a certain sense,
isomorphismtypes of the setpartitions of [0,1,2], where 0 and 1 are
identified.

For a reason I cannot explain here, I would like to have the operation
isomorphismTypes provided by the same domain, i.e., isomorphismTypes should be
exported by CombinatorialSpecies.

Does this explain things better?

Martin