I have not thought much about it, but it might be nice to allow for
multisets = unordered lists, too...
> (1) nset.lisp is a new set package that Stavros and I have been working
> on. It's main
> difference with the old package is that it treats lists and sets as
> distinct objects:
>
> (C6) s1 : set(a,[a,b],set(a,b));
> (D6) {a, {a, b}, [a, b]}
> (C7) adjoin(set(a),s1);
> (D7) {a, {a}, {a, b}, [a, b]}
>
> My favorite function in the package finds equivalence classes:
>
> (C8) equiv_classes(set(1,2,3,4,5,6),lambda([x,y],is(remainder(x-y,3)=0)));
> (D8) {{1, 4}, {2, 5}, {3, 6}}
> (C9)
this sounds very nice.
> (2) The file names in test.lisp are hardwired. Could we change that?
> Maybe
> test.lisp should look through all directories under ../share for files
> with
> names of the form test*.mac.
I do not have a file test.lisp. BUT: beware that there may be multiple
files of the same name in share/*/*. EG permut.lisp exists twice:
./combinatorics/permut.lisp
./sym/permut.lisp
A way out is to make autoload search only in share. sym/permut.lisp finds
that, combinatorics/permut.lisp the other. Fail otherwise. I think there
was some discussion on this ealier on...
Martin