Thank you, David and Raul, for those two solutions. Using David's
approach, the verb I was working on is complete. It lists all subsets of
an input list, a.k.a. all combinations of all possible lengths between zero and
totality.
cc=: 3 : 0
ty=. #y
zz=. <'' NB. same as <"_1 (0 comb ty)
for_h. >:i.ty do. zz=. zz , <"_1 h comb ty end.
zz {&.> <y
)
The verb comb (all combinations) is drawn from the J Dictionary|Control
Structures|For
http://www.jsoftware.com/help/dictionary/cfor.htm
as usual.
The technique Raul proposed only works with literal and numeric input.
It's not a good fit here because I want this verb to receive a list of boxes
and produce results parallel to what it produces with other data-types. (I
don't yet understand the niceties of why this is so.)
If anybody thinks the for.-loop of this verb (cc) would be neatly replaced by
some other structure, I'd be interested in hearing about it.
Tracy
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
