David Lambert wrote:
> I wrote the rosetta code red-black tree entry.
I haven't read/reviewed the Red-Black tree implementation on RosettaCode,
but it did remind me of an argument I once made for supplying I. with an
inverse [1]:
> I.-space is useful, particularly when the domain (input & output)
> is boolean, but the translation is most easily expressed as arithmetic
> on indicies. I've wanted this to locate the children in a binary tree
> stored as a vector, for example:
>
> assert 'r' = red_black_tree #~ (# {. ([: , 1 2 +/~ +:)&. I) 'b' =
> red_black_tree=.'brrbbbbrrr'
This is an example of the kind of accomodations I believe we'd have to
build into the language to make working with trees (again, as trees, not
as a type of orthotopic array) simpler and more elegant.
-Dan
[1] "Proposed inverse for monad I.":
http://www.jsoftware.com/pipermail/general/2006-April/026876.html
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
