Roger Hui wrote:
> icol1=: # - +/@(+./\)@(1&=)
>
> See also:
> http://www.jsoftware.com/jwiki/Essays/Collatz_Conjecture
Thanks, Roger!
Perhaps you could also help me on another question from this context. I
have a function f to evaluate on an array a. Since f is expensive, I
evaluate it on the nub and distribute the results. The NUB adverb in the
Dictionary does not work because the arrays are too big, so I use another
verb g which is much faster, but not optimal.
For example:
a=:10 | i.100
f=:*:
g=: (~. i. ]) { f @:~.
(f-:g) i.100
1
I know I ought to be able to do better by only evaluating the nub once,
but I can't see how to do it.
Best wishes,
John
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
