Aai,
Yes, it seems to work very well and quickly.
I will study it some more now. Apparently I. is unnecessary. Thanks very much.
On Mon, Apr 27, 2009 at 10:21 AM, Aai <brady...@xs4all.nl> wrote:
> May be it's not really the type of solution you have in mind, but I'll
> give it a try.
>
> If the domain of the args is i.n , then this is more efficient for:
>
> LHA with 'large' gaps, and 10 ( much < ) # RHA
>
> trap17=: 4 : 0
> 'u h d' =. 2 -~/\ x
> 'lz dr' =. 0 >. x (-&{.,-~&{.) y
> (#y){. dr}. (lz$0), ((%*i.)u), (1$~h+1), (%...@-)d
> )
>
> 200 400 480 600 (trap-:trap17) i.1e3
> 1
>
--
(B=)
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
