I wrote:
> it is a non-trivial example of a true, natural dyadic fork.
> it continues your [statistical] theme
> it has some interesting features ... [which] might make
> the example too complex for your needs.
Actually, you could retain the former two advantages without the drawback
associated with the latter, by using the standard library definition of
corr instead:
load 'stats'
corr
cov % *&stddev
-Dan
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
