On Tue, Aug 26, 2008 at 3:17 PM, Brian Schott <[EMAIL PROTECTED]> wrote:
>        I found something fishy by trimming down l to my l1
> I see there is no term that equals 0. Doesn't that seem
> incorrect?
>
>   l1=: - %"1 -/[EMAIL PROTECTED]
>     3.5 l X
> _0.00488281 0.0410156 _0.170898 0.683594 0.512695 _0.0683594 0.00683594

Note that you did not use l1 here -- you used l, which I had defined as
   l=: - */@(-.&__ 0 _)@:%"1 -/[EMAIL PROTECTED]
(and X was i.7).

The bit you removed was */@(-.&__ 0 _), which removes the number which
came from the diagonal of the matrix which l1 would give you, and then finds
their product (these diagonal values will always have had 0 in their denominator
and would be __ if l1's left argument is less than the corresponding value from
l1's right argument, 0 if l1's left argument equals the corresponding value from
l1's right argument, or _ if l1's left argument exceeds the corresponding value
from l1's right argument).

I would expect a 0 on the diagonal only when l1's left argument appears in l1's
right argument.

FYI,

-- 
Raul
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to