> You can look at polynomial inverses another way: 

(This is already getting out of my depth.)

> Evaluation and interpolation are inverses.  Here is a simple example.  x 
> L y produces the polynomial with values y at points x using 
> Lagrange interpolation.  

But if you're saying there is another meaningful definition for  p.^:_1  then 
great!  Let's use that.  Or let's discuss our
alternatives.  

I'm really more interested in knowing why an error is better than an 
arbitrary-but-valid result.

-Dan

PS:  If you (or the rest of the Forum) like, we can discuss this off-line.

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

Reply via email to