On Tue, Mar 16, 2010 at 10:52 PM, Brian Schott <[email protected]> wrote:
> Yes, your 1+F does seem to produce results identical to Idotr and Ier
> behaves as you suggest. But look at the following experiments which
> suggest that the difference is a little more subtle as suggested by
> the final example. ...
Yes, when you introduce differences less than epsilon, my Ier
misbehaves. (And, for integers, epsilon is 1).
> Do you think the pair of I.'s in F could be made more efficiently
> calculated to require a single calculation? Years ago, I believe it
> was Henry who showed me how such an efficiency can be
> accomplished in simpler verb trains.
The result of I. is used on x and on y, which I think prevents the
kind of expression you were suggesting.
That said, an explicit implementation can deal with this
issue without problem.
F=: 4 :' I - y < (I=.x I.y) { x,_'
That said, J also may in some future version analyze
tacit expressions during parsing, and reuse results where
the same verb gets used on the same arguments?
--
Raul
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
