Robert Bernecky wrote:

> I am not sure how much of APL or J could be extended to
> infinite arrays, in the sense of avoiding the need for
> array expansion often enough to be useful. That is, do
> the benefits warrant the effort required to implement it?
> Or do we end up with a toy with just enough power to be
> irritating?

Actually, I was not proposing a change to J.  I was asking how we can
use J as a notational model when we wish to discuss mathematics.  K. E.
Iverson wrote several texts where APL or J were the core notation.  I'm
reaching toward getting good at using this sort of notation as I study
math.  Normally, of course, the things I encounter aren't already put in
those terms.

In the case at hand, John Randall advised that the comparison should
involve equivalence classes, not distinguished representatives of those
classes.  Representatives are scalars but equivalence classes are
infinite sets, so I probably can't render the distinction without some
departure from executable J.  That's fine.  I'm just looking to share
ideas as to how to handle such departures when we choose to use some
manner of Iverson notation on a blackboard or whiteboard to play with
statements, theorems, or proofs.  This need not imply any pressure on J,
the executable notation, to accommodate that use.

Tracy

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

Reply via email to