Henry clarified:
> n is any noun, not just an atom.
Well, I would say n is an atom if y is a vector, and in general n has the
shape of a item of y . I don't want to get into
the discussion of cutting on lengthy frets like CRLF (that would complicate
its definition, and therefore its already slim
chances of adoption).
The question is whether we want to "undo" the append. That is, do we want the
inverse to be as I originally proposed:
> <;._2!.n <==> <;._2@:(,&n) :. ([: ; ,&n&.>)
Or do we want to curtail the result, as in:
<;._2!.n <==> <;._2@:(,&n) :. ([: }:@:; ,&n&.>)
I note that this latter is much more commonly desired, but also that while the
former is easy translateable to the latter, the
reverse is not true (without redundant logic).
-Dan
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
