First an assertion of facts:
(A) The items of an array have dimension one less than the array
itself. So for an array with shape m n p, the items will have
shape n p
Now, my point of contention.
RE: http://www.jsoftware.com/help/jforc/loopless_code_i_verbs_have_r.htm
We read:
(B) Think of the operand as an array with shape f whose items are cells
of rank r.
The text then goes on to say:
+: i. 2 2
0 2
4 6
(C) The steps to get this result are:
The verb rank is 0 and the noun rank is 2, so we will be applying
the verb to 0-cells. The frame f is 2 2
So my problem is this:
If I think of the operand as an array with shape f. Then per (B) I
would think of the operand as an array with shape 2 2 who items
have rank 0. But (A) implies the items of an array with shape 2 2
will be rank-1 with shape 2.
So my suggestion is:
Reword (B) to say:
Think of the operand as an array with shape f whose innermost
dimension contains cells of rank r.
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
