On Sun, Jul 3, 2022 at 4:44 AM Elijah Stone <elro...@elronnd.net> wrote: > I am not sure how to feel about this. Mainly, I am not sure what the > difference is between x m}y and (for high-ranked numeric m) x (<"1 m)} y. I > am also not sure what the difference is between your proposed semantics and > the existing ones.
After thinking about this a bit, ... perhaps this rephrasing will make sense: Before J901, m (in x m} y) indexed items of y (and position in m had to conform with x). My thinking here was that the leading column of a rank 2 m should still index items of a high ranked y. In the variant introduced in J901, that's only the case when the last dimension of m equals the rank of y. There's definitely some other issues to consider, involving conformance, but I think that analogy with the J8 behavior of amend should serve, there. -- Raul ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
