Any computer (today) has got to process data in register-sized chunks. If the register size varies across machines, you expect your language processor to handle the migration without the need to change your code.
I wouldn't call it "abstracting from the programmer" if a given language processor accepts fancier data items than ordered sets of precisely 8 bits or 16 bits in its "variables". Even the original BASIC lets you process data items longer than the machine register (up to 255 I recall), as in: 100 LET A$ = A$ + "--this soon gets too long to fit in one register" But you remain free to model your strings as arrays of characters -- and perform operations on them char-by-char. BASIC has merely relaxed that particular restriction. J is simply more relaxed still, restricting you to (its notion of) an "array". And what is an acceptable "element" of an "array" is J's choice, not yours. (Or, for that matter, the machine's.) On Sun, Jul 10, 2011 at 8:02 PM, David Vaughan <[email protected]> wrote: > How can J perform operations on entire arrays without doing it to each > element sequentially? Or does it in fact do this, but abstracts it from the > programmer? > > ___________________________ > > David Vaughan > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm > ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
