On Sun, Jul 24, 2022 at 2:39 AM Elijah Stone <elro...@elronnd.net> wrote:
> On Sun, 24 Jul 2022, Raul Miller wrote:
> > space/time tradeoff
>
> There should not be a significant time penalty for blocking, compared with
> fully eager iteration.

For the example of +/ 1 + i.100 ? certainly.

But if that's as far as things go, it would be rather useless, Especially since
   ({{0 0.5 0.5 p. y}} -: {{+/1+i.y}}) 100
1

And you can replace 100 with an arbitrary natural number without
changing that result (at least up to the point where you run into
significant resource limits from the i. implementation.)

But if the language supported streaming as a general case, you would
start running into other issues. And, which of those issues would be
the primary time consuming bottleneck would vary from case to case. I
imagine that there could be some advantages in some cases (like if the
system could dispatch code to multiple cores), but there's going to be
costs and it's not always going to be an advantage.

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

Reply via email to