June Kim wrote:
> Probably the simplest approach is make a 1000x1000 table and
> fill in the boxes with 1s and count 1s. Or calculate the size
> of the set of the coordinates of covered integer points(x,y).

I think this is the right solution (not the 1000x1000 grid,
but the grid based on enumerated coordinates).

> Is this a conventional way of handling geometric problems
> (points, rectangles, ...) in J? Any place to improve?

I don't know about how conventional this is or is not.
It's conventional in the sense that it reduces potential
complexity to a relatively simple form.

As for potential improvements -- I have a few rough
ideas, but they need a bit more thought before I can
tell if they're worth talking about.

-- 
Raul

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

Reply via email to