[Home] [Groups] - Message: [Prev in Group] [Next in Group]
11629: Re: [MUD-Dev] Spatial datastructures and their application (Was Re: Hilbert Curves)
[Full Header] [Plain Text]
From: Miroslav Silovic <silovic@zesoi.fer.hr>
Newsgroups: nu.kanga.list.mud-dev
Date: 12 Nov 1999 11:09:55 +0100
References: [1]
Organization: Kanga.Nu
"Brandon J. Rickman" <ashes@pc4.zennet.com> writes:
> Your raycasting is going to bog down after a few tens of thousands
> of checks, though, so if that happens you'd want to limit the
> processing to a smaller subset of neighborhoods, &c.
Raycasting is done in time proportional to n^(1/D), where n is the
number of objects you have to work with before you hit the first wall
and D is the number of dimensions (2 or 3).
So 1,000,000 objects in space is only 10 times harder than 1,000
objects. Compare with the fact that object position updates are linear
in number of objects. So the slowdown is not much of a deal.
On another issue, if I were in a room with an armed human and 1,000
small herbivores, I'd pay attention to the human... unless you're
talking about Quest for the Holy Grail rabbits.
--
How to eff the ineffable?
_______________________________________________
MUD-Dev maillist - MUD-Dev@kanga.nu
http://www.kanga.nu/lists/listinfo/mud-dev