[MUD-Dev] Re: AFAP: As fast as possible, non linear...

[Mik Clarke <mikclrk#ibm,net>]

quzah [softhome] wrote:
> 
> Hiya all.
> 
> I'm finishing up a tiny maze generator, and all I need is a
> tad bit of a speed gain. (Or more than a tad. :) I'm using
> a 10x10x10 cube, generated a plane at a time (10x10). I am
> linking them through one exit per plane (up/down), with the
> bottom level's exit leading "out". Here's the basic rundown:
> 
> void mazePunch( char startX, char startY, char startZ, bool nuke){
>    static char path[100];
>    static char used;
> 
>    /* if nuke, zero the path, simple linear for() loop */
>    /* then trash 'used' */
> 
>    /*
>     startX and startY determine this planes 'current' location.
>     {
>     If any neighbor is zero (unused), note it (valid).
>     Pick from the list of valid neighbors at random to use.
>     Knock out the wall, them to us.
>     Increment 'used'.
>     If current == 0, increment 'used'.
>     Knock out the wall, us to them.
>     }
>     {
>     If no neighbors were zero, and we are zero, punch a hole
>     from us to any valid neighbor (valid this time being an
>     in bounds coord.) Increment 'used' since we are no longer
>     zero.
>     }
>     if( used < 100 )
>        recursive call with new coords for x and y.
>     else return
>     */
> }
> 
> That about does it. The only thing I don't like about it is that
> that I have a crappy way of getting the new coords for the next
> recursive call. I'm currently using:
> 
>    mazePunch( number_range(0,9), number_range(0,9), FALSE );

Thinking about it, I'm not convinced that your algorythm will always
generate a solvable maze.  Condiser:

 A-O-O-O O-O
   | | | | |
   O-O-O O-O
   |     | 
   O-O O-O-O
   | | | | |
   O-O O-O-O-B

I think this would meet your loop exit crietria (used = 100), but it is
not possible to get from A to B.  This could pose something of a problem
for your players...

Mik
-- 
http://www.geocities.com/SoHo/Cafe/2260