Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!deimos.cis.ksu.edu!uxc.cso.uiuc.edu!uxc.cso.uiuc.edu!ux1.cso.uiuc.edu!uxe.cso.uiuc.edu!burkie
From: burkie@uxe.cso.uiuc.edu
Newsgroups: comp.lang.c
Subject: Re: Mandlebrot set
Message-ID: <225800193@uxe.cso.uiuc.edu>
Date: 25 Jun 89 23:16:10 GMT
References: <2492@blake.acs.washington.edu>
Lines: 99
Nf-ID: #R:blake.acs.washington.edu:2492:uxe.cso.uiuc.edu:225800193:000:3247
Nf-From: uxe.cso.uiuc.edu!burkie    Jun 25 04:37:00 1989


>double
>func(real, imag)
>double real, imag;
>{
>	int	n;
>	double newreal, newimag, re2, im2;
>
>	for (n = 0; n < MAX_ITER; n++) {
>	 	if ((re2 = real * real) + (im2 = imag * imag) > 4.0)
>			break;	/* point is not in Mandelbrot set */
>
>		/* Xn+1 = Xn^2 + Xn */
>		newreal = re2 - im2 + real;
>		newimag = 2.0 * real * imag + imag;
>		if (newreal == real && newimag == imag) {
>			n = MAX_ITER;
>			break;
>		}
>	}
>	return n;
>}
I think there is something missing in the above code.

-----------------------
The general algorithm is:
given complex number: c 
z = 0 
{ do (z = z ^ 2 + c) until |z| > 2 }    <---	 don't check beyond MAX_ITER
						iterations; assume beyond
						MAX_ITER is for all practical 
						purposes 'stuck forever'
return the number of iterations it took for |z| to exceed 2
and if never exceeded 2 during the MAX_ITER iterations that we tried then 
return MAX_ITER
------------------------
double
func(c_real, c_imag)
double c_real, c_imag;			/* c = (c_real + i*c_imag) */
{
	int	n;			/* iteration count */
	double	z_real, z_imag,		/* z = (z_real + i*z_imag) */
		re2, im2;		/* z_real^2 , z_imag^2 */

	z_real = 0;
	z_real = 0;
	n = 0;
	while ( (n < MAX_ITER) && 
		((re2 = z_real * z_real) + (im2 = z_imag * z_imag) <= 4.0) ) {

		/* iterate z = z^2 + c	*/
		/* = (z_real + i*z_imag)^2 + (c_real + i*c_imag) */
		z_imag = 2.0 * z_real * z_imag + c_imag;
		z_real = re2 - im2 + c_real;
		n++;
	}

	/* if n < MAX_ITER then |z| exceeded 2 after iteration n
	 *		(and 'c' is definitely not in mandelbrot set)
	 * if n = MAX_ITER then we were not able to see it diverge 
	 *		(so we may assume it is part of mandelbrot set)
	 * 		(for lack of knowing any better)
	 */
	return(n);
}
----------------------------
I realize most people already know this but here is a short explanation:

the mandelbrot set is the set of c's in the complex plane which never
diverge when iterated in the manner described above.
(i.e. the z remains finite, oscillating between values, converging
to a value etc. destined never to escape beyond the circle) 

in most pictures of the mandelbrot set - the set is colored black and other
areas are colored according to the number of iterations it took
for the condition |z| > 2 to be reached using the 'c' value at that
point.

the condition |z| > 2 guarantees that future iterations of z = z^2 + c 
will diverge more and more away and so we use it as a convenient condition 
to test for divergence.

however for certain points 'c' it may take a very long time to verify that
|z| eventually exceeds 2 and for some this condition may never become true
(those points which are truly within the mandelbrot set)
So, to avoid waiting forever for this condition, we only test upto MAX_ITER 
iterations and then if |z| still < 2 , we assume for our purposes that the 
z never diverges and give up testing further for divergence.  (and include 'c'
 in the mandelbrot set)

the more detailed your examination of the boundary of the mandelbrot
set, the higher MAX_ITER will have to be to give a good representation.
For a given pixel resolution, MAX_ITER beyond a certain value
will not increase the apparent accuracy of the picture and will only slow 
down the program.


-Salman
burkie@uxe.cso.uiuc.edu