Welcome to Fractal Forums

Fractal Software => Programming => Topic started by: Duncan C on February 13, 2007, 06:51:04 PM




Title: Boundary following for mandelbrot and Julia sets - John Bailey?
Post by: Duncan C on February 13, 2007, 06:51:04 PM
I found a very cool link on the net to a Java applet that plots Mandelbrot and Julia sets (http://www.ibiblio.org/e-notes/MSet/BigM.htm) using a boundary following algorithm to find and fill contiguous areas with the same iteration values. (click the link to see it.)

This code creates plots INCREDIBLY fast. The Java applet is much faster than my carefully factored C code, even though Java is interpreted and has a lot more overhead. I'm impressed, and would like to implement this in my code.

The code was written by a guy named John Bailey, but the links on the page, to his website and to his email, have gotten stale.

I have some questions for him about his code. Does anybody here know him, and/or how to contact him? I found a few different addresses for him, but none are valid.

Faling that, have any of you implemented boundary following routines for plotting mandelbrot and Julia set neighborhoods? If so, would you mind answering some quesitons?

Duncan C


Title: Re: Boundary following for mandelbrot and Julia sets - John Bailey?
Post by: Nahee_Enterprises on February 14, 2007, 12:23:49 AM
Duncan C wrote:
>
>    The code was written by a guy named John Bailey,
>    but the links on the page, to his website and to his email,
>    have gotten stale.   ....    Does anybody here know him,
>    and/or how to contact him?

If you are talking by John M. Bailey, then he may be found in several ways.  First of all, did you try the "Fractal Census (http://www.Nahee.com/Census/)" ??  At least one of those web sites is valid, as is one of the email addresses (which may also be found on the website).

Also, He has been known to frequent the Usenet Newsgroup news:sci.fractals (http://news:sci.fractals)
 


Title: Re: Boundary following for mandelbrot and Julia sets - John Bailey?
Post by: web2k on February 17, 2007, 05:09:43 AM
Hi Duncan,
Just to let you know, on this forum John Bailey is web2k, me.

Checking the URL you gave (http://www.ibiblio.org/e-notes/MSet/BigM.htm), it politely cites my page which describes an edge tracing process but Evgeny Demidov (http://www.ipm.sci-nnov.ru/~Demidov/) is presumably the author of the fast edge tracing actually shown on the page.

My edge tracing programs are the victims of technological advancements.  I have not kept them updated as systems changed. I am delighted to see the technique employed.

John aka web2k


Title: Re: Boundary following for mandelbrot and Julia sets - Found it.
Post by: Duncan C on May 03, 2007, 06:19:09 PM
Hi Duncan,
Just to let you know, on this forum John Bailey is web2k, me.

Checking the URL you gave (http://www.ibiblio.org/e-notes/MSet/BigM.htm), it politely cites my page which describes an edge tracing process but Evgeny Demidov (http://www.ipm.sci-nnov.ru/~Demidov/) is presumably the author of the fast edge tracing actually shown on the page.

My edge tracing programs are the victims of technological advancements.  I have not kept them updated as systems changed. I am delighted to see the technique employed.

John aka web2k


John,

Thanks for the reply. I finally managed to track down information on the boundary following algorithm. it turns out the source code for the Java applets is available on the parent site for the Mandelbrot boundary tracing applet. It can be found at the bottom of the page here: The Mandelbrot Set Anatomy Page (http://www.ibiblio.org/e-notes/MSet/Contents.htm/original)

Dr. Evgeny Demidov is listed in the header of the source code as the author, but it's been long enough that he doesn't remember the details, and suggested to me that the code came from Michael R. Ganss, author of "AlmondBread".

In any case I was able to figure out the boundary tracing algorithm from the java code at the above site, and get it implemented in my app. The code is very dense and a little cryptic, so it was a challenge for me to figure out. For plots with large contiguous areas at the same iteration value this algorithm increases plot speed dramatically. (sometimes by an order of magnitude or more.)

I was able to modify the algorithm slightly so that it doesn't miss "hollow" areas in iteration areas, such as the lowest iteration bands for Mandelbrot and connected Julia sets, and the hollow regions in disconnected Julia sets.

If anybody else needs help figuring this out I'd be happy to help.



Duncan C