Ruff & Ready docs (?) for MANDLE51.PAS == a first shot at mapping the
Mandlebrot Set. The source was hacked on tube as more a mental doodle
than any really productive and planned attack on the problem.

MANDLE5A Drives the EGA display directly and so YOU MUST HAVE AN EGA TO
SHOW THE MAPS (EGA/VGA Mode 15)! Also the .EXE uses a co-processor.
For the uninitiated, Mandlebrot or Julia set plotting requires TONS of
floating point calculations. My suggestion is don't even think about
it if you do not have an 80x87 (or better) co-processor.

Some routines from the Borland Grafix Toolbox (Ecchhh) were used to
give a quick distribution histogram.

Data are stored on files as integers all-in-a-row for the image. One
group at the front of the file are reserved for info on size,
repetitions and scale of the image in question.

Obviously the software could be cleaned up and given direction. I
focussed pretty hard on the iteritive calculation of exp(-z) for
efficiency. I have even written some .ASM direct 80x87 drivers to
substitute for the Borland v5.0 compiler calls. Guess what -- mine
were only a few licks better. But, I am a hacker, so perhaps a few of
you real wizards out there could attack the problem. FP Calculation is
really the Achilles' heel of these toys (Oh! For a Cray!).

Any thoughts, please give me a call. Again, I will not hear remarks
about sloppy code or immoral programming practices ( I think there are
actually two GOTO's in here!).


Jeff Thompson
9058 Gorman Road
Laurel, MD 20707
(301) 490-5077