########################################################################### Documentation To Mand_Tst Program V1.1. ########################################################################### History of releases: --------------------------------------------------------------------------- First version 1.0, released august/september 1992. --------------------------------------------------------------------------- Version 1.1, a bug fix up, released november 1992. --------------------------------------------------------------------------- Version 1.1 has corrected a bug which showed up on some machines during exit to DOS and save/load of mandel files. The operating system reported "Memory allocation error". This bug was not seen on PS/2 machines running DOS 5.0!! The functions for DOS exit and save of mandelfiles are modified, they will run a bit faster on Version 1.1! PS!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Mandelfiles saved with Version 1.0 program, can not be loaded by Version 1.1 program. I am sorry about that, but one parameter saved with the Version 1.0 program may cause the above mentioned problem when loaded into a Version 1.1 program. The program will not load a file from the former version!!!!!!!!!!! ########################################################################### Using The Program: The Program is soo easy that no documentation will be needed, but some small notes will follow: When You run Mand_tst in a DOS session under Windows things goes slowly! Run the program under DOS for max speed! If You have a MathProcessor installed use floating point calc!! If You do calculations with max pixels i.e. 650, the program needs 420 Kbytes free mem after it has loaded. To get maximum free memory, use QEMM / 386MAX / BLUEMAX or some other MemMgr. The color EPSF files are good when printed on a fullcolorprinter like Canon CLC500, I have not tested other colorprinters. On top of the EPSF image all the parameters for the actual image are printed in two lines. If You are calculating at max pixs and suddenly wants to break the calc, You can save the partial calculated image, quit program, resume some time later, load the partial image and resume calculation where You left off! The same applies to entering DOS command line, but then automatically an temp image file will be created and saved, when You leave DOS this file will be loaded and calculation will resume where You left. The display routines use real coordinates, and lines only. At low resolution, say 120 pixs You see the lines on screen. When You increase the pixs, say 370 pixs you will still see breaks in the display. This is due to mismatch between resolution of the screen (640 x 480) and the real coordinate point, which only occasionally will fall on an exact screen point. The EPSF files will never have breaks in the image, because they use the iteration value at the real point to generate color/greyscale value at that point in the image. The Program is rather simple and uses brute force under calculation. The speed is just pain slow when using floating point and no mathprocessor!! It is no competitor to FractInt, but the user interface may be easier. The max zoom is 1000 due to the use of float and only 16 bits behind the point in fixed point calc. The error for fixed point calc and large zoom is making rough images, use floating point to get smooth and good images at these high zooms and high pix values!!! Use low pix values to navigate to an interesting area, then increase pixs to a high value to get good images! Good Luck!! vatne@edb.uib.no ##############################################################################