(Comp.sys.handhelds)
Item: 2308 by ervin at pinbot.enet.dec.com
Author: [Joseph Ervin]
  Subj: New HP48SX Screen Dissolver
  Date: Mon Mar 04 1991 22:54 

[NOTE: For a cute demo of DISS in action, see DISSDEMO.  Also, the CLK
program by Dave Marsh uses DISS within a library.  -jkh-]

Sorry for reposting this program so soon, but I found a potential bug and I
wanted to kill it before it bit anyone.  While I was at it, I shortened the
program by around 50 bytes, and improved the speed of the bit-wise dissolve
from 5 seconds down to around 4.2 seconds.  

Enjoy!

>>>Joe
**************************************************************
The following is a new version of my screen dissolver.  This version (0.2) is
slightly faster than the original version posted recently, and is roughly 50
bytes smaller.

A future release may allow dissolving into the menu area when viewing the
stack, if people think that would be useful (let's hear some feedback :-).

The program takes two arguments:

2: real number (zero or one)
1: GROB (131x64)

The program will do different things with the GROB depending on whether a 1 or
a 0 is in level 2.  If level 2 contains a 0, then the program will do a
bit-wise dissolve of the GROB on level 1 into PICT.  The idea here is that you
would use DISS in a program  after previously displaying some graphic image
via PVIEW.  When you execute DISS, the new graphic on level 1 will replace the
graphic currently on the display 1 pixel at a time, in a random fashion.  The
program currently requires both the GROB on level 1 and PICT to be 131x64. 
Any other dimensions will generate an error message.  

If a 1 is on level 2 when DISS is executed, then instead of doing a bit-wise
dissolve, it will do what I call "tiling".  In this mode of operation, the
program will move random _nibbles_ from the level 1 GROB into PICT, rather
than doing one bit at a time.  The effect is slightly different, and it runs
around 5 times as fast as the bit-wise dissolve (4.2 seconds for a bit-wise
dissolve, and around .9 seconds for tiling).  I think both modes look pretty
cool (of course). :-)

It is written in a combination of system RPL and machine language (like
Bill Wickes' ASC routines themselves).

I thank Jan Brittenson for his assistance during this endeavor.  I could not 
have finished this project without his wonderful STAR assembler and his 
expert advice on writing machine language and system RPL code.  Also to be
credited are Alonzo Gariepy, who did much of the pioneering work on the saturn
processor, and Derek Nickel, who has provided a wealth of information
regarding system RPL programming on the 48SX.

I will be watching the conference dilligently for feedback, and ideas for
improvement. I found it a simple matter to include this function into TETRIS
so that the startup screen dissolves into view, producing a nice effect.

Enjoy!

>>>Joe Ervin