PHD  Data Compression
                          ~~~~~~~~~~~~~~~~~~~~~
                      Copyright:  Paul Douglas  1992
                      ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


First things first - this is only a DATA FILE compression utility
If you want to compress your executables then use Powerpacker or
Imploder or similar. The de/compression routines have been written
in 680X0 assembly language on Commodore Amiga but could be ported
to any 680X0 platform (Apple Mac, Atari ST etc) but only with the
my permission. You have permission to use the decompression routine
in your own programs as long as it is released as Public Domain
or Shareware (a mention of my name would be nice but thats entirely
up to you  .You could also send me a small contribution if you like.)
If you wish to use the decompression routine in a commercial program
then you'll need my permission.

If you wish to distribute the above routines then you must include
this text file and all associated files.

Now thats the nasty stuff out of the way.



1. About PHD Data Compression

The routines use lossless data de/compression algorithms, this 
basically means after compression and then decompression the output
file will exactly match the input file so can be used on any type of
data. I'm not going into heavy details but basically it uses an
enhanced version of the Lempel-Ziv, Adaptive Huffman combination
as used in LHArc etc. For this reason its not as fast as some of
the more simple packers, but usually compresses data more
efficiently than any other data compression utility on the Amiga.
The decompression routine has been heavily optimised to run as
fast as possible on a standard 68000. Please contact me if you
want an enhanced version for 68020+ processors.



2. Advantages/Disadvantages

Generally this compression utility provides better compression
ratios than any similar utility on the Amiga, especially for
graphics and sound data. ( Handy for A1200 programmers, 256
colour graphics sure do burn up disk space!) 

On the downside is the decompression speed. This shouldnt be a
problem if you're using 68020+ processors (A1200 or A4000) but
on a bog-standard 68000 you'll have to decide if the timings
are acceptable or not. This could depend on how much disk
accessing your program does.



3. Using The Compression Utility

I've included a simple utility that allows you to load a file
compress it and then save it as a compressed file. This uses
an ARP requester for file input/output so is pretty simple to
use. This works fine on a WB1.3 machine but I've not had the 
chance to test it on a WB 2.0 or 3.0 Amiga.
To use it you'll need at least 1 meg of memory, perhaps more
if you wish to multitask or your memory is divided half meg
fast, half meg chip ram (luckily decompression is a lot less
greedy on the memory).

OK, assuming you have enough memory (and if you're developing
Amiga programs you should have as much as you can get!) then its
simple to use.


Loading a file

click on gadget marked LOAD and use arp requester as in devpac etc


Compressing a file

There are 6 gadgets marked 0 to 5. Click on any of these to start
compression once you've loaded a file. It's probably best to
experiment with different values on each file, but most of the time
selecting 5 will produce the best results.

Once a file has been loaded it can be compressed at different ratios
without having to reload the file.

Once compressed, the compressed length and compression percentage
will be shown.


Saving a file

Jeez, do I really have to explain this! Click on save and use Arp
requester as per usual. This saves the compressed file by the way!!



4. Using The Decompression Routine

This is probably the most important bit for all you programmers out
there. The binary dump of the decompression code is included so can
be easily incorparated in your own programs (I've not included the
source code because its pretty irrelevant, all you have to know is
how to use it). This code is relocatable and re-entrant so you can put
it anywhere on a word boundary.

I've included an example of how to load and decompress a file using
DOS, which shows the basics.

The decompression routine needs about 4k of memory ( again word-aligned
but doesn't have to be cleared ) see the include doc for actual size.

Total memory useage for decompression is about 5k, plus the memory for
the uncompressed file. For example you can uncompress a file that was
originally 200k in size with only 205k, with the extra 5k becoming
free to use after decompression.

On entry to the decompression program you'll need pointers to
                   1. Work buffer
                   2. Compressed file
                   3. Output uncompressed data
 

Compressed File Format is as follows

   Bytes 0-3       I.D. for PHD compression. This is so you can tell
                   the file apart from normal uncompressed files or
                   say Powerpacker files.

   Bytes 4-7       Length of original ( uncompressed ) file in bytes.           

   Bytes 8-End     Actual compressed data.


Please note, you must load the compressed file onto a Word boundary
else it'll be Guru time. 

I've not included any form of checksum on the compressed file, as I
think Amiga DOS does this anyway??

Anyway see the example code which should explain things more easily.

         Please send bug reports, comments etc to,


                                   Paul Douglas
                                   147 Winkworth Road
                                   Banstead
                                   Surrey SM7 2JP