The ZipCode format

  Disclaimer: The description below is a mere extrapolation from the files and
program sources I have encountered.

  The ZipCode algorithm is used to split and compress  disk  images  and  1541
disks so that the chunks may fit onto a single 1541 disk. This way even people
connecting to the Net with Commodore machines and 1541  drives  can  download,
extract and use such disks.



                         The diskpacked ZipCode format

  The algorithm creates four separate files out of one disk. On  the  PC,  the
base name of the ZipCode archives comes from the name  of  the  original  disk
image. The four files are numbered this way:  '1!BASE.EXT', ..., '4!BASE.EXT'.
An extension of this format is to handle 40 track disk  images,  as  well,  in
which case an additional archive, '5!BASE.EXT' is created. These files contain
the following data:

  FILE          POSITION        DESCRIPTION
  1!BASE.EXT    $0000-$0001     Load address (usually $03FE)
                $0002-$0003     ID of the original disk (not the  one  in  the
                                BAM but the one in block  headers);  in  files
                                created on a PC it may contain the string '64'
                $0004-          Data of tracks 1-8

  2!BASE.EXT    $0000-$0001     Load address (usually $0400)
                $0002-          Data of tracks 9-16

  3!BASE.EXT    $0000-$0001     Load address (usually $0400)
                $0002-          Data of tracks 17-25

  4!BASE.EXT    $0000-$0001     Load address (usually $0400)
                $0002-          Data of tracks 26-35

  5!BASE.EXT    $0000-$0001     Load address (usually $0400)
                $0002-          Data of tracks 36-40


  Blocks in track data are usually recorded with a soft sector  interleave  of
10-11 blocks and not sequentially. This way is much  faster  to  compress  and
decompress on a Commodore machine. Here is the structure of a block:

  POSITION      DESCRIPTION
  $0000         BITS 6-7: Compression method:
                  00    Block data is stored unchanged (byte by byte)
                  01    Block is homogeneous (filled by a single byte)
                  10    Block data is RLE-compressed
                  11    Unused
                BITS 0-5: Track number
  $0001         Sector number
  $0002-        Block data

  Unchanged block data:

  $0002-$0101   Data of the original block (256 bytes)

  Homogeneous block data:

  $0002         Fill byte

  RLE-compressed block data:

  $0002         Length of compressed data (LEN)
  $0003         Mark byte
  $0004-LEN+3   RLE-compressed data (LEN bytes)
                A 'mark byte - repeat count - data byte' byte  sequence  tells
                that there was a 'repeat count' long sequence of  'data bytes'
                in the original block; sequences shorter than four  bytes  are
                not compressed



                         The filepacked ZipCode format

  This format is rarely used and is similar to diskpacked ZipCode archives and
Lynx archives, as well. Instead of storing all the  sectors  on  the  original
disk, only the sectors used by the files are stored.

  The file names of the archives are similar to those of the diskpacked  ones,
but they start with letters like A!*.*, B!*.* and so on:

  FILE          POSITION        DESCRIPTION
  A!BASE.EXT    $0000-$0001     Load address (usually $03FF)
                $0002           Number of blocks in the file (max. 166 blocks)
                $0003-          Continuous block data

  B!BASE.EXT    similarly, if 166 blocks were not enough for the archive
  ...

  The block data is ZipCoded but there is no track and  sector  number  before
the data, the compression method is in the track link of the  sector  instead.
All 256 bytes of the sector are stored, even in the last block of files:

  POSITION      DESCRIPTION
  $0000         BITS 6-7: Compression method:
                  00    Block data is stored unchanged (byte by byte)
                  01    Block is homogeneous (filled by a single byte)
                  10    Block data is RLE-compressed
                  11    Unused
                BITS 0-5: Track number of next sector (0 if last block)
  $0001         Sector number (or number of used bytes in last block)
  $0002-        Block data (the same as with diskpacked archives)

  There is a special file with the name X!*.* which contains a little  program
listing the contents of the archive:

  ADDRESS       DESCRIPTION
  $0801-$09FD   Listing program
  $09FE         Number of archive data files (e.g. 1 if only A!*.* exists)
  $09FF         Number of files contained in the archive
  $0A00         Directory entries (21 bytes per entry)

  A directory entry has the following structure:

  POSITION      DESCRIPTION
  $00-$0F       Original name of the file, with trailing Shift-spaces
  $10           Type of the file (PETSCII uppercase character D, P, S or U)
  $11-$12       Length of the file in blocks
  $13-$14       Original track/sector position of first block of the file

  Note that filepacked ZipCode archives are not  designed  to  store  relative
files.