MIXIT has been updated. The /WORD=n and /GROUP=m options now work (on
input files only, those options on output filenames are ignored for now).

DIO accepts n bit words (multiples of 8) and will try to program whole
words into n/8 EPROMS. If word size times EPROM size is greater than the
available memory in the DIO, it will automatically switch to programming
1 EPROM and will prompt for group number.

A word about group numbers. From the VAX's and DIO's point of view, all
words are organised the same, that is, bytes are stored in the file in
increasing significance and bits within the each byte are mapped directly
to the corresponding bit in the EPROM. In other words, higher addressed
bytes are assumed higher significance in the word. The Gangpak will
program n-byte words into n EPROMS simultaneously. Assuming 16 bit words,
the Gangpak will program:

  fileoffset +0  +1  +2  +3  +4  +5  +6 ...
	     -----------------------------------
  socket #    1   2   1   2   1   2   1 ...

If the word size were 24, then the Gangpak would program:

  fileoffset +0  +1  +2  +3  +4  +5  +6 ...
	     -----------------------------------
  socket #    1   2   3   1   2   3   1 ...

MIXIT and DIO (in group mode) behave exactly the same. MIXIT will extract
every nth byte from the input file starting with the group_number divided
by 8. Presently the group_number is the power of 2 significance of the
bit in the word and must be a multiple of 8. I.e. (in VAX/T11/6502 addressing)

	Group	bits	fileoffset (=byte number)
	0	0-7	+0
	8	8-15	+1
	16	16-23	+2
	24	24-31	+3
	etc.

If the majority would prefer that the group number be a byte number (0-3
in the above example), I would be willing to change it.

The important thing to remember is that the 68000 operates contrary to this
scheme. Bytes are addressed by DECREASING significance so the concept of
low byte/high byte should not be used. Instead use the fileoffset value
as an indicator of the EPROM contents (meaning I should really change the
group code to be a byte number).

ds