Several emails arrived as a result of the item in MMDigest 091202.
Here are answers to some of the queries.

My organ presently has two ranks of small pipes, 37 Bourdon, and
33 reed oboes, from a church organ, plus a single long folded pipe
covering two bass octaves driven by four polyphonic oscillator ICs.
The 94 distinct tones are decoded and driven by two J-Omega MTP-8
MIDI to parallel converters each with 64 outputs, from
http://www.j-omega.co.uk/ 

MIDI music files on an SD card are streamed on 4 channels to the
decoders by an SD MIDI Controller.  Unused decoder outputs are used
to light LED lights mounted on pipes to show where the tune is being
played.

My software differs from Noteur and other MIDI editing software in
that you do not meddle with individual notes.  This is not a tool for
the wannabe composer, it is for the "Life-Is-Too-Short."  You define
your instrument, and give general directions about what should be done.

For instance, I get results that please me with the following.  Brass
instruments go to my oboes on channel 2(which are adjusted to sound like
trumpets); All instruments go to the Bourdons on chan 1; All bass notes
and percussion go to the bass on channel 3; The tune goes to the LEDs
on channel 4; If less than 10% of the notes are on the oboes, then add
the tune there too.  (The last rule makes piano arrangements quite
pleasant as I can use the manual stop to turn the oboes on and off in
interpreting the music as it plays.)

I let the software convert entire folders of MIDI music at one bash.
Many pieces sound okay without further adjustment, the rest I just
don't use.  With a polyphonic instrument of wide compass, I seldom need
to transpose anything, but this can be easily, even automatically done.

There are snags, though.  Firstly, to make the complexities of MIDI
manageable, the software had to be written in a particular computer
language for which a license fee is payable when shipped to anyone else.
It means I would have to charge maybe $70 depending on the level of
interest.  Secondly, it must be understood that being software, there
will be bugs.  These are only eliminated by testing, and it is the
users mainly who do the testing.

Thirdly if there is a demand it will take me a few months to polish
and package the software for distribution to run on other computers.
April 2010 might be a good target for this, but it might be useful to
distribute a few what are called "Betas" earlier.  The response to my
post in MMDigest is encouraging.

John Heath