[GreenKeys] Error-correcting codes

[Dave Emery]

On Thu, Mar 10, 2005 at 09:17:25AM -0500, Tim McNerney wrote:
> When is the earliest anyone here can remember of error-correcting 
> equipment being used in RTTY communications?

	I worked for Codex Corp in the summer of 1966 when I was a
senior in high school and at that time they had introduced forward error
correcting RTTY equipment using Massey convolutional coding at rate 1/2.
I know it was employed on at least a few NASA and military/spook HF
circuits at the time - typically using two channels on a VFT 16 channel
fdm mux for two bitstreams - one parity and one data.   I forget whether
the box was capable of combining parity and data bits into one stream,
but I think so...

	At the time it was thought to be a bleeding edge product and a
relatively revolutionary technology (which it sure is, digital
cellphones, digital TV and even hard disks would be worthless without
forward error correction as the raw error rate of most channels is
nowhere near good enough to support reliable transmission without it).

	I remember playing with the units with a couple of 28 ASRs in
the company lab... indeed they would work very well cleaning up a
garbled transmission from a tape on one machine to the other over a bad
channel simulator that would produce almost unreadable copy otherwise.

	This would establish roughly 1964 or 1965 as the time this
technology was first deployed.

	Whether the NSA or other classified folks used forward error
correction before this in crypto applications is not something I know
the answer to - and would be curious to hear...  I do vaguely remember
the WW II era revolutionary pioneering secure digital HF voice terminal
SIGSALY used some mighty way ahead of the curve ideas and might possibly
have implemented error correction somewhere.

	And I know that a lot of the pioneering hf spread spectrum work
done at MIT Lincoln Labs in the 50s must have touched on these ideas but
whether they deployed anything in actual gear I do not know. They did
invent other useful stuff such as the RAKE receiver, widely used in
modern SOTA wireless systems such as CDMA base stations.

	Of course use of error DETECTING codes (eg parity) was much
earlier in HF RTTY - some time in the 30s or early 40s when ARQ was
introduced - certainly operational by the late 40s or early 50s.  But
that stuff (such as Moore code) merely detected errors and requested
a repeat, not corrected them.

-- 
   Dave Emery N1PRE,  die at dieconsulting.com  DIE Consulting, Weston, Mass 02493