History Early Data transmission technical considerations [telecom]

The term "bit" still describes the speed of the raw "data transmission over private lines". Things like characters, words, bytes, packets, frames, etc are formed at higher levels of the 7 layer OSI stack by the DTE equipment attached to the data circuit.

see

--Reed

Sometimes "yes", sometimes "no".

Consider analog line signalling where one 'symbol' represents multiple 'bits' in the original data stream.

Thus the 'classical' distinction between 'bits' and 'baud' in more commplex analog data links.

For an extreme case, consider the Telebit Corp "PEP" protocol, used in their 'trailblazer' (and compatible) modems.

Well, yes, various methods have been used to increase the usable bit rate of an otherwise band-limited channel. QAM, Trellis coding, etc. However I think the the usable "speed" of the channel would still be expressed in bits per second.

Also I took "L or J's" comment as aimed at bytes, etc. Perhaps they meant baud, since that always has been a common point of mis-understanding when speaking of modem speeds.

Thanks for the reminder, Reed

The word "bit" actually comes from Nyquist back in the twenties, but the actual relationship between channel capacity in bps, bandwidth, and noise, was worked out by Shannon in 1948.

--scott

Actually, I believe that Lisa and/or Jeff were pointing out that 1957 was a very _early_ use of the term "bit." This is in fact quite interesting: Many people today do not realize that transmission of binary information over electrical circuits has been going on for a good

170 years or so, but the term "bit" was not commonly used until about fifty years ago.

The 1953 edition on "Principles of Electricity applied to Telephone and Telegraph Work" [aka the AT&T Big Green Book] does not use the word "bit" anywhere in its 350+ pages, except to refer to "bits of paper" and "bits of iron." However, they use the following terms extensively:

Make or Break

Mark and Space

Dot and Dash

I will leave it to others to expand this thread further. When I get started about technology and semantics there is significant risk of a

5,000 word diatribe flowing from my keyboard...

Jim

************************************************** Speaking from a secure undisclosed location.

I suppose so, *IF* you forget about Nyquist in the 1920s, Shannon in the

1940s, and everybody else in between.

Nyquist was a "mark and space" man. He also used the term "signal element" a lot.

The earliest known use of "bit" in a published paper was by Shannon in

1948, but he didn't coin the term. That honor goes to J.W. Tukey [a Massachusetts boy], who proposed that "bit" was an acceptable [and clever] contraction of the term "binary digit."

It would be a good decade before the term was in common usage, with the BSTJ paper cited by Lisa and Jeff being an excellent example.

Jim Bennett

****************************************************** Checking my facts in a secure undisclosed location.

Was he also the one...binary digits are associated with logs to the base 2, and someone suggested 'nit' for a digit resulting from logs to the base e, and perhaps 'dit' for decimal digits?

***** Moderator's Note *****

Dit-dah-dit-dah-dit, dah-dit-dah

Bill Horne Moderator