Re: Standard codes (out of topic !?)

miK wrote:
> 
> Neil Franklin wrote:
> 
> > "Meph" <UZS...@uni-bonn.de> writes:
> > >
> > > 1963 ASCII (American Standard Code for Information Interchange)
> > > - specified as 7-bit code for telecommunication and data exchange
> > > - 100 of 128 positions are used (only upper case)
> >
> > Prepend this with: U.S./AT&T Teletype code. Existant at least in 1953
> > (I have used a device built then). Had 64 printable characters (codes
> > 32-95) identical with todays ASCII with a few substitution (no. 95 (_)
> > was then a leftwards facing arrow, no. 94 (^) was an actual upwards
> > facing arrow, possibly others I do not know of).
> 
> Was this the 1st ASCII-alike 7-bit code, to your knowledge? (meaning:
> were the 0's and 1's arranged in the same order in US/AT&T as in ASCII?)
> 
> TMK (at least in the early days) teletypes used 5-bit Baudot-code,
> "bracking" each sign with an independent start- and stop-element, thus
> using 7 bits rather bluntly. And not being able to transmit a lowercase
> char, cos' of being Baudot...
> 
The five-bit Baudot code (extant in the 1930s and said to have been developed
out of stock-ticker codes that began as Morse) preceded each code with a
two-bit Mark (high) and followed with a one-bit Space (low).  An RC filter
charged by the double Mark tweaked a synchronous motor that had to be set with
a tuning fork at each machine on the net; if this slipped more than a wee in
use, you got gobbledygook.  The regular double Mark is responsible for the
uniform "outboard motor" tone of the RTTY signals that used to be ubiquitous
all over the short-wave bands, and were standardised at 60 or 100 5-char
words/minute.  /All/ machines on any net had to be within 1/4 Mark of the same
speed.
  The code permitted 32 combinations (2**5), allowing most of the lower case
QWERTY keyboard in three rows of keys; however the whole character set was
rather shorter than 64 "printable" characters as only 30 codes were assignable
in any case:
  The "blank" (00000) was precluded by any stretch of blank tape or "air
time."
  The "rubout" (11111) was reserved for correcting errors on the tape, and
doubled to slam the carriage back into the lower case (LETTERS).
  The "space" (01000) had to be physically produced in either case.
  A NUMBERS code slapped the carriage into the upper case, the first row of
which was the QWERTY numbers.  Most of the other two rows of keys were machine
functions.
  " ' . , ; : were included in the second row of the upper case, but few used
them as it was faster and less-noisy to type QT/UQ AP(S) PD CM SMC CLN,
respectively, in the text stream (followed by two spaces).
  Punctuation had to be put in the upper case because their usual positions
were taken by machine functions that had to operate in both cases (CR-CR-LF,
BS, e.g.).
  CR acquired its modern position on the Baudot keyboard, replacing the
semicolon/colon (;/:) (the AT keyboard has one more rank of keys on the
right).  It was ALWAYS pressed/punched twice to allow the carriage time to
return to battery, else the next character would print in the middle of the
line on the way back.  (This protocol persists in the Gutenberg distribution
files and some (unidirectional) printers.)
  LF acquired its separate function on the Baudot TTY, CR being handled by a
motor and LF by a solenoid.  LF replaced the period (.).
  NUMBERS (if my fingers recall correctly 35 years later) replaced the comma
(,).  LETTERS was a muscle button on the face of the console, but also
sent/punched the rubout.
  Because LF was a separate function, overprinting was an available "art"
tool.  Much more popular than "ASCII art" was using the varying blackface of
different letters (enhanced by overprinting) to make "halftone" art, generally
of known subjects; existing tapes of such pages could be transmitted directly
to another tape punch simultaneously with being printed, and most ops had a
boxful of them.  Printing M over W gives a very-black halftone "dot," where
screen type yields something less than 50% black.
  BTW, the Baudot codes bear /no/ relation to later codes with more bits,
i.e., later codes were not arrived at by adding a bit or two to the Baudot
series; ASCII and most others are in numerical sequence, while Baudot isn't
even close.  It is said that the codes were distributed to allow the hammer
commutator maximum-average recovery time between characters, but a dirty
commutator will drop the second of certain pairs anyway.
-- 
  >^,,^<
The number you have reached is imaginary.
 Rotate your modem 90 degrees and reconnect.
http://blitz.org/fishhook/