Anyone recognize this Baudot encoding?
Eric Fischer
across an article, "A New Keyboard Perforator for the Baudot Printing
Telegraph System" by A. E. Thompson, in _Electrical Communication_,
volume 3, pages 295-300. The article says that "the Baudot system in
its usual form is well known to most telegraph engineers," but its
table of the Baudot code (which, prior to the invention of the keyboard
perforator described in the article, had to be typed in raw binary!) is
nothing like any of the Baudot variants I've ever seen before:
code letters figures
------- ------- -------
00001 letters space (do these shift cases *and* produce a space?)
00010 figures space
00011 * *
00100 Y 3
00101 S ;
00110 B 8
00111 R -
01000 E 2
01001 X ,
01010 G 7
01011 M )
01100 I _O_ (that is, an underlined, superscript O)
01101 W ?
01110 F _F_
01111 N #
10000 A 1
10001 _t_ . (yes, an underlined superscript *lowercase* t)
10010 J 6
10011 K (
10100 U 4
10101 T !
10110 C 9
10111 Q /
11000 E' &
11001 Z :
11010 H _H_
11011 L =
11100 O 5
11101 V '
11110 D 0
11111 P %
Has anyone ever seen this encoding in use anywhere before? Another
article, by the same author, thirteen years later, uses what seems
to be the standard Baudot encoding, the "international" version.
eric
Robert Billing
> 01100 I _O_ (that is, an underlined, superscript O)
This may be the dead giveaway. This symbol is used in IIRC Spanish to
write N _o_ as an abbreviation for "number".
--
I am Robert Billing, Christian, inventor, traveller, cook and animal
lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/
"Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock
phasers on the Heffalump, Piglet, meet me in transporter room three"
John Varela
wrote:
> > 01100 I _O_ (that is, an underlined, superscript O)
>
> This may be the dead giveaway. This symbol is used in IIRC Spanish to
> write N _o_ as an abbreviation for "number".
Spanish uses superscript underlined o to represent masculine ordinal numbers.
For example, 2_o_ is segundo. But it also uses superscript underlined a for
feminine ordinals, as in 1_a_ for primera. This set has no superscript
underlined a and the F, H, and t are a whole 'nother thing.
--
John "try again" Varela (delete . between mind and spring to e-mail me)
John Kdllin
Robert Billing wrote in message ...
>Eric Fischer wrote:
>
>> 01100 I _O_ (that is, an underlined, superscript O)
>
>This may be the dead giveaway. This symbol is used in IIRC Spanish to
>write N _o_ as an abbreviation for "number".
_a_ and _o_ are used in Spanish to designate ordinals: 3_o_ means "tercero",
or "third" for a masculine ordinal, and 3_a_ means "tercera", or third for
feminine ordinal. Being able to write N_o_ is an added bonus.
--
John Kdllin -- johnkal at microsoft dot com
My opinions. Mine, MINE, *MINE*!
Tony Duell
[...]
: table of the Baudot code (which, prior to the invention of the keyboard
: perforator described in the article, had to be typed in raw binary!) is
: nothing like any of the Baudot variants I've ever seen before:
I have here volume 6 of 'Modern Electrical Engineering'. The date is not
given, but I would guess early this century. In Chapter VIII 'The
Baudot', it gives essentially the same coding. However, you're giving the
bits in a strange order, namely bit 1 first and then on to bit 5.
The table I have gives the bits in order 5, 4 ,1, 2, 3. This is not as
strange as it sounds, those are the order of the 5 keys on the chording
keyboard.
Bit finger
5 Middle, left hand
4 Index, left hand
1 Index, right hand
2 Middle, right hand
3 Third, right hand.
: code letters figures
: ------- ------- -------
: 00001 letters space (do these shift cases *and* produce a space?)
: 00010 figures space
Yes. Space and change case. The printer printed on a strip of paper, so there
was no need for CR or LF, BTW
[...]
: 01111 N #
That's N_o_ (superscript lower case O), not a # in my table
[...]
: Has anyone ever seen this encoding in use anywhere before? Another
See above. I think this is the original Baudot code.
Also, the conventional baudot system of the time time-multiplexed 4 sets
of 5 bits onto a single line. There was a rotating mechanical commutator
for this.
-tony
Dik T. Winter
> The article says that "the Baudot system in
> its usual form is well known to most telegraph engineers," but its
> perforator described in the article, had to be typed in raw binary!) is
> nothing like any of the Baudot variants I've ever seen before:
What you had was indeed the original Baudot code.
> Has anyone ever seen this encoding in use anywhere before? Another
> article, by the same author, thirteen years later, uses what seems
> to be the standard Baudot encoding, the "international" version.
And that second article did not contain Baudot code but Murray.
The original Baudot code was CCITT #1. The modified (and more widely
known and erroneously called Baudot) code was the Murray code aka
CCITT #2. In most of these codes the non-alphanumerical symbols were
a bit indeterminate.
--
dik t. winter, cwi, kruislaan 413, 1098 sj amsterdam, nederland, +31205924131
home: bovenover 215, 1025 jn amsterdam, nederland; http://www.cwi.nl/~dik/
D. Peschel
Dik T. Winter <Dik.W...@cwi.nl> wrote:
>What you had was indeed the original Baudot code.
[...]
>The original Baudot code was CCITT #1. The modified (and more widely
>known and erroneously called Baudot) code was the Murray code aka
>CCITT #2. In most of these codes the non-alphanumerical symbols were
>a bit indeterminate.
Isn't the formal name "CCITT International Telegraph Alphabet #1" or something
like that? (BTW, I just read about #3 and #5 but not #4! Does anyone have
any info on those?)
My real question is: Does anyone know what method Baudot used to allocate the
letters in his code? I once sat down with a table in order to find the logic
in the code, but the whole thing looks very ad hoc. At the time I came up with
the theory that the figures were assigned first. (There's some correspondence
between the code numbers for the figures and their numerical values, IIRC.)
Maybe some of Baudot's notes would clear things up.
The code given in the original post (now snipped) had some special characters
in the alphabetic case, which I don't remember seeing in my table. So either
there was some variation in the lower case as well as the upper, or CCITT made
some changes when they adopted Baudot's code as their own.
-- Derek
Eric Fischer
> My real question is: Does anyone know what method Baudot used to
> allocate the letters in his code? I once sat down with a table in
> order to find the logic in the code, but the whole thing looks very
> ad hoc. At the time I came up with the theory that the figures were
> assigned first. (There's some correspondence between the code
> numbers for the figures and their numerical values, IIRC.)
Actually, there is a logic, however weird, to most of the code,
though some of the punctuation marks seem to be thrown in wherever
they would fit. Remember that in the original Baudot system the
telegraph operator was pressing a key to correspond to each bit,
not typing on a typewriter-style keyboard, so there had to be some
logic to the code to allow people to memorize it. Here is (most of)
the code again, this time with the proper bit order and sorted by
category:
Control characters:
o. ... space (letters)
.o ... space (figures)
oo ... erasure
Numbers:
.. o.. 1 .o o.. 6
.. .o. 2 .o .o. 7
.. ..o 3 .o ..o 8
.. o.o 4 .o o.o 9
.. ooo 5 .o ooo 0
Vowels: Middle third of the consonants:
.. o.. a oo o.. k
.. oo. é oo oo. l
.. .o. e oo .o. m
.. .oo i oo .oo n
.. ooo o oo ooo p
.. o.o u oo o.o q
.. ..o y oo ..o r
Other consonants:
.o ..o b o. ..o s
.o o.o c o. o.o t
.o ooo d o. ooo v
.o .oo f o. .oo w
.o .o. g o. .o. x
.o oo. h o. oo. z
.o o.. j
I don't know what the logic was for the ordering within each group, but
the second half of the numbers use the same right-hand pattern as the
first half, and the last third of the consonants similarly parallels
the first third. I'm not going to try to come up with an explanation
for why the middle third and the vowels go in the opposite order from
the first and last thirds.
Thanks to everyone who's been answering these Baudot questions! I
should have done a little more reading before I posted.
eric
Dik T. Winter
> Isn't the formal name "CCITT International Telegraph Alphabet #1" or
> something like that?
Indeed: International Alphabet #1. No Telegraph in there.
> (BTW, I just read about #3 and #5 but not #4! Does anyone have
> any info on those?)
That is the one I have been missing a long time, never found it.
> My real question is:
And I have no answer to this.
> The code given in the original post (now snipped) had some special characters
> in the alphabetic case, which I don't remember seeing in my table. So either
> there was some variation in the lower case as well as the upper, or CCITT
> made some changes when they adopted Baudot's code as their own.
There was variation in the non-alphanumeric symbols. So both cases show
variation. I know of quite a few variants of #2.
jsa...@freenet.edmonton.ab.ca
: Has anyone ever seen this encoding in use anywhere before? Another
: article, by the same author, thirteen years later, uses what seems
: to be the standard Baudot encoding, the "international" version.
The standard 5-level Teletype code, although it's often called "Baudot",
after Emile Baudot, the inventor of the 5-unit code printing telegraph,
doesn't at all resemble the code he actually used - and this code is the
one you probably saw in that article.
Later, one John Murray, in London, devised a printing telegraph using a
code very similar to the standard 5-unit code used today. However, like
Baudot's original code, it also used space characters that shifted, and
the figures case was mostly different. (It included characters like 1/ ,
3/ , 5/ , and 7/ , raised, for making fractions.)
You can see a chart of the standard 5-unit code including the original
Murray code on which it was based on my web site at
http://www.freenet.edmonton.ab.ca/~jsavard/tele03.html
So some people refer to today's 5-unit code as the Murray code, and they
are closer to being correct than those who call it Baudot.
John Savard
John Savard
>In article <F11nJ...@midway.uchicago.edu> dpes...@u.washington.edu (D. Peschel) writes:
> > Isn't the formal name "CCITT International Telegraph Alphabet #1" or
> > something like that?
>
>Indeed: International Alphabet #1. No Telegraph in there.
>
> > (BTW, I just read about #3 and #5 but not #4! Does anyone have
> > any info on those?)
>
>That is the one I have been missing a long time, never found it.
I thought I saw 4 in the James Martin book "Telecommunications and the
Computer". #3 was a 6 bit version of #2, with a parity bit added, and
#4 was the Moore ARQ code, IIRC.
There were two completely different 7 bit codes, one a 3 out of 7 code
and one a 4 out of 7 code, used to represent 5-level code. The newer
one is given in the ARRL Handbook. I think both of them are in the
table on my web page too:
http://www.freenet.edmonton.ab.ca/~jsavard/tele03.html
John Savard
http://members.xoom.com/quadibloc/index.html
John Savard
>In article <F11nJ...@midway.uchicago.edu> dpes...@u.washington.edu (D. Peschel) writes:
> > Isn't the formal name "CCITT International Telegraph Alphabet #1" or
> > something like that?
>
>Indeed: International Alphabet #1. No Telegraph in there.
>
> > (BTW, I just read about #3 and #5 but not #4! Does anyone have
> > any info on those?)
>
>That is the one I have been missing a long time, never found it.
I checked my page, and refreshed my memory. #3 was Moore ARQ; #4 was
the one with the parity bit.
John Savard
http://members.xoom.com/quadibloc/index.html
Dik T. Winter
> Dik.W...@cwi.nl (Dik T. Winter) wrote, in part:
> > > (BTW, I just read about #3 and #5 but not #4! Does anyone have
> > > any info on those?)
> >That is the one I have been missing a long time, never found it.
>
> Computer". #3 was a 6 bit version of #2, with a parity bit added, and
> #4 was the Moore ARQ code, IIRC.
>
> There were two completely different 7 bit codes, one a 3 out of 7 code
> and one a 4 out of 7 code, used to represent 5-level code. The newer
> one is given in the ARRL Handbook. I think both of them are in the
> table on my web page too:
What I have found as CCITT #3 is a 3 out of 7 code. I also do have the
Moore 3 out of 7 code, but that one is completely different. In addition
there is the Telex 4 out of 8 code. And I have one additional 4 out of 8
code.
> There were two completely different 7 bit codes, one a 3 out of 7 code
> and one a 4 out of 7 code, used to represent 5-level code. The newer
> one is given in the ARRL Handbook. I think both of them are in the
> table on my web page too:
>
> http://www.freenet.edmonton.ab.ca/~jsavard/tele03.html
Indeed. And CCITT #3 (ITA 3) is indeed the code I had. ITA 4 is there
too and it apparently is ITA 2 with a leading 0 bit added plus a few
symbols where the leading bit is 1 (character 32, beta and SYNC).
John Savard
>What I have found as CCITT #3 is a 3 out of 7 code. I also do have the
>Moore 3 out of 7 code, but that one is completely different. In addition
>there is the Telex 4 out of 8 code. And I have one additional 4 out of 8
>code.
Interesting. I'll have to visit your site again, and check into that;
I know my sources have given ITA 3 and the Moore ARQ as the same.
Another 3 out of 7 code is on my site - and I know of the IBM 4 out
of 8 code for BCDIC, but that's unrelated. (And there's AUTOSPEC too,
which I happened to run across, as yet another variant of 5-level code
for bad conditions; this one repeats each character twice, but
inverting the repetition depending on the character's parity.)
>Indeed. And CCITT #3 (ITA 3) is indeed the code I had. ITA 4 is there
>too and it apparently is ITA 2 with a leading 0 bit added plus a few
>symbols where the leading bit is 1 (character 32, beta and SYNC).
Even when I looked at my site, I didn't recheck it closely enough!
John Savard
http://members.xoom.com/quadibloc/index.html
John Savard
>What I have found as CCITT #3 is a 3 out of 7 code. I also do have the
>Moore 3 out of 7 code, but that one is completely different.
Not _quite_ completely different. Your tables show these codes in
octal; from your table, the codings are:
Moore ARQ CCITT #3
Q 130 1011000 015 0001101
W 122 1010010 045 0100101
E 016 0001110 070 0111000
R 023 0010011 144 1100100
T 121 1010001 105 1000101
Y 124 1010100 025 0010101
It's just that the source you used for Moore ARQ presented it in the
opposite bit order from that used by your source for CCITT #3.
Because ASCII is transmitted least-significant-bit first, it isn't
clear whether 5-level code and its relatives should be shown with the
bits in 54321 order or 12345 order.
John Savard
http://members.xoom.com/quadibloc/index.html
Dik T. Winter
> Not _quite_ completely different. Your tables show these codes in
> octal; from your table, the codings are:
> It's just that the source you used for Moore ARQ presented it in the
> opposite bit order from that used by your source for CCITT #3.
Indeed, except for a Z <-> U reversal, but I think that is an error in
my Moore table. I will have to check (first check where I found it in
the first place back in about 1985). I thought I had checked for bit
reversal, apparently I had not.