Early history of ASCII?

35 views
Skip to first unread message

Tim Shoppa

unread,
Dec 9, 1996, 3:00:00 AM12/9/96
to

In article <RSM.96De...@platinum.math.arizona.edu>,
Robert S. Maier <r...@math.arizona.edu> wrote:
>What was the early history (1970's) of the ASCII encoding? In particular,
>did it ever undergo a revision?
>
>I distinctly remember seeing, back in the early 1980's, an old
>character-cell terminal that represented the ^ character (caret, octal
>\136) by an up-arrow, and the _ character (underscore, \137) by a
>right-arrow. The terminal must have dated back to the early 1970's.
>Someone told me that the terminal was interpreting \136 and \137 according
>to an early version of the ASCII standard, which was different from the one
>that was finally adopted. Can anyone confirm this? Unfortunately, I don't
>remember which company manufactured the terminal.

The printing characters you mention are shared by my model 33 Teletypes.
(In addition, mine has a slash through the letter O "Oh" and no slashes
in the digit 0 "Zero".) Literally dozens of different type elements
were available for the model 33 - the customer had the choice.

Looking through my model 33 manuals, the ANSI Standard X3.41-1974 is
mentioned, though there is no indication of which type element follows
this specification most closely.

Tim. (sho...@triumf.ca)

James Carlson

unread,
Dec 9, 1996, 3:00:00 AM12/9/96
to

In article <58hfvg$l...@nntp.ucs.ubc.ca>, sho...@alph02.triumf.ca (Tim Shoppa) writes:
[...]


|> Looking through my model 33 manuals, the ANSI Standard X3.41-1974 is
|> mentioned, though there is no indication of which type element follows
|> this specification most closely.

X3.41 is for the ANSI control codes (like CR, LF, FF). X3.4 is the
ASCII code itself.

--
James Carlson <car...@xylogics.com> Tel: +1 617 272 8140
Annex Interface Development / Xylogics, Inc. +1 800 225 3317
53 Third Avenue / Burlington MA 01803-4491 Fax: +1 617 272 2618

ds...@cameonet.cameo.com.tw

unread,
Dec 10, 1996, 3:00:00 AM12/10/96
to

Robert S. Maier (r...@math.arizona.edu) wrote --

> What was the early history (1970's) of the ASCII encoding? In particular,
> did it ever undergo a revision?

ISO and ANSI (then called ASA) recognized the need for data processing
standards in 1960 and began serious work on text encoding methods and
other issues in 1961-62. The first version of ASCII, which used six bits
and lacked lowercase letters and certain symbols, was approved in 1963.
Seven-bit ASCII, I think, was approved by ANSI in 1968 and was modified
very slightly in later revisions (there were several, but in some cases
at least, only function or character names were changed).

Material about the early history of ASCII can be found at the University
of Minnesota's Charles Babbage Institute. Last I checked, the URL was --

http://www.cbi.umn.edu/welcome.htm

I have a German document that _seems_ to say seven-bit ASCII was adopted
in Europe in 1965, three years before the U.S. Anybody who reads German
willing to take a look at it for me? There is just one paragraph that I'm
eager to get a full translation of.

Dan Strychalski
ds...@cameonet.cameo.com.tw
(Reposted Tue Dec 10 11:01:50 CST after first attempt appeared to fail.)

Markus Kuhn

unread,
Dec 10, 1996, 3:00:00 AM12/10/96
to

ds...@cameonet.cameo.com.tw wrote:
> I have a German document that _seems_ to say seven-bit ASCII was adopted
> in Europe in 1965, three years before the U.S. Anybody who reads German
> willing to take a look at it for me? There is just one paragraph that I'm
> eager to get a full translation of.

Just post the paragraph.

I guess, it refers to ECMA-6, a standard that you can order for free
(see www.ecma.ch and ftp.ecma.ch). The modern 1991 ECMA-6
standard is of course aligned with the modern ISO 646 (the
US-ASCII as we know it today), but the foreword talks about the
early history if I remember correctly.

There is also a lot of discussion in old Communications of the ACM
issues in the 1960s going on about ASCII.

Markus

--
Markus Kuhn, Computer Science grad student, Purdue
University, Indiana, US, email: ku...@cs.purdue.edu

ds...@cameonet.cameo.com.tw

unread,
Dec 11, 1996, 3:00:00 AM12/11/96
to

I wrote --

> I have a German document that _seems_ to say seven-bit ASCII was adopted
> in Europe in 1965, three years before the U.S. Anybody who reads German
> willing to take a look at it for me? There is just one paragraph that I'm
> eager to get a full translation of.

And Markus Kuhn (ku...@cs.purdue.edu) responded --

> Just post the paragraph.

How can I refuse the keeper of the International Standards FAQ?
Especially when his name appears in the document I mentioned....

Yes, it refers to ECMA-6, and it's the only source I've seen that gives a
date -- April 1965 -- for the adoption of that standard (it's kinda hard
to get info on these kinds of things when you're in Taiwan). Several
people have already mailed me translations of the passage, and they all
agree: it says that ECMA-6 was approved, passed, promulgated, issued,
published, whatever, in April 1965. Six-bit ASCII came out in '63, but as
far as I can tell, ASA/ANSI (and whatever it called itself in between --
it briefly had another name) didn't approve a seven-bit character
encoding method until '68. Can anyone confirm or refute that?

Thanks for the tips about the ACM discussion and the foreword to the 1991
ECMA-6 standard, Markus. How can I eavesdrop on the ACM discussion?

Here's the passage containing the paragraph that intrigued me. That
paragraph is the second, the one that starts with "Ende April 1965." Many
thanks to those who offered to translate it, and especially to those who
have already sent me translations. I should work so fast....

Dan Strychalski Technical Writer and Translator
ds...@cameonet.cameo.com.tw Hsinchu, Taiwan, Republic of China
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
===========================================================================
2.1 Geschichte der Zeichenkodierungen
===========================================================================

Ich bin mir bewusst, dass diese Auflistung sowohl unvollstaendig als auch
einigermassen willkuerlich ist - man moege mir das nachsehen.

Ende April 1965 wurde ECMA-6 verabschiedet. ECMA-6 ist die 7-bit
Zeichenkodierung, die auch als US-ASCII oder auch als ISO 646 bzw.
als DIN 66003 (Juni 1974) bekannt und weit verbreitet ist.
(Eigentlich sollte man denken, dass der US-ASCII auch aus den USA
stammt, jedoch ist mir keine US-amerikanische Quelle bekannt, die
aelter als April 1965 ist - kann das sein?)

Mit der Zeit wurde den EDV-Treibenden das Umschreiben von Umlauten
zu laestig ;) und sie entwarfen nationale Varianten von ISO 646.
Das fuehrte dazu, dass man sich z. B. in Deutschland zwischen Umlauten
und eckigen sowie geschweiften Klammern entscheiden musste oder mit
Escape-Sequenzen arbeiten.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
P.S. I'd like to post a translation, but I'm not qualified to judge
which of those I've received does the original the most justice.
All agree that ECMA-6 as an official standard dates from 1965,
and that the author of the passage is as curious as I am about
whether the U.S. had a seven-bit standard before that.

Oh, almost forgot. The site and directory path of the document containing
the above passage (put 'em together and you'll have a URL):

http://sunsite.nijenrode.nl
/ftp/packages/rtfm/de.comp.standards/Umlaute_im_deutschsprachigen_Usenet_FAQ

Markus Kuhn

unread,
Dec 11, 1996, 3:00:00 AM12/11/96
to

ds...@cameonet.cameo.com.tw wrote:
>======================================================================
> 2.1 Geschichte der Zeichenkodierungen

>======================================================================
>
> Ich bin mir bewusst, dass diese Auflistung sowohl unvollstaendig als auch
> einigermassen willkuerlich ist - man moege mir das nachsehen.
>
> Ende April 1965 wurde ECMA-6 verabschiedet. ECMA-6 ist die 7-bit
> Zeichenkodierung, die auch als US-ASCII oder auch als ISO 646 bzw.
> als DIN 66003 (Juni 1974) bekannt und weit verbreitet ist.
> (Eigentlich sollte man denken, dass der US-ASCII auch aus den USA
> stammt, jedoch ist mir keine US-amerikanische Quelle bekannt, die
> aelter als April 1965 ist - kann das sein?)
>
> Mit der Zeit wurde den EDV-Treibenden das Umschreiben von Umlauten
> zu laestig ;) und sie entwarfen nationale Varianten von ISO 646.
> Das fuehrte dazu, dass man sich z. B. in Deutschland zwischen Umlauten
> und eckigen sowie geschweiften Klammern entscheiden musste oder mit
> Escape-Sequenzen arbeiten.
> - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Ok, here is my translation:

======
2.1 History of character codes

I am aware that this list is both incomplete and somewhat arbitrary -
please forgive me for this.

In the end of April 1965, ECMA-6 has been finalized. ECMA-6 is the 7-bit
character code, which is known and widely used as US-ASCII or as ISO 646
or DIN 66003 (June 1974), respectively. (One should think that US-ASCII
originated in the USA, however I am not aware of any U.S. source which
predates April 1965 - is this possible?)

[... the rest talks about the missing German Umlaut characters, etc.]
======

Kosta Kostis <ko...@live.robin.de>, the author of the "German Umlaut
Characters in the German parts of USENET FAQ", has a copy of ECMA-6 and
has got this info from the foreword there. I have already forwarded
him Dik's posting with the 1964 reference to the article about the ISO
character set in The Computer Journal 7 (1964), Nr 3, 197-202, but it
still seems that ECMA-6 was the first formal standard describing ASCII
that has been published.

Kosta Kostis

unread,
Dec 12, 1996, 3:00:00 AM12/12/96
to

ds...@cameonet.cameo.com.tw wrote:
> P.S. I'd like to post a translation, but I'm not qualified to judge
> which of those I've received does the original the most justice.

You could send me those translations and I judge them (I wrote that
original German version). ;)

> All agree that ECMA-6 as an official standard dates from 1965,
> and that the author of the passage is as curious as I am about
> whether the U.S. had a seven-bit standard before that.

That's right. ECMA is one of those incredibly friendly organisations
who send you a copy of their standards if you ask them and they even
do so for free (at least they did when I asked some time ago).
My source was the foreword of ECMA-6.

Since I'm also curious about the matter, please keep us informed
about your findings. :)

Best Regards

Kosta
--
kos...@acm.org, ko...@live.robin.de, ko...@blues.sub.de
Kosta Kostis, Talstr. 25, D-63322 Rödermark, Germany
http://ourworld.compuserve.com/homepages/kosta/

Kai Henningsen

unread,
Dec 12, 1996, 3:00:00 AM12/12/96
to

ku...@cs.purdue.edu (Markus Kuhn) wrote on 10.12.96 in <32ADF8...@cs.purdue.edu>:

> ds...@cameonet.cameo.com.tw wrote:
> > I have a German document that _seems_ to say seven-bit ASCII was adopted
> > in Europe in 1965, three years before the U.S. Anybody who reads German
> > willing to take a look at it for me? There is just one paragraph that I'm
> > eager to get a full translation of.
>

> Just post the paragraph.
>
> I guess, it refers to ECMA-6, a standard that you can order for free
> (see www.ecma.ch and ftp.ecma.ch). The modern 1991 ECMA-6
> standard is of course aligned with the modern ISO 646 (the
> US-ASCII as we know it today), but the foreword talks about the
> early history if I remember correctly.

ECMA-6 adopted 30 Apr 65, revised Jun 67, Jul 70, Aug 73
3rd ed gave us ISO 646-1972 and CCITT V.3 "International Telegraph
Alphabet No. 5"

ISO stuff revised 1983 and 1991.

All from that foreword of the 1991 issue of ECMA-6.


Kai
--
Internet: k...@khms.westfalen.de
Bang: major_backbone!khms.westfalen.de!kai
http://www.westfalen.de/private/khms/

Keith F. Lynch

unread,
Dec 15, 1996, 3:00:00 AM12/15/96
to

Robert S. Maier <r...@math.arizona.edu> wrote:
> I distinctly remember seeing, back in the early 1980's, an old
> character-cell terminal that represented the ^ character (caret,
> octal \136) by an up-arrow, and the _ character (underscore, \137)
> by a right-arrow. The terminal must have dated back to the early
> 1970's.

Are you sure the underscore wasn't a LEFT-arrow?

According to Eric Raymond's _The New Hacker Dictionary_, those
characters changed from up-arrow to caret, and from left-arrow
to underscore, in June 1966.

(Posted and mailed.)
--
Keith Lynch, k...@access.digex.net
http://www.access.digex.net/~kfl/

Tim Shoppa

unread,
Dec 15, 1996, 3:00:00 AM12/15/96
to

In article <591p1s$k...@access2.digex.net>,

Keith F. Lynch <k...@access2.digex.net> wrote:
>Robert S. Maier <r...@math.arizona.edu> wrote:
>> I distinctly remember seeing, back in the early 1980's, an old
>> character-cell terminal that represented the ^ character (caret,
>> octal \136) by an up-arrow, and the _ character (underscore, \137)
>> by a right-arrow. The terminal must have dated back to the early
>> 1970's.
>
>Are you sure the underscore wasn't a LEFT-arrow?
>
>According to Eric Raymond's _The New Hacker Dictionary_, those
>characters changed from up-arrow to caret, and from left-arrow
>to underscore, in June 1966.

I pulled out one of my old Model 33 print sets and there were all
sorts of print variations available: left arrow vs right arrow,
caret vs up-arrow, slashed letter "O" vs slashed numeral "0",
versions where "[" and "]" are replaced by the Paragraph symbol
and the pound sign, etc. Undoubtedly all these were produced
not because of any "standard" but because "the customer paid for
it."

Tim. (sho...@triumf.ca)

ds...@cameonet.cameo.com.tw

unread,
Dec 16, 1996, 3:00:00 AM12/16/96
to

Kosta Kostis (ko...@live.robin.de) wrote --

> ds...@cameonet.cameo.com.tw wrote:
> > P.S. I'd like to post a translation, but I'm not qualified to judge
> > which of those I've received does the original the most justice.
>
> You could send me those translations and I judge them (I wrote that
> original German version). ;)

Done!

Another big thank-you to the folks who sent me translations. They have
graciously given me permission to pass their renderings on, and who would
be a better recipient than the author himself? Thanks also to those who
offered but couldn't get around to it. I know how that is.

Following is a broad outline of what I've been able to find out (and
surmise) about the early history of ASCII (some of the abbreviations may
be unfamiliar; I'll explain them further on):

Standards organizations began turning their attention to information
technology in 1960. At the international level, ISO initiated plans to
set up TC 97. In the U.S., BEMA, recognized by the ASA as the sponsor
for standardization work in data processing, brought together a number
of companies to form a data processing group; in 1961 this group set
up the committee known as X3.

Immediately after starting operation, TC 97 set up six working groups,
and X3 set up six subcommittees. Both sets of subdivisions included
bodies for the standardization of data transmission methods,
programming languages, computer-related terminology, character
recognition techniques, and character encoding practices.

As part of an international organization with three official
languages, TC 97 needed first to define the terms in which standards
would be described. At the top of the list of TC 97's subdivisions was
Working Group A, Glossary.

X3, organized by BEMA in accordance with ASA guidelines, included
representatives of producer, consumer, and "general interest" groups
in approximately equal proportions. Commercial providers and users of
data processing goods and services made up almost two thirds of X3's
membership. At the top of the list of X3's subdivisions was Standards
Subcommittee X3.1, Character Recognition.

Only one pair of subdivisions had both the same area of responsibility
and the same apparent rank in both hierarchies: TC 97 Working Group B,
Character Sets and Coding, and Standards Subcommittee X3.2, Coded
Character Sets.

It seems clear that both ISO and ASA attached great importance to the
standardization of character codes, and that big business had a big
influence on ASA's priorities....

X3's job was to develop U.S. standards, submit them to ISO, and develop
U.S. positions on proposals coming from ISO. The original X3 committee
was made up of representatives from --

10 manufacturers, one of which was Honeywell Minneapolis.
10 consumer groups, including the Air Transport Association and the
American Bankers Association.
11 general interest groups, including the Association for Computing
Machinery, the Department of Defense, and the American Management
Association.

Subcommittee X3.2 worked on coded character sets; the task group working
on ASCII appears to have been designated X3.2.4. Membership in X3.2 and
X3.2.4 was based on technical qualifications rather than affiliation with
any of the X3 member bodies.

Other than that, basically all I know is that ASA approved six-bit ASCII
in 1963, ECMA passed the seven-bit ECMA-6 standard in 1965, and USASI
published a seven-bit version of ASCII in 1968. Now for those
abbreviations and some related history:

ISO The International Organization for Standardization. Founded in
1947; based in Geneva. ISO is made up of "member bodies," one
per country. The name "ISO" is derived from the Greek "isos,"
meaning "equal," and is pronounced EYE-so in English.

TC 97 ISO technical committee on computers and information processing.
Now dissolved. Its work is carried on by ISO/IEC JTC 1, the
Joint Technical Committee of ISO and the International
Electrotechnical Commission.

ASA The American Standards Association. Became the United States of
America Standards Institute (USASI) in 1966 and the American
National Standards Institute (ANSI) in 1969. U.S. member body of
ISO; private and non-profit.

BEMA Business Equipment Manufacturers Association (U.S.). Later
became the Computer and Business Equipment Manufacturers
Association (CBEMA); known since 1994 as the Information
Technology Industries Council (ITI).

X3 ASA sectional committee on computers and data processing. Later
called American National Standards Committee X3 (ANSC-X3); now
known as Accredited Standards Committee X3 (ASC X3). Accredited
by ANSI and administered by ITI.

Also in the early sixties, IBM formulated a single, representative (I'd
rather avoid words like "standard" and "official") version of EBCDIC;
previously, each department or project had used a different version.
Whether IBM tried to get EBCDIC approved by ASA I have no idea. I do know
that most of the values in the control range (18 of the 32, if I recall
correctly) have identical definitions in ASCII and EBCDIC.

It looks like ASA was bending over backward to please IBM, but was
prepared to go only so far. And we all know that IBM didn't make an ASCII
machine until 1981....

So what was going on in ASA/USASI between 1963 and 1968?

Dan Strychalski
ds...@cameonet.cameo.com.tw

John Everett

unread,
Dec 16, 1996, 3:00:00 AM12/16/96
to

In article <591p1s$k...@access2.digex.net>, k...@access2.digex.net says...

>
>Robert S. Maier <r...@math.arizona.edu> wrote:
>> I distinctly remember seeing, back in the early 1980's, an old
>> character-cell terminal that represented the ^ character (caret,
>> octal \136) by an up-arrow, and the _ character (underscore, \137)
>> by a right-arrow. The terminal must have dated back to the early
>> 1970's.
>
>Are you sure the underscore wasn't a LEFT-arrow?
>

Back in the mid '60s, most of DEC's software depended on the presence of a
left-arrow. The PDP-6 PIP syntax used a command string like:

RESULT.FIL<-FILE.FOO,FILE.BAR

to combine two files, only the "<-" was a single character left-arrow; or as
we called it then, a backarrow. When presented with a TTY with an alternate
keyboard we substituted the underscore key. It was only when terminals
without backarrows became common that we changed the CUSPs (anyone still
remember that term) to accept the equals sign in place of the backarrow. At
least with an arrow it was clear which argument was source and which
destination.

TOPS-10 had a COPY command which invoked PIP. When I first started to use
DOS, I had a terrible time getting COPY to do what I meant.

--
jeve...@wwa.com (John V. Everett) http://www.wwa.com/~jeverett


Johnny Billquist

unread,
Dec 17, 1996, 3:00:00 AM12/17/96
to

In <59406n$6...@kirin.wwa.com> jeve...@wwa.com.defeat.uce.bots (John Everett) writes:
>
>Back in the mid '60s, most of DEC's software depended on the presence of a
>left-arrow. The PDP-6 PIP syntax used a command string like:
>
>RESULT.FIL<-FILE.FOO,FILE.BAR
>
>to combine two files, only the "<-" was a single character left-arrow; or as
>we called it then, a backarrow. When presented with a TTY with an alternate
>keyboard we substituted the underscore key. It was only when terminals
>without backarrows became common that we changed the CUSPs (anyone still
>remember that term) to accept the equals sign in place of the backarrow. At
>least with an arrow it was clear which argument was source and which
>destination.

CUSPs... Sure. Still use 'em. :-)

Anyway, you skipped a step. The change wasn't immediately from the backarrow
to the equal sign. In between, dec used the '<' instead of the backarrow
in CUSPs.

It's still the character you use in OS/8, if you don't feel like using
the underscore (unless you have a real TTY).

Now, how many knows what a TTY *really* is? :-)

Johnny
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: b...@update.uu.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol

J. Chris Hausler

unread,
Dec 17, 1996, 3:00:00 AM12/17/96
to

Keith F. Lynch <k...@access2.digex.net> writes:

>Robert S. Maier <r...@math.arizona.edu> wrote:
>> I distinctly remember seeing, back in the early 1980's, an old
>> character-cell terminal that represented the ^ character (caret,
>> octal \136) by an up-arrow, and the _ character (underscore, \137)
>> by a right-arrow. The terminal must have dated back to the early
>> 1970's.
>
>Are you sure the underscore wasn't a LEFT-arrow?


>According to Eric Raymond's _The New Hacker Dictionary_, those
>characters changed from up-arrow to caret, and from left-arrow
>to underscore, in June 1966.

The first language I learned to program, Algol, used both of these
characters on the TTY, the up-arrow for exponentiation and the
left-arrow for assignment ie A<-B+C; meaning the sum of the values
B and C is stored into A. I also recall a Basic implementation
which used the up-arrow to indicate an exponent.
Chris

Magnus Olsson

unread,
Dec 17, 1996, 3:00:00 AM12/17/96
to

In article <RlGVTK-....@delphi.com>,

J. Chris Hausler <jcha...@delphi.com> wrote:
>Keith F. Lynch <k...@access2.digex.net> writes:
>
>>Robert S. Maier <r...@math.arizona.edu> wrote:
>>> I distinctly remember seeing, back in the early 1980's, an old
>>> character-cell terminal that represented the ^ character (caret,
>>> octal \136) by an up-arrow, and the _ character (underscore, \137)
>>> by a right-arrow. The terminal must have dated back to the early
>>> 1970's.
>>
>>Are you sure the underscore wasn't a LEFT-arrow?
>
>
>>According to Eric Raymond's _The New Hacker Dictionary_, those
>>characters changed from up-arrow to caret, and from left-arrow
>>to underscore, in June 1966.
>
>The first language I learned to program, Algol, used both of these
>characters on the TTY, the up-arrow for exponentiation and the
>left-arrow for assignment ie A<-B+C; meaning the sum of the values
>B and C is stored into A.

Actually, Algol 60 (which I assume is what you're talking about) is
two languages: the "definition language" and the "implementation
language". The definition language uses a character set not often
seen on terminals: boldface for reserved words, a single := character
for assigment, a below-the-line "10" (also a single character) to
denote "10 to the power of (where Fortran, C and Pascal use 'E'),
special characters for the Boolean operators, and so on.

Each implementation of Algol 60 can have its own mapping of these characters
to ASCII, EBCDIC or whatever. Early implementations could use the backarrow
for assignment. When backarrow was replaced by underscore, implementors
switched to the two-character combination := (still used by Pascal, Ada
and other languages).

However, some compilers didn't bother to switch, and used underscores for
assignments. The resulting "a _ b + c;" looks pretty strange...

> I also recall a Basic implementation
>which used the up-arrow to indicate an exponent.

And this carried over to modern Basics, which usually use the caret for
exponentiation.

Actually, Motorola's 6847 video controller chip (ca 1980) had a built-in
character set which used the up arrow rather than the caret. So there
were computers using the up arrow well into the 1980's.


--
Magnus Olsson (m...@df.lth.se)

J. Chris Hausler

unread,
Dec 18, 1996, 3:00:00 AM12/18/96
to

Magnus Olsson <m...@bartlet.df.lth.se> writes:

>Actually, Algol 60 (which I assume is what you're talking about) is
>two languages: the "definition language" and the "implementation
>language". The definition language uses a character set not often
>seen on terminals: boldface for reserved words, a single := character
>for assigment, a below-the-line "10" (also a single character) to
>denote "10 to the power of (where Fortran, C and Pascal use 'E'),
>special characters for the Boolean operators, and so on.

Hmmm... I don't remember the "10". Somewhere I have a formal BNF
description of Algol 60 I'll have to check out. The "implementation
language" in my case was "Algol 20" running on Bendix G-20's.


>Each implementation of Algol 60 can have its own mapping of these characters
>to ASCII, EBCDIC or whatever. Early implementations could use the backarrow
>for assignment. When backarrow was replaced by underscore, implementors
>switched to the two-character combination := (still used by Pascal, Ada
>and other languages).
>
>However, some compilers didn't bother to switch, and used underscores for
>assignments. The resulting "a _ b + c;" looks pretty strange...

I remember (I think :-) that the Algol for the UNIVAC 1108 would
accept the two character sequence := and since IIRC the keypunches
we used did not have colons, it would accept the three character
sequence ..= (lazy colon :-) as well. However it would also
accept the = alone and everyone I knew used this. BTW the three
character sequence ..= was converted to := when printed on the
line printer.


>And this carried over to modern Basics, which usually use the caret for
>exponentiation.
>
>Actually, Motorola's 6847 video controller chip (ca 1980) had a built-in
>character set which used the up arrow rather than the caret. So there
>were computers using the up arrow well into the 1980's.

Interesting. I have a couple RS COCO II's which use this chip.
I'll have to wake one of them up and see what the ROMed basic does
when I try to represent exponentiation.

Chris

Richard S. Shuford

unread,
Dec 20, 1996, 3:00:00 AM12/20/96
to

In message <592eag$l...@reader.seed.net.tw> at 16 Dec 1996 03:05:52 GMT,
Dan Strychalski <ds...@cameonet.cameo.com.tw> wrote:
>
> ...the early history of ASCII...
>
> ...in 1960. At the international level, ISO initiated plans to

> set up TC 97. In the U.S., BEMA, recognized by the ASA as the sponsor
> for standardization work in data processing, brought together a number
> of companies to form a data processing group; in 1961 this group set
> up the committee known as X3.
>
> ...

>
> X3, organized by BEMA in accordance with ASA guidelines, included
> representatives of producer, consumer, and "general interest" groups
> in approximately equal proportions. Commercial providers and users of
> data processing goods and services made up almost two thirds of X3's
> membership....

>
> Only one pair of subdivisions had both the same area of responsibility
> and the same apparent rank in both hierarchies: TC 97 Working Group B,
> Character Sets and Coding, and Standards Subcommittee X3.2, Coded
> Character Sets.

At least as of today, I think this is actually known as X3L2. Have a look at:

http://www.x3.org/tc_home/x3l2.htm

The bureaucracy of international standardization is formidable!
A description from one vendor's point of view may be found at:

http://www.unisys.com/TechnologySolutions/WhitePapers/07_s02b.html

> ...


>
> Other than that, basically all I know is that ASA approved six-bit ASCII
> in 1963, ECMA passed the seven-bit ECMA-6 standard in 1965, and USASI
> published a seven-bit version of ASCII in 1968.


I think some part of the development of ASCII should be credited to
the Teletype Corporation (Skokie, Illinois). I can't say authoritatively
how much, however, you can read some discussion at:

http://www.cs.utk.edu/~shuford/terminal/teletype_news.txt


> It looks like ASA was bending over backward to please IBM, but was
> prepared to go only so far.


IBM historically has played the standards "game" very well.


> And we all know that IBM didn't make an ASCII
> machine until 1981....
> So what was going on in ASA/USASI between 1963 and 1968?


Ah, there is one very important development in ASCII that I have not seen
mentioned in this discussion.

In about 1965 or 1966, President Lyndon B. Johnson issued an Executive
Order requiring that all computer systems purchased by the United States
federal government be capable of communicating in ASCII.

(Years ago, in the appendix to a large book published by McGraw-Hill--
maybe it was "Encyclopedia of Data Processing"--I saw a facsimile of LBJ's
Executive Order mandating ASCII.)

I assume, but don't know for a certainty, that the Seven Dwarfs computer
manufacturers (see below) were the parties who instigated this. The U.S.
government had become very dependent on the mainframe products of IBM, all
of which were EBCDIC-speaking.

(Rumor says that IBM relocated its headquarters to Armonk, New York, as a
patriotic gesture, so that when Manhattan was leveled in the anticipated
nuclear war, IBM would still be intact to help the nation rebuild.)

But with ASCII mandated for federal government use, it was somewhat easier
for the Seven to win procurement competitions. IBM had to submit bids
that mostly included kluge measures, such as the 7171 Protocol Converter,
to translate from ASCII to EBCDIC. (I think that IBM did build some
special-purpose computers that used ASCII--before the famous model 5150!)

A book that might shed more light on this subject is

"Coded Character Sets: History and Development" by C. E. MacKenzie.
Reading, Massachusetts: Addison-Wesley, 1980.

If you don't remember an IBM product number of "5150", you may recall the
machines themselves: they came from a division in Boca Raton. :-)

--- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- ---

* The Seven Dwarfs were those intrepid enterprises that dared to
challenge IBM in selling data-processing machinery: General Electric,
RCA, Sperry Univac, Honeywell, Burroughs, Scientific Data Systems,
and Control Data Corporation

One recent book tells the story of the largest of the Seven:

http://www.computer.org/cspress/catalog/bp07383.htm
or http://www.davison.net/cgi-bin/vlink/0818673834

Read the preface:

http://www.computer.org/cspress/catalog/bp07383/dwf-pre.htm

Actually, some of the Seven (maybe Honeywell?) did build some EBCDIC
computers themselves.

--- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- ---

A variety of video-terminal-related information may be found at:

http://www.cs.utk.edu/~shuford/terminal_index.html

--- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- ---

...Richard S. Shuford |"He who oppresses the poor to increase his wealth and
...shu...@cs.utk.edu | he who gives gifts to the rich--both come to poverty.
...Info-Stratus contact| Proverbs 22:16 NIV

Heinz W. Wiggeshoff

unread,
Dec 20, 1996, 3:00:00 AM12/20/96
to

"Richard S. Shuford" (shu...@cs.utk.edu) writes:
>
> Actually, some of the Seven (maybe Honeywell?) did build some EBCDIC
> computers themselves.

In the summer of 1980, I did some IBM System /360 assembler on a
Univac 9xxx machine. Damn hot summer too, here in Ottawa! B-)

Gord Campbell

unread,
Dec 20, 1996, 3:00:00 AM12/20/96
to

Quoting Richard S. Shuford:

RS>* The Seven Dwarfs were those intrepid enterprises that dared to
RS> challenge IBM in selling data-processing machinery: General Electric,
RS> RCA, Sperry Univac, Honeywell, Burroughs, Scientific Data Systems,
RS> and Control Data Corporation

RS>Actually, some of the Seven (maybe Honeywell?) did build some EBCDIC
RS>computers themselves.

RCA for sure.
---
ş DeLuxeı 1.25 #7680 ş Where service isn't just a word; it's a noun.

Marco S Hyman

unread,
Dec 20, 1996, 3:00:00 AM12/20/96
to

shu...@cs.utk.edu said:

> In about 1965 or 1966, President Lyndon B. Johnson issued an Executive
> Order requiring that all computer systems purchased by the United States
> federal government be capable of communicating in ASCII.

From FIPS PUB 7, Dated March 7, 1969:

-------------
PURPOSE.--To provide further details concerning the implementation and
applicability of the Federal Information Processing Standards (FIPS), Code
for Information Interchange (FIPS 1), Perforated Tape Code for Information
Interchange (FIPS 2), and Recorded Magnetic Tape for Information Interchange
(800 CPI, NRZI)(FIPS 3).

EXPLANATION.--White House memorandum to heads of departments and agencies,
dated March 11, 1968, signed by President Lyndon B. Johnson approved as
Federal Stabndards the United States of America Standard Code for
Information Interchange and assiciated standards for recording the code
on perforated and magnetic tape media. ...
-------------

A copy of the memoranda is attached to this FIPS. If anyone *has* to
have it I suppose I could type it in.

> (Years ago, in the appendix to a large book published by McGraw-Hill--
> maybe it was "Encyclopedia of Data Processing"--I saw a facsimile of LBJ's
> Executive Order mandating ASCII.)

I'm using "McGraw Hill's Compilation of DATA COMMUNCATIONS STANDARDS,
Edition II" dated 1982. It's just under 2000 pages.

// marc

Marco S Hyman

unread,
Dec 20, 1996, 3:00:00 AM12/20/96
to

Quoting Richard S. Shuford:

RS>* The Seven Dwarfs were those intrepid enterprises that dared to
RS> challenge IBM in selling data-processing machinery: General Electric,
RS> RCA, Sperry Univac, Honeywell, Burroughs, Scientific Data Systems,
RS> and Control Data Corporation

So when did the Seven Dwarfs turn into the BUNCH (Burroughs, Univac,
NCR, CDC, and Honeywell)?

// marc

Heinz W. Wiggeshoff

unread,
Dec 21, 1996, 3:00:00 AM12/21/96
to

Marco S Hyman (ma...@dumbcat.codewright.com) writes:
>
> So when did the Seven Dwarfs turn into the BUNCH (Burroughs, Univac,
> NCR, CDC, and Honeywell)?

About the time that Fairchild Semiconductors merged with Honeywell
Computers to form Fairwell Honeychild?


Nick Spalding

unread,
Dec 21, 1996, 3:00:00 AM12/21/96
to

."Richard S. Shuford" <shu...@cs.utk.edu> wrote:
>
> Ah, there is one very important development in ASCII that I have not seen
> mentioned in this discussion.
>
> In about 1965 or 1966, President Lyndon B. Johnson issued an Executive
> Order requiring that all computer systems purchased by the United States
> federal government be capable of communicating in ASCII.
>
360 was announced in 1964 (May?) and there was a 'Use ASCII' bit in
the PSW from the beginning wasn't there?
--
Nick Spalding

William G. Royds

unread,
Dec 21, 1996, 3:00:00 AM12/21/96
to Bill_...@pch.gc.ca

Nick Spalding wrote:
>
> ."Richard S. Shuford" <shu...@cs.utk.edu> wrote:
> >
> > Ah, there is one very important development in ASCII that I have not seen
> > mentioned in this discussion.
> >
> > In about 1965 or 1966, President Lyndon B. Johnson issued an Executive
> > Order requiring that all computer systems purchased by the United States
> > federal government be capable of communicating in ASCII.
> >
> 360 was announced in 1964 (May?) and there was a 'Use ASCII' bit in
> the PSW from the beginning wasn't there?
> --
> Nick Spalding The IBM 360 series had a bit for ASCII/EBCDIC (bit 12) that was 0 for EBCDIC and was
initialized to EBCDIC. THe problem was that much of the OS (DOS) would not run properly
in ASCII mode. The 370 series removed its meaning since it was not used. It seemed to be
there only to allow IBM computers to be sold to U.S. government, but then ignored.

HAncock4

unread,
Dec 21, 1996, 3:00:00 AM12/21/96
to

> RS>* The Seven Dwarfs were those intrepid enterprises that dared to
> RS> challenge IBM in selling data-processing machinery: General Electric,
> RS> RCA, Sperry Univac, Honeywell, Burroughs, Scientific Data Systems,
> RS> and Control Data Corporation

In the early 1950s, IBM was quite nervous about this competition. Data
processing was IBM's lifeblood (time clocks and typewriters were marginal).
The other companies had very substantial resources in other businesses--
Sperry Univac (originally Remington Rand) and Burroughs were major office
products suppliers.

IBM won out because it HAD the data processing experience and a major foot
in the door from its tab machines. Further, its computers were a logical
extension of existing tab machines.

It should be noted that while electronic computers got the headlines in
the early 1950s, there were very, very few of them and they were very
expensive. The automated workhorse in those days was the punch card
machines, including the IBM 604 calculator which was a poor man's computer.

Nick Spalding

unread,
Dec 22, 1996, 3:00:00 AM12/22/96
to

.az...@freenet.carleton.ca (William G. Royds) wrote:
> The IBM 360 series had a bit for ASCII/EBCDIC (bit 12) that was 0 for EBCDIC and was
> initialized to EBCDIC. THe problem was that much of the OS (DOS) would not run
> properly
> in ASCII mode. The 370 series removed its meaning since it was not used. It seemed to > be there only to allow IBM computers to be sold to U.S. government, but then ignored.

The point I was trying to make was that this was 1964, before the
1965/66 date the previous poster said was when such a requirement came
about. The fact that it never actually worked is neither here nor
there.
--
Nick Spalding

Michael D Shapiro

unread,
Dec 23, 1996, 3:00:00 AM12/23/96
to

In article <591rjs$s...@nntp.ucs.ubc.ca> sho...@alph02.triumf.ca (Tim Shoppa) writes:
>In article <591p1s$k...@access2.digex.net>,
>Keith F. Lynch <k...@access2.digex.net> wrote:
>>Robert S. Maier <r...@math.arizona.edu> wrote:
>>> I distinctly remember seeing, back in the early 1980's, an old
>>> character-cell terminal that represented the ^ character (caret,
>>> octal \136) by an up-arrow, and the _ character (underscore, \137)
>>> by a right-arrow. The terminal must have dated back to the early
>>> 1970's.
>>
>>Are you sure the underscore wasn't a LEFT-arrow?
>>
>>According to Eric Raymond's _The New Hacker Dictionary_, those
>>characters changed from up-arrow to caret, and from left-arrow
>>to underscore, in June 1966.
>
>I pulled out one of my old Model 33 print sets and there were all
>sorts of print variations available: left arrow vs right arrow,
>caret vs up-arrow, slashed letter "O" vs slashed numeral "0",
>versions where "[" and "]" are replaced by the Paragraph symbol
>and the pound sign, etc. Undoubtedly all these were produced
>not because of any "standard" but because "the customer paid for
>it."
>
>Tim. (sho...@triumf.ca)

The slashed zero versus oh controversy goes back to data versus text
input. Data input assumed many more zeros than ohs so the convention
in handwritten text was to slash the oh. Text assumed more ohs than
zeros so slashed the zeros in handwritten text. You could buy your
Teletype terminal with either convention.
--
Michael D. Shapiro, Ph.D. Internet: msha...@nosc.mil
Code 4123, NCCOSC RDT&E Division (NRaD) San Diego CA 92152
Voice: (619) 553-4080 FAX: (619) 553-4808 DSN: 553-4080

Michael D Shapiro

unread,
Dec 28, 1996, 3:00:00 AM12/28/96
to

In article <32c5eb10....@news.iol.ie>,

Nick Spalding <spal...@iol.ie> wrote:
>."Richard S. Shuford" <shu...@cs.utk.edu> wrote:
>>
>> Ah, there is one very important development in ASCII that I have not seen
>> mentioned in this discussion.
>>
>> In about 1965 or 1966, President Lyndon B. Johnson issued an Executive
>> Order requiring that all computer systems purchased by the United States
>> federal government be capable of communicating in ASCII.
>>
>360 was announced in 1964 (May?) and there was a 'Use ASCII' bit in
>the PSW from the beginning wasn't there?

That ASCII mode in the 360 wasn't really ASCII code. It was a way of
encoding certain information with non-contiguous values that looked
something like ASCII. It flopped and disappeared later.

Michael D Shapiro

unread,
Dec 28, 1996, 3:00:00 AM12/28/96
to

Through the 1960s, most mainframe computers did not use ASCII as an
internal code. They used some internal code and translated to or from
it as part of the input and output process. I believe the first
mainframe to use ASCII as its internal code was the NCR Century series.

The reason for this choice was that most mainframes were used for data
processing, and this was punched card based for the most part.
Mini-computers, when they arrived, were more communications oriented
and so used ASCII, a communications code.

Dr. Peter Kittel

unread,
Dec 28, 1996, 3:00:00 AM12/28/96
to

In article <BlMWL9D....@delphi.com> J. Chris Hausler <jcha...@delphi.com> writes:
>Magnus Olsson <m...@bartlet.df.lth.se> writes:
>
>>Actually, Algol 60 (which I assume is what you're talking about) is
>>two languages: the "definition language" and the "implementation
>>language". The definition language uses a character set not often
>>seen on terminals: boldface for reserved words, a single := character
>>for assigment, a below-the-line "10" (also a single character) to
>>denote "10 to the power of (where Fortran, C and Pascal use 'E'),
>>special characters for the Boolean operators, and so on.
>
>Hmmm... I don't remember the "10".

Oh, we had it, too, on our Electrologica X1 made by Philips in the
early 60's and running from 5-channel paper punched tape. We had old
reworked telex machines to punch our tapes. The rework consisted of
exactly exchanging some of the characters from normal teletex use
to things like the base 10 and square brackets. We used ":=" as two
characters and had no single one for it. Later, on punched cards for
an ICL mainframe, they had a single "<-" sign which could be used for
such an assignment.

> However it would also
>accept the = alone and everyone I knew used this.

Oops, no, our compiler didn't allow this.

>>Actually, Motorola's 6847 video controller chip (ca 1980) had a built-in
>>character set which used the up arrow rather than the caret.

Yup. Though I know it from the 6845 and its Commodore version 6545,
do you mean the same?

--
Best Regards, Dr. Peter Kittel // http://www.pios.de
Private Site in Frankfurt, Germany \X/ office: pet...@pios.de
Einen guten Rutsch! A happy new year!

Don Payette

unread,
Jan 7, 1997, 3:00:00 AM1/7/97
to

"Richard S. Shuford" <shu...@cs.utk.edu> wrote:

>Actually, some of the Seven (maybe Honeywell?) did build some EBCDIC
>computers themselves.

Unisys (Burroughs) machines have been EBCDIC for years. In fact
I don't know that Burroughs ever made an ASCII machine.


Don.P...@mv.unisys.com
I speak only for myself; not my employer


Douglas W. Jones,201H MLH,3193350740,3193382879

unread,
Jan 8, 1997, 3:00:00 AM1/8/97
to

"Richard S. Shuford" <shu...@cs.utk.edu> wrote:

>Actually, some of the Seven (maybe Honeywell?) did build some EBCDIC
>computers themselves.

The SDS/XDS Sigma 7 was an EBCDIC machine with a strong IBM 360 flavor
but a far better virtual memory mechanism at a time when IBM was having
problems with that.

Doug Jones
jo...@cs.uiowa.edu

Reply all
Reply to author
Forward
0 new messages