[GreenKeys] Teletype ASR33 - approximate date of introduction: 1963

[Jim Haynes]

ASCII was the work of a lot of people and organizations.  A good summary
is an article by Fred Smith in Western Union Technical Review 18:2,
April 1964, p. 50.  Another reference is to R. W. Bemer, Datamation,
August 1963 and September 1963, and Honeywell Computer Journal, 1972.

Baudot (actually Murray) of course we have had forever.  The computer
industry had a number of 6-bit codes more or less related to punched
card codes.  The Signal Corps had developed Fieldata as an attempt
to standardize a code suitable for both communication and computing.
The American Standards Assn. in the U.S. and the International Standards
Organization in Europe both started work on a new code around 1960.

The computer industry tends to regard Bob Bemer as the "father of ASCII"
and indeed he had a car vanity license plate "ASCII".  The Bell System
tended to regard John Auwaerter of Teletype as the "father of ASCII"

There were an awful lot of issues.

1. Eventual support for upper and lower case alphabets.

2. Space for additional alphabetic characters used in some languages.

3. You'd like to have the decimal digits represented in binary
numbers.  You'd also like to have the alphabet represented in proper
sequence in binary numbers.

4. However all the codes based on IBM punched cards had gaps in the
middle of the alphabet, with other characters in there, when they were
converted to binary numbers in a simple way.  This was because
A through J were represented as zero through nine with some "zone
punches" to distinguish them from digits.  Then K through T were
represented as zero through nine with other zone punches, and so on.]
So in a binary representation there are ten alphabetic characters and
then there are six more bit combinations used for something else and
then the alphabet begins again with zero.

5. Typewriter keyboarding - There was a certain amount of standardization
already about which typewriter keys went with which punctuation marks.
So the new code had to try to preserve these associations, while keeping
the binary codes for the upper and lower case characters the same
except for a bit or two.  Ultimately there were two keyboard standards,
called "typewriter paired" and "bit paired" that were close to standard
typewriter usage in the first case or easy to implement in a mechanical
keyboard in the second case.

6.  And others that Bemer goes into.

The ASA committee working on all this was necessarily large, since IBM
and the other computer companies were concerned, as was the Bell System
and all the other communication companies, as was Teletype and other
equipment manufacturers, and the Department of Defense and other
government agencies.  It was very hard to get a consensus, and especially
with IBM playing the role of 800-pound gorilla and insisting on easy
conversion between punched card code and the new character code, which
would have required gaps in the alphabet.  Nearly everybody else
wanted a code with the entire alphabet contiguous in binary, to facilitate
sorting by using the codes as binary numbers.

Suddenly IBM withdrew its objections and the new code came together and
was standardized as a 7-bit code.  Then in 1964 IBM announced its new
System/360 computer line and, guess what?, a new code called Extended
Binary Coded Decimal Interchange Code EBCDIC as its new standard.
EBCDIC has the gaps in the alphabet.  Having failed to prevail in the
standards committee work, IBM hoped to use its dominant position in
the computer industry to establish EBCDIC as a de-facto standard and
have ASCII go awy.

It all seems pretty silly now.  A mere 256 bytes of memory can map any
8-bit character into any other.  But when IBM was producing computers
with only 1400 bytes of memory total it must have been a big issue.




jhhaynes at earthlink dot net