5.0149 Responses: etaoinshrdlu (4/100)

Elaine Brennan & Allen Renear (EDITORS@BROWNVM.BITNET)
Tue, 18 Jun 91 10:41:54 EDT

Humanist Discussion Group, Vol. 5, No. 0149. Tuesday, 18 Jun 1991.

(1) Date: Fri, 14 Jun 91 0:31:56 EDT (25 lines)
From: "Paul N. Banks" <pbanks@cunixa.cc.columbia.edu>
Subject: Useless information

(2) Date: Fri, 14 Jun 91 08:37:57 MDT (9 lines)
From: koontz@alpha (John E. Koontz)
Subject: Re: 5.0139 Rs: Letter Freq.

(3) Date: 13 Jun 91 21:44:26 EST (25 lines)
From: James O'Donnell <JODONNEL@PENNSAS>

(4) Date: Thu, 13 Jun 91 19:36:52 LCL (41 lines)
From: "Dana Cartwright, Syracuse Univ" <DECARTWR@SUVM>

(1) --------------------------------------------------------------------
Date: Fri, 14 Jun 91 0:31:56 EDT
From: "Paul N. Banks" <pbanks@cunixa.cc.columbia.edu>
Subject: Useless information

Robin Smith asks if there is a more useless bit of information than the
origin of etaoin shrdlu. Yes!

For whatever reason, the Linotype's keyboard (and those of Intertypes
and similar machines) was oriented vertically rather than (as a
typewriter or computer keyboard) vertically. The equivalent of the
spacebar was on the left, operated with the left little finger, and
etaoin was the first vertical row of keys, shrdlu the second, and so
forth. So in fact these were the left-most rather than the top two rows
of the keyboard as Smith states.

There was no such thing as touch typing, in the sense of a "home row" on
which one rests fingers for orientation for the rest of the keyboard.
Keyboarding on these machines was more like playing the piano--totally

Paul N. Banks | Conservation Education Programs
Research Scholar | School of Library Service
pbanks@cunixa.cc.columbia.edu | 516 Butler Library
212 854-4445 | Columbia University
212 865-1304 | New York NY 10027
(2) --------------------------------------------------------------19----
Date: Fri, 14 Jun 91 08:37:57 MDT
From: koontz@alpha (John E. Koontz)
Subject: Re: 5.0139 Rs: Letter Freq.; E-Enc.; Mail; Nota Bene; Grammar

Robin Smith reports that etaoinshrdlu is the not the frequency list for
letters in English, but rather a string derived from the keyboard of the
linotype. As far as I can recall from reading elementary books on cypher
when I was a kind, it is also the start of (a) frequency list.

(3) --------------------------------------------------------------25----
Date: 13 Jun 91 21:44:26 EST
From: James O'Donnell <JODONNEL@PENNSAS>

Now funny that Robin Smith took up the, strictly irrelevant, point about
English letter frequency. When I saw the original query, I got to
wondering whether ETAOINSHRDLU was a myth or not and set about testing
it on some of my own prose with the help of the computer. Simple matter
to insert blank spaces between every letter in a sample of English prose
(my own), alpha-sort, then count. The results were *very* close (given
that the sample was only about 30,000 char.) to the conventional list:
N was quite a bit lower, H a bit lower, and F rather higher, but ETAOIS
was definitely the top six. Someone must have a larger database and
more accurate (and more recent) information, but this is all of great
use for any of us who wind up as contestants on *Wheel of Fortune*. As
to the linotype arrangement, my guess is that this represented
somebody's guess a hundred years ago as to order of frequency.

QWERTYUIOP, on the other hand, is famous as the end-result of a very
non-scientific fiddle to try to put the most common letters in the most
convenient places, complicated by having no accurate idea as to the most
common letters and no consistent idea as to the most convenient places.
The right index finger, arguably the strongest and most coordinated, has
little work to do on ETAOINSHRDLU, at any rate, and the left hand is
uncommonly busy.
(4) --------------------------------------------------------------48----
Date: Thu, 13 Jun 91 19:36:52 LCL
From: "Dana Cartwright, Syracuse Univ, 315-443-4504" <DECARTWR@SUVM>

The story I was told to explain the sequence etaoinshrdlu (and in
my mind this is no more than a story, I have no direct way of verifying
it) is that it is rooted in the Linotype machine and some of its
peculiarities of operation.

As I understand it, as an operator "types" at a keyboard, small molds
for the corresponding letters are assembled into a single line of text.
At the end of the line, the operator presses what would be equivalent to
what we call "return," whereupon the machine pours molten typemetal over
the row of molds, forming a solid bar (slug) of typemetal for a single
line of text. The molds are broken free and recirculated into the

Now, as each key on the keyboard is pressed, the mold literally *drops*
into position. Once in place, a mold cannot be individually retrieved.
And the assembly place is buried inside a maze of hot metal. In short:
there is no "backspace" or "oops" key. Once you hit a key on a Linotype,
that character is going to be typeset in the slug.

Of course, real Linotype operators make mistakes. I believe a standard
way of handling errors was to mark the offending slug. The
method used at the New York Times (and doubtless elsewhere) was simply
to run one's fingers down the upper row of keys, thus setting the character
sequence etaoinshrdlu into the slug. The operator then re-did the line,
this time avoiding the error.

So now this single line of text has been translated into two slugs, one
correct and one with an error, marked by containing the sequence.

At a later stage in handling, someone was supposed to pick out and throw
away all the slugs containing the sequence etaoinshrdlu. And of course
they were overlooked. So readers of the Times would be treated to an odd
line of text containing the sequence. And thus it became widely known.

This story was told to me by a computer scientist who named a robot of
his "Shrdlu." So not all computer scientists think the sequence
is the letter frequency of the English language.