Hacker News new | past | comments | ask | show | jobs | submit login
Etaoin Shrdlu (wikipedia.org)
116 points by tosh on Nov 1, 2021 | hide | past | favorite | 33 comments



The documentary of the same name, which chronicles the last ever hot metal typeset edition of the New York Times, is fascinating and I highly recommend it.

It's 30 minutes and you can watch it for free on Vimeo: https://vimeo.com/127605643


I discovered a couple personally fascinating things stemming this documentary.

According to the film, the last paper published with hot type was July 2 — my grandfather’s birthday. He was an avid reader of the New York Times and loved doing the Sunday crosswords — in ink!

In sharing the fun coincidence about the date of the last typeset paper with my mom, I learned one of my grandfather’s uncles worked at the New York Times doing the exact typesetting seen in the documentary. He retired in 1968, though, so never saw the transition shown in the film. It gives pause to think of the front-page headlines he may have been involved in preparing (i.e. Pearl Harbor, JFK, etc.).

I’ve worked on the web for 25 years. I wonder what my (great-great) uncle would think of digital publishing. :)


The Linotype machine has become a kind of obsession for me. It is dizzying to think about the impact this machine had on the entire media landscape of the last 130 years. Not only did it allow for regular newspapers that were longer than a few pages, it also dramatically decreased the cost of both newspapers, books and magazines. In a way, an entire century stands on the shoulders of this machine, invented by a single man (Ottmar Mergenthaler) who was obsessed with it to an unhealthy point:

> One of Mergenthaler’s greatest setbacks was that he could not stop inventing. He constantly wanted to improve upon the machine. This made investors angry because they wanted to ship the machines and make money right away, but Mergenthaler kept coming up with new ideas—delaying the shipping process. A publisher of the New York Tribune, named Whitelaw Reid, tried to make money fast off of Mergenthaler’s invention. He began to criticize the Linotype design, in hopes that the stock price would decline so that he could buy more and take complete control of the devices. This put a lot of pressure on Mergenthaler, who was already a perfectionist and worked late into the night. Eventually, Mergenthaler’s health began to suffer and he died of Tuberculosis by age 44. His premature death meant that he never lived to see the success of his Linotype invention [0]

But it is also the technical aspects of the machine that are so fascinating: the bucket sort mechanism for the matrices used a 7-bit code for each character, encoded in the matrix teeth [1]. The machine supported text justification in a completely mechanical way. There are just so many small engineering gems in this machine, and at the end it prints out a cast letter mold. All that in a mechanical machine that fits into the corner of an office.

Mergenthaler was the Gutenberg of the 19th century. Unfortunately, almost nobody knows his name.

[0] https://sites.google.com/a/umich.edu/from-tablet-to-tablet/f...

[1] https://upload.wikimedia.org/wikipedia/commons/4/4c/Linotype...


One of the things I love about linotype text justification is that it’s actually really a solution to a mechanical problem - the machine needs to make a complete line of text every time without leaving gaps at the end of the mold, and squeeze all the character blanks together tightly to prevent lead from squirting out - so driving wedges in between words to squeeze them together is a neat solution to that problem. That it has the side effect of creating that characteristic newspaper column justified aesthetic is a consequence of an engineering decision, not so much a deliberate design choice.

Interesting article on this tech and its possible connection to our ongoing confusion about how many spaces go after a period: http://widespacer.blogspot.com/2014/01/two-spaces-old-typist...

It’s not conclusive, but there is definitely something to this and the linotype’s mechanical justification that I think overall emphasizes how much what we think of as aesthetic judgements about what makes for good typography are actually just things we’ve grown used to from the technical limitations and operational choices made by people using particular technologies in the past.


> a mechanical machine that fits into the corner of an office.

Could these machines be used in an office?

I watched the documentary mostly to see how they were used, and it looked much more like an artisan's workshop -- fairly noisy, hot type metal (mostly lead) in the machine, hammering the final page flat.


A Mergenthaler company survived into the 1990s. Its L-202 laser typesetting output unit was widespread in the early 1980s, and its L-300 allowed Postscript output.


The same 1978 documentary on YouTube, without a need for logging in:

https://www.youtube.com/watch?v=1MGjFKs9bnU


It should be noted that this version has considerably poorer video quality.


Thanks for noting this. I couldn't compare since I have no Vimeo account.


They very briefly mention deaf printers using sign language. I had no idea being a printer was common for deaf people.

https://www.dailymoth.com/blog/the-washington-post-recognize...


I think it was a practical decision. The room was so loud that even if you could hear normally you wouldn't be able to hear someone talking anyway.


Oh lord, the computerized solution still results in dozens of men having to cut out the computer print-outs with an xacto and glue them onto the pages by hand. WTF. All the technology they had in those roomfuls of computers and no-one had yet written the software to lay out the page too.


OK but it's 1978. The processing power just wasn't there yet for actual page layout.


I did not think about the fact that lines were set in molten lead before! Did typesetters experience health issues from this?


Yes - I’d heard of ‘hot metal printing’ for years before I actually saw a linotype in action and realized how literal it was. A linotype machine literally molds lead slugs into a ‘line o’ type’ then stacks them up to make a printable plate. It’s hot when it comes out because it was molten a moment earlier.

And by lining up those plates next to one another you get columns. Newspaper layouts just naturally fall out of the capabilities of the technology.

So much of what we take for granted in computer typography has direct lineage to mechanical linotype concepts - all web developers should have an understanding of how linotypes worked because the machine’s function defines the form of so much of what we expect of typographical layout.


I'm getting "Video is not rated. Log in to watch."


Use yt-dlp https://github.com/yt-dlp/yt-dlp/releases

    yt-dlp https://vimeo.com/127605643
The film is also available on the Internet Archive - https://archive.org/details/FarewellEtaoinShrdlu

Cheers :)


Some past submissions. Others?

Farewell, Etaoin Shrdlu - https://news.ycombinator.com/item?id=28464368 - Sept 2021 (1 comment)

Farewell Etaoin Shrdlu (1978) [video] - https://news.ycombinator.com/item?id=23538028 - June 2020 (46 comments)

1978 – 'Farewell, Etaoin Shrdlu' - https://news.ycombinator.com/item?id=16907778 - April 2018 (2 comments)

The last day of linotype at the NYT (1978) [video] - https://news.ycombinator.com/item?id=16904770 - April 2018 (65 comments)

Farewell – ETAOIN SHRDLU (1978) - https://news.ycombinator.com/item?id=15806072 - Nov 2017 (1 comment)

Etaoin shrdlu - https://news.ycombinator.com/item?id=9958143 - July 2015 (1 comment)


My dad left school at 15 to support his family during the great depression, his dad was blind, he became a linotype operator.

When he went off to war he survived in relative safely, always a few km behind the lines in an HQ company, only got shot at a few times. This was because he had a rare skill for a man at that time - he could type (on a typewriter as well as a lino machine).

Skilled lino operators were very rare and were treated well they could quit, go travelling and always know they had a job to come back to. (Dad went to university under the NZ equivalent of the GI bill after the war and became an accountant - for the same newspaper).


The image of the linotype keyboard is worth 1000 words: https://upload.wikimedia.org/wikipedia/commons/4/46/Linotype...

The ligature ff gets its own key! And em space and thin space too. Very cool.


Any character without its own key required manually inserting the appropriate mat in the line which was a time consuming process. You’ll notice if you look closely that there are keys for fi fo ffi and ffl as well. f was the Achilles heel of the Linotype machine. Because of the way it worked, it wasn’t possible for one character to overlap the next so even with the f ligatures, it was possible to get ugly combinations like fifty fifes. One 1920s design had f-ligatures available for most combinations of f+a letter but was incapable of being used to typeset Kafka or Afghanistan.


I found on the Wikipedia page that "pi matrices" referred to all the gylphs that didn't have their own key, due to rare usage. Fractions, foreign letters, presumably £ for the American machine shown.

https://en.wikipedia.org/wiki/Linotype_machine#Pi_matrices


SF Author Fredric Brown published a story in 1942 titled "Etaoin Shrdlu". I believe he worked as a linotype operator at one point.


ETAOIN SHRDLU was the frequency counted manually from a corpus of 20,000 words in the 60's.

The actual letter frequency is more like: ETAOIN SRHLDCU [1]

1. English Letter Frequency Counts: Mayzner Revisited - http://norvig.com/mayzner.html


ETAOIN SHRDLU was the standard ordering on linotype keyboards from the 1880s, so it seems unlikely the corpus it was based on was analyzed in the 1960s.

Norvig’s article doesn’t actually attribute ETAOIN SHRDLU to Mayzner’s study, which was more interested in developing statistical models for bigrams, n-grams, and letter positions. In fact it seems more likely to me that if Mayzner’s 20000 word corpus managed to match the ETAOIN SHRDLU sequence it might be because they actually used it to calibrate the corpus to ensure it was representative so that they had some faith in the n-gram analysis.



In my case, I saw "Shrdulu" and "Zardulu" came to mind [1] :-). There's a fun podcast about them at Reply All [2].

1: https://en.wikipedia.org/wiki/Zardulu

2: https://gimletmedia.com/shows/reply-all/o2ho6j


Speaking of alternative keyboard layouts, I'd love to keep a Blickensderfer [1] around the house and learn DHIATENSOR. Unfortunately it's one of several aspirational hobbies of mine that's a complete and total anachronism with zero utility (like learning Morse Code, running an SGI Indy, or building tube amps) that I can't justify spending time and money on :-/.

[1] https://en.wikipedia.org/wiki/Blickensderfer_typewriter


I learned about this mnemonic as a kid for doing frequency analysis on simple substitution ciphers.

(Also when I saw my first R rated movie I apparently quipped: “It should be fetaoin shrdlu”)


It’s my go-to for Hangman or other letter-guessing games.

Love the quip :)


In more recent letter frequency analysis[1] the letter C makes the cut. The distribution also reminds me of anomaly detection using Benford's Law[2]. Wikipedia provides equivalent strings for other languages[3].

[1]http://norvig.com/mayzner.html

[2]https://en.m.wikipedia.org/wiki/Benford%27s_law

[3]https://en.wikipedia.org/wiki/Letter_frequency


I saw a demo of one of these Linotype machines long ago. It totally fascinated my ten-year-old self. IIRC, I tried to get the operator to give me a line of type. He wouldn't -- something about lead being toxic.





Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: