The Great 202 Jailbreak - Computerphile

Before laser-printers, high quality print-outs were the domain of typesetters, expensive and tightly controlled. In 1979 a Bell Labs team reverse engineered one in their summer vacation. Professor Brailsford has the details.
EXTRA BITS - More on Printing and Typesetting History: • EXTRA BITS - Printing ...
Mainframes and the Unix Revolution: • Mainframes and the Uni...
To find out how the 1980 Bell Labs `vacation memo' was completely rebuilt,
in 2013, visit:
www.eprg.org/papers/202paper.pdf
If you want to see the `vacation memo' itself -- in its original scanned-in form and after being rebuilt -- then visit:
www.cs.princeton.edu/~bwk/202
The memorandum gives an outline of how the `reverse engineering' of the 202 was accomplished.
/ computerphile
/ computer_phile
This video was filmed and edited by Sean Riley.
Computer Science at the University of Nottingham: bit.ly/nottscomputer
Computerphile is a sister project to Brady Haran's Numberphile. See the full list of Brady's video projects at: bit.ly/bradychannels

[@babis8142]

Just noticed in the beginning of the video it says and at the end I'm pleased

[@0pyrophosphate0]

Can we just have an 8 hour livestream of the professor telling stories?

[@bigun402]

This gentleman that is in this video needs to voice children movies. Can I purchase him to be my grandpa? Does he voice any books on tapes? I think I smelled roses and saw a unicorn just listening to his voice.

[@ZeedijkMike]

Yes he does have a very pleasant voice.

[@picosdrivethru]

+Zeedijk Mike I totally agree. He timbre is quite lovely.

[@Disthron]

....This guy is like the David Attenborough of computing history! XD

[@BertGrink]

+Disthron It's quite funny you should say that, because i was thinking of David's brother Richard, while i was watching this video.

[@JulianIlett]

I used to repair the 202. Floppy disk alignments, CRT alignments and even the occasional software patch using hand-punched paper tape. It had a very heavy PSU - it was no fun carrying that up 3 flights of stairs.

[@TechyBen]

"But why would you want to do that. No one wants to do that." Famous last words in business. Famous beginnings for successful new starts.

[@LMacNeill]

Anyone younger than 30 years old just has no clue how difficult it used to be to get specific fonts printed. I'm 43, and I started my computer career in the mid 1980s -- certainly an easier time than the 1970s that Professor Brailsford is speaking about, but still far more difficult than today. True Type fonts simply did not exist. Your printer had to have the fonts built-in -- you had to buy specific (and very expensive) font-cartridges that contained the fonts, and then the software had to recognize those cartridges to be able to utilize the fonts. And the fonts could not be "infinitely" sized -- they had only specific sizes built-in. True Type Fonts are one of the best things ever invented, and all this work that these gentlemen did back in the late '70s is the earliest step taken along that path. Without their work, and the subsequent work of others that followed in their footsteps, printing would still be an EXTREMELY expensive, time-consuming, and difficult process.

[@thepumpkingking8339]

LMacNeill Coming form the Stationery side in the early 90's . I Remember the changeable Daisy Wheel's with different fonts. So even into the 90's .. The Font Issue, as we have it today. Still had not been resolved.

[@SyphistPrime]

Looking at old tech just makes you appreciate how far we've come.

[@JuanPabloCarbajal]

The original memo uses a very scientific statement about the 202 machine: "considerable pain in the posterior"

[@coomcake]

$50,000 in 1979 is the equivalent to about $160,550 today

[@MartijnvandeStreek]

This is a great piece of Unix/typesetting history! :) Modern "man" pages (manual pages) on Unix and Unix-like systems are still written in the language that's described here (troff/nroff), and they're easy to typeset/print or display in a terminal window.

[@Seegalgalguntijak]

Wow, computer science must be the only field of science where you could be an actual scientist with the up-to-date technologies and then later on become a historian and talk about things no one knows a thing any more... ;-)

[@Madsy9]

This is an actual problem in my opinion. Proprietary software and libraries/APIs die off and the documentation with it. Not only does it make maintenance of old software difficult, we lose a big chunk of our history and culture every time it happens. Imagine how important SGI was in the field of computer graphics when they made the first OpenGL API. Those first library specifications are all gone, as well as most of the documentation on WGL. Imagine what we lose whenever computer games become abandonware, when no one has the source code anymore and the game wasn't selected for preservation. Games that have filled people's lives with joy and meaning just suddenly vanish to be forgotten forever. A few dedicated groups in the world collect old C64 and Atari games exactly to preserve an important part of our (pop)-culture.

[@Seegalgalguntijak]

***** I think that is a learning curve the world has to take, until every code is free (as in open, "free speech", not necessarily as in "free beer").

[@capt-morgan276]

Best professor on computerphile!

[@HennerZeller]

Bravo! As a typographile and believer in naturally being able to get the full access to the sourcecode of machines you purchased and *own*, I enjoyed this description of a historic jailbreak very much. About 30 years ago, when this reverse-engineering happened, the dark ages of vendor secrecy and proprietary closed source just began, and we're in an even deeper shit nowadays. This is when I am very thankful for people such as Richard Stallman and Linus Torvalds that played key roles in showing the world that keeping software accessible for you to learn from, improve upon and share is creating a tremendous value for society: GNU/Linux is by far the most common operating system deployed today (most notably in places many people don't realize such as phones, appliances, routers or datacenters of essentially all big internet companies ...)

[@rich1051414]

It's story's like the 202 jailbreak which really puts into perspective how much the 'open source' ideology meant to the unix team.

[@MatthewPotter]

All that font recreation reminds me of when I was at Yellow Pages Group and needed to convert Inuit font families into opentype for the new platform we were releasing. Not only was it a completely unknown language to me but the glyphs were also not roman based nor pictographic. On top of that there were, as with your chess piece trick, zero width glyphs that acted like accents. Luckily I had PDFs to start from that I was able to capture the vector from but going through for the spacing and proper unix character tables took ages to do. Great video!

[@Tom-lf5og]

04:06 "smelly developer" ... at first I was like... "what ? like a PHP developer ?" But then I realized he was referring to the chemical LOL :)

[@un2mensch]

After this video, Dave Brailsford has gone from awesome guru to fucking legend in my mind. I loved this video so hard.

[@thecassman]

Great video! I'm sure i'm not alone in saying that all videos should be this long!! 10 minutes isn't always long enough to fit all of the info into it - if this video was halved in length it wouldn't have covered anywhere near enough.

[@clearmenser]

This extreme level of academic geekary is impressive to say the least. Wow, also wow, and wow.

[@joshcryer]

Are you effin' kidding me? They jailbroke the thing, reverse engineered it, then, as a hobby, at a whim, they REPRODUCED THE ORIGINAL MEMO BY REVERSE ENGINEERING A PHOTOCOPY. That's some inception level hacking there. God I love this guy.

[@davidnixon4509]

As someone who worked at both Mergenthaler's UK affiliate Linotype Paul 1978-79 and also at Bell Labs 1980-1990 this was a fascinating video and I can attest to the lack of quality in the Mergenthaler software of the time. I never worked on the 202 but I did work on its big brother the Linotron 606. There was an interesting problem on the 606 called "character breakup" where seemingly at random and out of the blue small sections of text would intermittently display as if the letters had been chopped up in a blender. It was never repeatable and never the same section of text. Some senior developers had been tasked with finding the problem but had failed to do so and there was speculation it might be a hardware problem or possibly a timing problem. I was working on a customer site commissioning a new installation of 606s and we saw the problem. With the help of the 606 hardware engineer we found the cause was a software bug that only got executed when there was a disk read error transferring the font data from disk into the memory used by the CRT to display the text. So the data was reread from disk but because of the bug the data was placed at an offset address to where it should be so the characters displayed would be part of one character and part of another more or less depending on how many bytes each character used. Solving that was the most interesting thing I did at Linotype. The most interesting thing I did at Bell Labs was attending John Horton Conway's talks, and the second most interesting was hacking the curses library version of the Aliens game to play Pacman on our terminals.

[@gulllars4620]

I'd love more videos from Professor Brailsford. I've watched all videos on computerphile, and there are many interesting people in the videos, but Professor Brailsford is without a doubt the best story teller. I love how he makes stories about very specific historic technical challenges come to life. I wonder what he thinks about typing on keyboards, specifically layouts (qwerty, dvorak, etc), and mechanical keyboards vs rubber dome or other types of key switches.

[@steevee1945]

I love hearing this man tell his stories.

[@HarleyPebley]

Wow. I used to work on a Merg 202. It was a cool bit of kit in the day; a big step up from the previous generation of typesetters. My dad headed a team at a commercial printshop that interfaced it to a DG Nova computer to bypass the paper tape input. Fun times.

[@ianchard]

In 1981 my mum worked at Hove greyhound track typesetting their programmes on a similar machine -- an Addressograph Multigraph. It was much more primitive than the 202: the CRT was replaced by a spinning disk with characters on it, and a light shone through the disk onto the bromide, exposing one character at a time. Changing fonts meant unscrewing the disk! The results were impressive, though, and I still remember the smell of developer and fixer as the long strip of bromide lay drying on the floor, waiting to go next door to have offset plates made. Laser printers just make it too easy!

[@kevinwynott7755]

This gentleman has the absolute BEST Voice......I'm not even into computer tech.....I just enjoy hearing him speak!

[@fllthdcrb]

Fascinating bit of history here. I like it. I had no idea typesetting was _that_ much of a hassle back in the day. Seems rather shortsighted of Mergenthaler, though. But then I suppose that reflects a common attitude of companies at the time. One irritating little technical problem: the oblique camera angles on these documents, combined with the low depth of field, makes them pretty hard to look at. I wonder if something can be done about that in the future.

[@DaithiDublin]

Excellent video! I worked in print for several years in the 90's with some older guys who had been in the business 20 or 30 years at that stage. Man, could those guys talk about fonts and typsetting and composition! Fascinating subject.

[@TheTrueRandomness]

Reading the original complaint letter is totally worth the time... This is one of my favourite parts so far: "Lyle R, our (by now regular) repairman, suggested wiggling the chips [sic] on the code converter board. This apparently nonsensical suggestion in fact cured the problem for a while. Thursday, July 26: The same chip-wiggling exercise became necessary every few hours, from this day forward"

[@TheTrueRandomness]

Friday, August 3: The machine died finally; no amount of chip-wiggling had any effect. Sunday, August 5: The machine spontaneously cured itself, at least for a while.

[@TheEvertw]

I remember the first time I saw a computer print out a non-built-in character. I was awe struck. I must have been either 19 or 20. The printer was a dot-matrix printer that turned out to have a "graphical" mode where you could print bitmaps. I did play with it a little trying to output PCB etching masks, but soon gave up and went back to hand-drawing them. I greatly admire the perseverance of these people!

[@TheKuleluke]

computerphile is awesome

[@StuartPetty]

Please never stop making these videos, there great, many thanks

[@voveve]

You could do a PreiodicVideos on developing photos/this kind of prints!

[@aameen951]

I have a request for a video: "Splines and Bezier Curves"

[@sige333]

I love watching the videos with Professor Brailsford!

[@dawidrozmus301]

I love all of your videos. It is great that you interview people that walked their talk!

[@Diachron]

Fascinating bit of history. Thank you for all the effort to authentically restore the memo!

[@Galakyllz]

Wow, this video was very good. I love when you delve into historical aspects of computing.

[@HKragh]

I just love this channel. And Professor Brailsford!

[@Quarter2830]

Absolutely brilliant storytelling of little known happenings and photography. I dig the bird shadows on the blinds and the subtle lens flares.

[@jacderida]

Excellent stuff. This guy is so interesting, I could listen to him all day!

[@BenjaminCavileer]

I started printing in the dot matrix generation. I've had a fair share of font problems in my life, and now I understand where it all comes from. Youve filled a gap in my knowledge of printing history. I wish I was there alongside those guys, hacking that 202.

[@Hyreia]

I loved listening to this guy gush about the typeset printer. The energy in the way he describes the parts and describes how it works really sells me on just how impressive the technology was for its time beyond just his words in text would.

[@Herrisx]

As always, Prof. Brailsford is amazing! Thank you!

[@3snoW_]

I love the videos with Professor Brailsford! Keep up the good work!

[@TheRogerx3]

I was a courier for a linotronic bureau in the early eighties, I did find this faciniating.

[@abcde_fz]

Wow. I cannot believe this. I went to a computer tech school back when such things were almost the BEST way into early computer tech work. Control Data Institute. Thank god. Never would have made it through four years of what passed for computer tech college back in '85. And just now, reading the Vacation Memo, I am astounded to find out that something that the memos authors found a nonsensical, that turned out to slightly ameliorate a problem they were having with the 202, would have been a totally common suggestion coming out of MY mouth, back in those days. "Wiggle the chips" "What? Wiggle the chips?" "YES, dude, you're a PHD electrical engineer. You're dealing with a bunch of little non-heat-sinked IC chips probably pressed into a PCB board, WITHOUT BEING SOLDERED IN!!! Where some of those chips MAY have the character bit maps on them." (OK, I am confabulating a little. Some VRTs stored their character maps as little arrays of timing blips or whatever-the-hells in chips. Address that array on that chip, and that array gets sent to the CRT, the array ends up being a letter "A" when scanned out. Work with me here...) "So where do you THINK an annoying little not-quite-permanent sorta-intermittent problem WOULD live on a PCB full of 16-pin press-fit IC's???" IF it's software, it's yours, debug it. If it's hardware, it's likely to be quite cyclical in nature, and highly reproducible, but if it's FIRMWARE, and that WARE is spread across mulitiple PCB's, (Naked Mini?), then wiggle the damn chips. You probably won't HURT anything, and you might well fix that annoying little not-quite-permanent sorta-intermittent problem. Can't believe I just MIGHT have given a little advice to someone like Brian Kernighan, if I'd been working for Merganthaler at that time. I already knew who HE was. I wasn't nobody! :)

[@ThatNateGuy]

I could listen to Professor Brailsford speak all day.

[@powmod1]

Jesus! I never thought printing fonts would be so complicated...

[@utopialabsvideos9408]

Adding to a comentary below: I didn't watch many of your videos in Computerphile because they were so short. But I think now, with a length of 15 or more minutes, videos are more enjoyable. I love Computer Science history and I love that Computerphile does these videos long enough to enjoy and dream about old times... Thanks, Computerphile! You're like my "nerd Netflix" now!

[@shkolarac]

"web pointers" - such nice way to say link :)

[@jlinkels]

Very nice presentation. I have read the vacation memo with much pleasure. Those were the days that disassembly was still a feasible means to understand how a device worked. And not to forget building a piece of hardware to interface a PDP11 to an external device.

[@daedra40]

Only 2 minutes into the video, I had a feeling I'm going to get something extraordinarily great. Thanks professor, and everyone else, these videos are indeed a pleasure :D

[@nekaiionera]

Fantastic video! I love bits of history like this.

[@1DJLNR]

these video's are what i love youtube for and only these types of video's. memory lane!!

[@SetMyLife]

VERY interesting! Thank you for sharing such a jewel of stories!

[@Gabbos]

Amazing! I love longer more indepth videos like this. Keep it up!

[@TheDarkerPath]

Love it! The longer format is great - more please :)

Wow, this brings back some memories! I really enjoyed this episode! I so so so remember these fonts, printing like this and in one space I was in practically hallucinating via the chemicals, then cut and paste, then Xerox, etc. (those were the days!)

[@giulianobernardi4500]

Great video about a very fascinating story. Thanks for sharing!

[@gerbermarks]

As an nroff/troff user since 1978, I found this fascinating. At a university department (AGSM at UNSW, Sydney), we didn't have a 202 (and so used nroff rather than troff before 1986), but come the first laser printers, we fell on troff. I still use groff for _all_ my text processing.

[@FreeFallForFive5]

Amazing video! Great work on the font professor Brailsford!

[@AlexZenla]

+Computerphile I could listen to Professor Brailsford all day. He is my favorite person on all the Brady channels, second is Professor Poliakoff.

[@BS-bd5uq]

As a student learning type design, this piece of history almost makes me cry.

[@momlulu66]

Fascinating. Thank you for making this video

[@BeCurieUs]

Ha, only a C coder would call a link a pointer, love this channel :D

[@MinecreftTakeover]

I could listen to this guy talk for days

[@ericsbuds]

new video! how exciting!!

[@ElectricEvan]

Thank you so much for making this!

[@barefeg]

every video should be this long

[@FlipJanson_]

Is there any way we can get the recreated Print Out font? Not any real purpose, but because I like to look at it.

[@frankharr9466]

I like the typesetting histories.

[@SpiderWeb1965]

Great video. Also, both papers made for excellent reading.

[@feastures]

Great story and teller. Thank you!

[@ChallengeTheNarrative]

Appreciation of old tech, I am in my element :-)

[@666Tomato666]

and there I was thinking that I couldn't have any more respect to the UNIX guys... awesome story

[@jamesusespivot]

Sorry for the question but I'm too young. What is a typesetter?

[@Computerphile]

might be worth watching the 'extra bits' film as this is a background on typesetting and how the 'typesetter' machines evolved... >Sean

[@parttroll1]

Ah the Sicilian defence my favourite opening for black

[@jellevm]

That was really cool!

[@MarcelJEMol]

Funny, I did not know this story. But I did something similar in around 1985/86 as a student project at the Technical University of Delft where we wanted to setup our own fonts on I think a Xerox printer in use at that time. We managed to crack the code (with help of a few other people in the Uk I believe) and created a font to print out the TUDelft logo.

[@brianpso]

This story was so cool, this guy got so many good histories =D

[@MegaPeers]

Great story! Must have been awsome to ne part of the beginning of the computer era.

[@Exevium]

You didn't do too bad at all :) Even with todays tools, I'd have trouble recreating some of the new fonts. Very impressive, very cool 80's computertech, awesome vid :)

[@tarcal87]

Not gonna lie, I understood maybe 10% of it, but I was watching it in awe, seeing how enthusiastic he is about the past. Must be amazing to be (to have been) his student

[@NATESOR]

I won't feel so bad about playing videogames when i hear this genius of a man spent 100+ hours messing around with minor details on a font

[@joshuaunderwood7]

Amazing video.

[@robbiep742]

Needs more comic sans

[@GreayStone]

Can you share your rebuild of the linotron print out font. I'm sure someone could find a use for it

[@the-goat]

Really educational. Thanks

[@kcircuit8684]

2:30 birds out the window

[@Vospi]

This is positively nuts!

[@croyfer]

Great video!

[@timdiggerm]

Are your reconstructed fonts available for download?

[@philsbbs]

love the knowledge sharing..

[@harmonicresonanceproject]

Brilliant - thanks!

[@unvergebeneid]

Funny how the 2D graphics of that time had the same restrictions as today's 3D graphics and in fact because we have so awesome 3D hardware compared to 2D these days, even 2D games tend to have circles that are actually visibly comprised of line segments. I wonder when this will finally end and we'll have proper curves in 2D _and_ 3D.

[@smegskull]

Curves are actually easier to do on analogue systems than on digital systems so we are kind of headed in the wrong direction for that at the moment.

[@unvergebeneid]

smegskull Well, I'd be fine with using pixels as long as the object description itself wasn't restricted to polygons. All you need for that is computing power, lots of it. I think the next step we're going to see is real-time raytracing and it's about time. I'm sick of plasticky leaves that block all sunlight. But after that, I hope solid models will come back on the table (especially since raytracing would mean that it suddenly becomes important what's _inside_ of volumes).

[@MonophonicSurround]

We have curves in 2D and curve-segments in 3D for a long time already. Example: en.wikipedia.org/wiki/B%C3%A9zier_surface

[@unvergebeneid]

MonophonicSurround I'm talking about real-time 3D. And regarding curves in 2D ... well, I was referring to the video.

[@smegskull]

yes but you cant draw it. this is the problem with pixels. with analogue images like oscilloscopes you can draw actual curves.

[@Brok3nC4rrot]

Any chance Prof. Brailsford has released Printout as a ttf for people to install on their home computers?

[@JonMasters]

What's a good academic text with more of the history of typography/fonts, Dave, Steve?

[@JonMasters]

Chris Chapman or Steven Bagley hopefully have some recommendations...

[@unlokia]

COR! Wunderbar!!!

[@chrisharrison763]

Although I really like Tom Scott, Professor Brailsford is the real anchor of this channel. Excellent.

[@jdgrahamo]

It is extremely difficult to create a type-face that looks anywhere near acceptable. There are all sorts of optical illusions, which vary according to the point size, as well as the flexibility needed to produce bold and italic versions, numerals, caps and lower-case et al.