Archives for category: Typesetting

Here’s a typeface recognition game from Better web type.


A five-minute film which makes it all pretty straightforward.

If you see no video here, please click on the title of this post in order to view it in your browser.

We used to spend a lot of time focussing on the quality of the output we’d get from various typesetting systems. You’d get out your loupe and scrutinize raster lines in Pacesetter or Linotron 202 repros, as if something you could only detect under high magnification could ever matter at all. But in the early days of phototypesetting this was the focus of concern: the image produced by a bit of metal type was something we knew. You knew you could rely on that good old analog reality; you could always see it; you could observe by eye and touch how smooth those curves were; you could always feel it as it indented the paper.

From Hugh Williamson: Methods of Book Design, 3rd edn. © Yale University Press 1983

Because a digital image can only be 1 or 0, on or off, black or white, the image maker has to make a decision around those curved edges: is more of my little dot black, or is it mostly white? On a curve this will obviously lead to a jagged edge as the decision goes one way or the other as you move along the slope. But it’s all on such a tiny scale that you really can’t see it happening.

Nobody would waste their time like this any more.  When a book is processed through a modern-day text processing system it remains a digital entity until it reaches the printing press, or just before that when it is used to create a printing plate. There’s not really any output to look at until you see a set of printed sheets, so we’ve just stopped worrying about it. I think this illustrates two things:

  1. If people don’t have enough to do they will worry about needless stuff, and
  2. It is difficult for people used to a tactile process, literally a hands-on-workflow, to repress the urge to touch their work, just to make sure it’s still there.

Times Roman or Times New Roman? One assumes they are different, but the reason we have these two names is merely the fact that when the face was cut for the Linotype soon after Monotype had introduced it as Times New Roman in 1932, Linotype named their version Times Roman tout court. Perhaps the name change was intended to distinguish it from the Monotype version despite the fact that, to the non-specialist eye, it looked more or less identical. The typeface which The Times (of London) used before that, what we might now call Times Old Roman, was in fact Monotype Modern, cut in 1908.

Times New Roman

The distinction in nomenclature survives the passing of hot metal typesetting: both appear as options on the Mac — Times Roman coming from the Linotype Corporation and Times New Roman from Monotype. There are differences between the two faces, but they are slight. Here from TypeTalk at CreativePro is an illustration showing some of the differences — Times Roman at the top; Times New Roman below. Basically you can see that the counter of the cap P differs, and Times Roman has pointy bits at the top of the shafts of letters, while they have been leveled off in Times New Roman on the lower line.

The creation of Times New Roman came about as the result of an insult. Allegedly when the Monotype Company was invited in 1929 to advertise in The Times’ Printing Supplement, Stanley Morison, who was Monotype’s typographical consultant, replied that he’d rather pay them £1,000 not to set an ad for them as The Times’ typographic standards were so low. Ironically Morison, who had started his working life as a bank clerk, had first become interested in type and printing when reading The Times’ previous Printing Supplement in 1912, and this next supplement got him the job of redesigning The Times, whose management immediately picked up the gauntlet.

Aesthetically not altogether lacking, the face was, it should be remembered made for the functional purpose of jamming as much text into as small a space as possible, and in this it succeeded. Morison made drawings which he then gave to Victor Lardent of The Times who translated them to reproduction standard. Morison used a design by Christophe Plantin (1520-89) as his inspiration, though there are elements of Perpetua and Baskerville in its make up. It took till 1932 for the work to be completed.

Don’t bother checking. The Times no longer uses Times New Roman. According to Wikipedia they stopped using in 1972 and replaced it with Times Europa, then Times Roman took over in 1982, Times Millennium in 1991, Times Classic in 2001, and Times Modern in 2006. Times Roman, older or newer is of course still widely used.

Here, via the SHARP listserv, is an account of some of the props used in making this film. The information is provided by Dan Franklin of The Two Sisters Press in Belleville Illinois.

Davin Kuntze of the Woodside Press in Brooklyn had this to say on the Letpress list in November 2017:

Earlier this year, I was approached to create a couple props for an upcoming major motion picture. The item in question was a front-page lockup of the Washington Post from 1971 just how they would have created it back then. With a lot of help and advice from a few experts (namely, Frank Romano at the Museum of Printing and John Christensen from Firefly Press), I pretty much managed to it pull it off with as much authenticity as I was able. I cut a few corners, mainly with the banner information, but everything else was done with a sharp eye to making it as true to an original as possible.
The creation of these props led to a three-day shoot here in our shop where we were decked out to look like an early seventies composing room at a major newspaper, smoking pipes and high-waisted pants included. They shot the up-close and personal operation of our Blue Streak Comet and Model 31 as beautifully as I’ve had the pleasure of seeing on film (and they were shooting onto actual film).
While I’m not allowed to share photos of the two front-page lockups I created just yet, the trailer just landed yesterday and, much to my surprise and enjoyment, a number of the inserts from our shop were used. Starting at about the 2-minute mark, you can catch brief glimpses of my hands, John Christensen’s hands, some mats spelling out a dramatic phrase and the distribution mechanism.


Then, Davin posted this terrific story on Jan. 16:

Marc, a good friend of mine who plays one of the Linotype operators in the film, is set up in front of the Comet on the second day so I wasn’t in the background again and to give the illusion of a much larger composing department than we actually have. An hour or two into the shoot, and he’s plonking away like he knows what he’s doing, but we had the sword in the magazine and the plunger disconnected so there wouldn’t be any malfunctions during an otherwise good take.

They set up the next shot, and from the back I hear “Ok. Now we need to have the Linotypes casting for the next scene.”
Marc looks over at me, wide-eyed.
Now, Marc has been around the shop for years and seen the machines in operation, but has never actually RUN a Linotype. I walk over and give him a two-minute crash course. Then I get off screen and give him the thumbs up. So Marc has to run a Linotype for the very first time with Steven Spielberg watching closely on the monitors and Meryl Streep standing behind him with her hand on his shoulder talking to Tom Hanks about newspapers and making history and stuff. He totally pulled it off, much to everyone’s relief. While cleaning up a couple days later, we find line after identical line, like Jack Torrance (from The Shining) had been running one of the machines. I guess Marc just found something he could touch type quickly and stuck to that.
Other people you might catch in the Linotype scenes who some of you may know are John Kristensen (from Firefly Press) and Andy Birsh (the owner of Woodside Press). Behind the scenes, we had Rich Hopkins, who cast the large Bodoni we used as as a stand-in for the handset “Post-doni” headline type. Finally, Frank Romano provided invaluable historical information as well as the two large chases for the front pages, the turtles and some other odds and ends that ended up as set dressing. It probably wouldn’t have happened without Frank’s input and assistance or Rich Hopkin’s beautiful type.

The July 2, 1978 issue of the NYT was the last set on Linotypes.

In 2016, there was a delightful story in the NYT about Rudolph Stocker, who worked at the NYT for 50 years:

The trouble with the earliest typesetting machines was always the distribution difficulty (though justification also presented persistent problems). In a letterpress print shop distribution didn’t mean what we think it means today. To distribute type is to take the individual bits of type (sorts) after they’ve been used for printing and put them back into the place they had started out from so that they could be reused for the next pages waiting to be typeset. In a hand setting world this mean distributing the sorts into the type case, putting each individual sort into the appropriate box in the case, minding your ps and qs of course, so that the compositor could start work on the new copy. Distribution would tend to be done first thing in the morning by the apprentice who had to get in early to break up the type pages which had been printed on the previous afternoon and get the individual sorts back to the starting line.

Early inventions tended to founder on this problem rather than on the easier task of getting the right sort to respond to a keystroke and drop into the right position. Type was something a printer valued. It would be obtained, at a cost, from a separate business, the type foundery, and was to that extent a “given”, representing part of the printer’s capital investment. Not until Mergenthaler and Lanston figured out that the problem could just be circumvented by recasting type every time you needed a character, was machine typesetting made truly cost effective. However the Thorne machine came close to solving the problem and was quite successful with between 1,500 and 2,000 machines produced. It traded from 1880 till about the end of World War I under the names of Thorne, Simplex and Unitype.

The Thorne Typesetting and Distribution Machine used that tower to hold all the little bits of type in 90 channels where they could drop down for use in the new setting. The tower consists of two separate parts. The type for distribution would be loaded into the top, rotating part, face out, and sent to the appropriate channel in the lower part by using a series of nicks on the side of the type which distinguished every sort from every other one. Distribution could be continuous with the keyboard operator working away at the same time. After a line was set it had to be justified by a second operator, by the hand insertion of spaces and hyphens. This inefficiency was partially “justified” by suggesting that the justifier served as a sort of extra proofreader.

The Thorne Typesetting and Distribution Machine is covered well at Circuitous Root, which carries many illustrations and much detail, including a link to a promotional booklet available at the Internet Archive. This booklet includes samples of the typefaces available on the Thorne, as well as testimonials from satisfied customers, including this revealing table of output and cost from APA.

The emphasis placed in the Note on the absence of heads and signatures is there to point up the efficiency even more starkly. Obviously setting a head like, say “Chapter I” would give you a quicker, easier line than the full lines of text which would follow.


Optical character recognition seemed like a great idea when it was first invented in the 1970s. If you had an old book which you wanted to reformat OCR gave you the opportunity to use a computer to scan it so as to get you simply from A to B. Unfortunately it was never as simple as this might sound: OCR came with a guaranteed error rate, so what you’d get was only a step on the way to a reset book: an error-filled digital file. Next you’d have to go through a proof-reading and correction cycle. When ebooks came along demand for “re-origination” of books exploded. With a book published before ebooks had been thought of there was no digital file in existence which could be adapted to create the ebook, so you had to scramble to get one made. OCR should have held out an attractive methodology for creating such a digital file. However, in the end people figured out that it was more accurate, quicker, thus cheaper to have humans (often in the Philippines) re-key the book. Humans are much better than computers at recognizing that that smudge is a Capital “S”, not a “b” or an “o” or whatever the computer might decide it was. Errors like this one are pretty obvious, but others are less predictable: would OCR stutter over that “w” in the previous entry? As so often in the computer world we discovered that garbage in really did mean garbage out.

Now, it could be argued that better inputs will result in better outputs, and to the extent that you might argue that a QR code was being processed by a sort of optical character recognition this is obviously true. Publishers, in the early days of ISBNs, before (and immediately after) barcodes and barcode readers had become available would typeset their ISBNs in OCR-A a typeface which was alleged to be infallibly readable. Who knows whether anyone could or did machine-read such ISBNs?

I’m betting that we printed them like that just because we were told to, and that nobody ever utilized the theoretical function. Barcodes quickly came along to make OCR-A redundant, although, belts and braces believers all, we kept on using both for years.


“Mr. William F. Hill one of the early employees of Mr White after his move from Hartford watch maker a superior workman and an ingenious mechanic, conceived a method of making copper type by what may be called ‘swedging’ or pressing by steel dies the face upon the body.” This convoluted sentence comes from History of Typefounding in the United States by David Bruce. Mr Bruce was not a writer: he was the inventor of the first effective typecasting machine in America.

Swedging is defined as shaping metal using a hammer or other force. (Colloquially it can also mean leaving a restaurant or shop without paying. The OED asserts this usage derives from U.S. nautical slang, which sense  appears to have evolved from a meaning of doubling back and going around an object.)

The Oxford English Dictionary sends us to the noun “swage”, which it defines as “A tool for bending cold metal (or moulding potter’s clay) to the required shape; also a die or stamp for shaping metal on an anvil, in a press, etc.”. Swage also means “an ornamental grooving, moulding, border, or mount on a candlestick, basin, or other vessel”, or more remotely “the excrement of the otter”.

I suppose this means swedging copper type would involve just bashing a bit of metal till you’ve formed the shape of a character. I guess such a procedure wouldn’t seem too crazy in a world where the idea of melting the metal and pouring it into a mould was restricted to a one-off hand casting routine. However, hammering a bit of copper would seem to be a slower alternative. I wonder if what Mr Bruce is actually referring to is the making of a mould: swedging a bit of copper might well be a description of just such a punchcutting process.

One of the early problems with mechanized type founding was a tendency for title air bubbles to form in the metal. This made the types lighter, but lead to their collapsing when pressure was applied to them in the printing press. David Bruce’s typecasting machines No. 1 (1838) and especially No. 2 (1843) overcame this and many other problems. His machine and versions of it remained the workhorses of typecasting for a hundred and fifty years.

“Set the Page Free”, a Xerox® project, has just published Speaking of Work: A story of love, suspense and paperclips for the benefit of the 92nd Street Y and Worldreader, a literacy promotion organization. This collaborative work can be obtained at this link for free as an ebook. If you’re a died-in-the-wool traditionalist your Xerox rep can give you a printed copy. It’s really easy to get the ebook though: enter your name and email address and they instantly send you an email link to the book with options as to how you want to download it. Took about 2 minutes total.

Business Wire has an account, and The New York Times wrote about it on publication day, 23 October. The contributors are Jonathan Ames, Lee Child, Billy Collins, Sloane Crosley, Joshua Ferris, Jonathan Safran Foer, Roxane Gay, Valeria Luiselli, Alain Mabanckou, Aimee Mann & Jonathan Coulton, Joyce Carol Oates, and Gary Shteyngart. Chip Kidd designed the cover. I’ve enjoyed the book, particularly the story by Lee Child. And the price is right!

Contributions include short stories, one in the form of a tele script, essays, a poem, and a song. Whatever Jonathan Safran Foer’s piece is about (it’s more about words than work, though I guess you could argue that words are the work-tools of a writer) it includes this interesting disclosure. “Have you ever come across the word ‘esquivalience’? It’s a made-up word — a ‘ghost word’— in the New Oxford American Dictionary, created to detect breaches of copyright. (There would be no other way to know if another dictionary-maker had simply stolen Oxford’s list of words; ghost words prove plagiarism.)” I’m not altogether clear why you’d object to other dictionary-makers following your lead, but obviously OUP thought it worthwhile. The word did turn up at and has since been taken down. Not sure just what good this does to NOAD. Surely you can’t copyright words, just the form of words used to describe/define them.

Esquivalience is defined as “the willful avoidance of one’s official responsibilities”, something which seems ever more relevant to current political debate. Now while I can see that you can add fictitious data to a map to detect unauthorized copying, inventing words is surely less effective: it’s something that goes on all the time in any case, and gives lexicographers a good deal of their work. Surely trying to invent words is a risky procedure for a lexicographer: do it well and the word risks becoming real: do it badly and people will think your product is error-filled. Would it do me any good to write of the verbifurs at as a method of detecting whether anyone as willfully copying my posts? If people think that the idea of a word-thief is one which needs a word, then verbifur may make it. If not, not. And what harm have I incurred either way? Esquivalience may have been invented in 2001 by dictionary editor Christine Lindberg, but surely it’ll have to be included in new editions of the dictionary as it gets used more and more. Ms Lindberg told the Chicago Tribune that she finds herself using it regularly.

Speaking of Work is sort of a promotion for Xerox’s ConnectKey® technology, a suite of apps which facilitate document handling and production. This was the system used to put the Speaking of Work project together. Xerox, a pioneer in so much, deserves our respect for being in at the origins of print-on-demand for books. It is good to know that PARC is still innovating.

At the end of the book there’s a description of how the software was used:

ConnectKey® Technology and the VersaLink® C405 Multifunction Printer enabled secure collaboration and communication across countries and continents, with enhanced productivity and security. DocuShare® Flex made content collaboration effortless. Easy Translator Service translated content around the world at the touch of a button. Xerox Apps for Google Drive & Dropbox empowered digital sharing through the cloud. Print Authentication provided device security using a smart phone. Voice Recognition Technology made productivity as simple as speaking. @PrintbyXerox App enabled printing from virtually anywhere. XMPie® software made the eBook personalized for each recipient*. And the printed book was produced on an iGen4® Press and Xerox Nuvera® 144 EA Production System using a FreeFlow® Print Server.


* The book does indeed contain a dedication page reading “To Richard and everyone, anywhere, who works.”

Is it depressing that a Google search for papyrus will return a page filled with links to the chain of stationery stores, Papyrus? Maybe not; after all what right do we have to assume that the internet isn’t all about business and retailing stuff?

Papyrus is of course the precursor of paper (and indeed the word’s origin).  Cyperus papyrus is an aquatic plant native to Africa. Its pith, cut into strips, would be woven into flat flexible sheets by ancient Egyptians (and others more recent) on which one could write. After the woven sheet had dried out under a weight it would be burnished with a stone to make it smoother. As you can see from this video, the stem has a triangular cross section which almost demands this sort of treatment.

Papyrus “books” were formed of several sheets of papyrus, joined together and rolled up to form a book roll. Writing on papyrus, which although its surface is pretty smooth (the lady in the video tells us its derivation is from the word for baby’s skin), demands different techniques than writing on paper — brush rather than pen. The Wikipedia article is comprehensive. Oddly, papyrus was called wadj, tjufy, or djet in the ancient Egyptian language. I guess this means the Greeks named the paper after the plant.

Papyrus is also a rather over-ornamental typeface designed in 1982 by Chris Costello. It’s the typeface, used, as Ryan Gosling’s character in this Saturday Night Live video is unable to get over, for the title sequence of the film Avatar.

(Link thanks to Lois Billig.)