If you live long enough and pay some attention along the way, you can begin to see the direction of history and to get an idea of what is going to happen next. You’re not necessarily right, but humans like to make patterns, so here we go.
The easy prediction is that typography will be done mostly by machines—machine learning. (It’s funny, but when I was writing this, Microsoft Word tried to correct this assertion to: “Typography will be learned by machines.”)
Writing is done by hand—but with tools. Keeping your pen sharp, your paper clean, and your ink wet—are part of the craft of calligraphy. Humans like that kind of activity. If you can do it well, it’s fun; you get into the “flow.” The craft.
Writing, like cooking, began at home. I don’t know when the first chefs got jobs, but, judging from stone inscriptions, writing became a trade about 4,000 years ago.
Printed books started in the West only 700 years ago, with hand-carved wooden plates. Then, movable type, better paper, faster presses. Writing moved from homes to factories.
Typography, by definition, is an artifact of machines. Graphic designers are typographers when they put type into pages (or “projects” as they say nowadays), sometimes working with one typeface, sometimes several. Good typographers learn how type works in different media, for different situations.
Good type designers work the opposite way. They see how type is used and how it is read. In the 15th century, the aesthetic was shaped by handwriting, but type gathered its own aesthetic; the look and feel of type was influenced by machines.
Rather than just imitating the calligraphy, people designing type incorporated their experience making punches, matrices, typecasting, typesetting, and printing. They made rules to deal with the constraints and the artifacts at each step. Today we call these rules “algorithms.”
In rapid order—Gutenberg, Jenson, Garamond, Kis, Caslon—new masters improved the craft, while printing technology improved. Driven by new punch engraving machines, typecasting machines . . . and then typesetting machines, print became the first mass medium. Soon we had phototype, then digital type, desktop publishing, and the web. Machines did more and more of the work of typographers—with an occasional pause which confirmed our assumption that the way things are the way they will continue to be.
It takes more than a few hundred years for a species to adapt to this much change. As technology spins faster, human nature will continue to trip us up.
The rush of progress
When I started as a typographer (a graphic designer who uses type), I drew layouts on sketch pads. A tight sketch was called a comp, for “comprehensive” layout. I traced headlines carefully from foundry tracing cards, provided by ATF and Monotype. Text was indicated with lines drawn by a broad pencil. The copy was then sent out to the typesetters. I had to “copy cast”—estimate how long the text would be when it was set. You wanted to get it right, so you didn’t have to have to pay to set it again.
In the mid-70s I switched to “hard comps,” that is, layouts pasted together from photostats of type and pictures. I had sets of Stempel blindtext pads with fonts like Sabon in different sizes. We’d photocopy those to dummy up the text. And then we’d make a stat of the whole page, for presentation.
In the 80s, we switched from stats to Xeroxes. Then Xerox made a copier that could enlarge and reduce. Holy cow! And then, color Xeroxes! Anything was possible.
All along, typesetting technology was changing, and you had to adapt.
From the beginning, I had set a bit of type myself, first in hot metal. Then on an IBM Selectric Composer and a lot of Letraset press type. I got a Phototypositor for my first publication art department (LA, 1972). Next, I played around with “front ends” for typesetting machines, like the fantastic AKI terminal which had a screen that showed line length, where you could edit text with a keyboard and a trackball (!). You’d put in typesetting codes, and then output a 6-level punched paper tape.
A decade later, I was at The New York Times which had these “state of the art” Harris terminals. You would key in your type using the old IBM markup: “CC1,9,10,13,” which did a “Change Column” to Imperial in 9 point with 10 point leading, 13 picas wide. Then you would take the elevator down to the composing room, get the proof coming out of the film processor, and go back up to your desk and paste up your comp. (None of the type we set in the art department was ever used in a real page, due to union rules.) But it all seemed so much better than the old days. It was a new version of the old craft, and you could still get that hypnotic flow when you were working. Some liked that flow better than the actual typography. I said, that’s just the play. The sandbox.
And then the Mac arrived, turned that into reality. You could see what you are doing—you could make the copy fit without all that guessing. Soon all designers converted to the personal computer, except for a few luddites. There was some grumbling about the loss of all the “fun” we had done it the old way. They liked the analog methods and resisted using computers. At the Aspen Design Conference in 1989, I did a demo of Quark Xpress, and Milton Glaser said, “You don’t do this yourself, Roger?
For me, the real fun has always been in the design—seeing the final page or project emerge. I liked working on the screen better than the sandbox filled with designer toys (T-square, pica poles, dividers, ruling pens, illustrations boards and wax). But many designers resisted the technology and expressed horror that the general public was now, God spare us, doing design.
Type designers, too, were moving their work to the desktop. David Berlow (my partner at Font Bureau) was one of the first at Mergenthaler Linotype to use Ikarus interpolations. He jumped on Fontographer before it was released, predicting in 1988, “One day all of this will be on the screen. The design. The type. The images. And the reading.”
You can see what’s happening: In each step, type designers and typographers are getting closer to the readers. So much work and so many decisions were now being done by the code.
Coding is the new sandbox
The race was on. From Postscript fonts, to Multiple Masters, to TrueType and TT GX, to EOT and then WOFF for the web, and now to Variable Fonts. The devices with screens proliferated; there were digital fonts in your car, and on your refrigerator.
Computers weren’t just setting columns of type any more. In print, they were composing pages, flowing text, integrating images, and preparing for output. On the web, they were taking simple rules and substituting fonts, adjusting sizes, changing the widths.
Designers were writing more and more code. We saw Erik and Just blow up Multiple Masters with Times New Random and Beowulf. Matthew Carter designed Georgia from the bitmaps to the outlines, instead of the other way.
Fontographer begot Robofog. Then Fontlab appeared. Then Robofont. Then Glyphs. Their users wrote scripts.
The code was doing more and more of the work. It took on every task of graphic design, one at a time, and expedited it. Even with the apps Photoshop and Sketch, we’re just setting the specifications, and the code does the rest.
Just as Gerrit Noordzij lead the progression of letterform theory to algorithmic type design, his successors like Petr Van Blokland are pushing to the next step: algorithmic typography.
When Sam Berlow and Eduardo Danilo and I launched Ready Media in 2010, with InDesign templates for many kinds of publications, there was great howl from members of the Society of Publication Designers, who felt threatened by the change. Maybe they sensed that the old art department was a thing of the past.
Their sense of flow comes from threading code like a weaver threads yarn on a loom. Okay, that’s fun. But were that AI or machine learning could do more and more typography and design.
There are a bunch of apps that promise, with varying success. Check out The Grid, Wix, Firedrop, and Brandmark, to a name a few.
The most promising so far is Van Blockland’s PageBot, which we used to make some of this TYPE No. 3. Petr’s point is that all this effort is not about the code. It’s about the design.