After Fonts, Beyond Faces
Exploring generative typography
A recent article by Melody Weister on webdesignerdepot asks how generative typography will change the web. The video is a promotion for the Type Director’s Club typeface competition, and if you don’t want to watch the seven minute video, I’ll sum it up for you in a picture:
This font has been generated from data about my hometown, like wind speed and current traffic. In the video, various typography experts discuss generative typefaces and expound their benefits: they can adapt to their environments programmatically. But there’s something wrong with a typeface that changes according to its context.
Type is unique because it is not part of the environment; rather, it is part of the speaker.
Brandid, a clothing company, recently published a page describing their typography. While some of their design choices unfortunately counter what they’re trying to do, their idea is solid: create a typographic voice for each ‘speaker’ on the website. Brandid’s principles are based on the fact that each typeface naturally carries its own tone; for example, the following pictures:
Typefaces are naturally human, not environmental phenomenon. Their “font,” “a particular size, weight and style of a typeface,” conveys something about the speaker: their mood, intent, or simply technical competence.
What designers should be concerned with, instead of making typefaces that respond to the amount of traffic or wind speed, is the mood and intent of the speaker. In the past, any attempt to change font as often as speaker mood has come off as hackey, almost like the ransom note example below. A mixture of dissimilar fonts implies madness.
But a generative typeface can fix this. Instead of changing its font based on environment, however, it changes based on speaker tone. If you search for “kinetic typography,” you’ll find something like this, which is a step in the right direction. However, each one of these quotes was painstakingly animated in After Effects, which is no way to run a typeface. What if we could automate the process?
We have enough fonts — we know the way each curve, each serif, ascender, descender, and x-height affects the impression the font gives off. By programming these variables into a typeface, we can create typefaces that move beyond fonts. Perhaps we can even create type that moves beyond faces.The process would require markup, for example:
<disdain>Who are you talking to right now? Who is it you think you see? Do you know how much I make a year? </disdain>
<arrogance>I mean, even if I told you, you wouldn’t believe it. Do you know what would happen if I suddenly decided to stop going into work? A business big enough that it could be listed on the NASDAQ goes belly up. </arrogance>
<anger>Disappears! It ceases to exist without me. No, you clearly don’t know who you’re talking to, so let me clue you in. I am not in danger, Skyler. I AM the danger! </anger>
<disbelief>A guy opens his door and gets shot and you think that of me? </disbelief>
<arrogance>No. I am the one who knocks!</arrogance>
Obviously the tag names are opinionated and semantic, but you get the idea. The generative typeface would read each tag in context and transition between each as the document moved forward. The code would work somewhat like extended ligatures: instead of “c followed by t equals a certain ligature,” the typeface rule would be “arrogance followed by anger equals a certain transition.” The typeface would facilitate its own natural reading. Angry explosions in the middle of tense, but quiet conversations would come surprisingly, whereas angry explosions after arrogant tirades would facilitate a longer buildup.
And this sort of generative typeface could find its way into personal messaging as well: instead of smileys being relayed verbatim, they would change the typeflow of the surrounding sentences. A sentence with :p afterward would be set in a casual, joking font, whereas one with :( would convey gravity. Even the speed of typing could convey a typographic tone: slow, contemplated messages would appear in contemplative fonts and hasty, misspelled messages would be autocorrected but styled in a laid-back font.
A generative typeface of this design has a second bonus: it works the other way as well by facilitating better text-to-speech hints. The tags describing the tone provide clues to the AI in text-to-speech. It would be a new way to encode human language replete with its metadata, the space “between the lines” that people already read.
I’m not a type designer, but I see a great space for someone willing to take the time to design a truly adaptive typeface. If this post inspires you to create something, let me know!