Dan Rhatigan on Variable Fonts and the Future of Typography
Last fall, Adobe, Apple, Google and Microsoft jointly announced a specification for variable fonts, a potentially dramatic reinvention of the way fonts are digitally constructed and delivered. In his announcement post at Typekit’s blog, Adobe Head of Typography Tim Brown imagines “a single font file gaining an infinite flexibility of weight, width, and other attributes without also gaining file size.”
This will essentially allow a much greater range of typographic options for designers like you and me. For instance, if you need to condense or extend the widths of certain characters, adjust a font’s x-height slightly, or even subtly modify the sharpness or roundness of letterforms, a variable font will allow you to do that while tapping into the built-in rules and logic encoded by the type designer. The end result will be richly expressive typography customized in accordance with the aesthetic intentions of the designer of the typeface itself—no more ill-informed modifications using the brute force of vector editors. This animated GIF by Dutch type designer Erik van Blokland hints at some of the possibilities.
The specification is just the first step towards building this glorious future, though. I talked to my colleague Dan Rhatigan about what it will take for us to get there. Rhatigan is a Senior Manager for Adobe Type and a veteran of the industry, having worked as a typesetter, graphic designer and typeface designer and teacher, including stints in London and New York as Type Director for Monotype. He is a key figure in this critical early stage for variable fonts, having participated in the development of the specification and also now spending time helping to turn its promise into reality.
Khoi Vinh: My initial understanding of the variable fonts announcement was that it’s multiple master fonts done right. Maybe we can start there—what did the original multiple masters fonts specification get wrong?
Dan Rhatigan: Variable fonts build on the ideas of a couple of older font formats: not just Adobe’s Multiple Masters but also Apple’s OpenType GX.
It’s a shame that Multiple Master fonts never lived up to their potential as commercial products, because the idea of letting users pick their own preferences within a given typeface’s range of options was a great one. In reality, though, it was a clunky way to work. As a user, you got a file that contained all these possibilities, but you had to use Adobe Type Manager (I think I may have also used a QuarkXPress plug-in) to play around and choose your desired mix of options, then spit out a new font file with the results. You couldn’t get a live preview in your documents, and then you had to manage all these files you’d create. I played around with Multiple Masters when I worked in publishing, and it was a hassle to keep track of the custom styles everyone on the team would create.
There’s an elegance to this new variable font model: you use a single file all the time, and rely on your browser or software to generate the desired result out of that file. The trick now, though, is getting all the software we use to implement good controls for making all those possible adjustments.
The great legacy of Multiple Master fonts, though, has been that they changed the way type designers work. Most of us type designers have been designing big families using the organizing ideas of the technology: you draw the key weights or widths or other styles, and allow your font tools to interpolate the styles in between. So at least we have an entire generation that learned to think about designing type along the same principles we need to produce variable fonts.
Is it right to say that there hasn’t been a lot of innovation in type technology since then, or have there been significant changes happening out of the public eye?
OpenType was the last really major development in type technology—twenty years ago! Even web fonts have been more of a revolution in how we use and deliver fonts, rather than in the fonts themselves. This is a big challenge when it comes to the major leaps we’re seeing now, like variable fonts, like SVG color fonts (both considered extensions of the OpenType specification). Fonts are software, after all, and the pace of development was bound to catch up with the increasing sophistication of all our other software. It’s really the demands of the faster, more powerful digital world around us that’s requiring our fonts to become smaller and that’s requiring a more sophisticated means of delivering typeface designs to people.
In software terms, twenty years is a basically an ice age. Are we in for a major disruption with variable fonts? What’s going to change?
We might have a major disruption, depending on how much other pieces of software are willing to adapt their typesetting controls. A lot depends on how well the implementation helps people understand the new capabilities of the fonts. Applications might choose to stick with a simple mode that just presents a variable font like it’s a typical font family, and ignore the potential of working with the in-between areas of the design. What I hope, though, is that that apps will see that this could be a real opportunity to handle type differently, with better automation of things like copyfitting or size-specific adjustments, or giving users a way to create and manage their custom type styles.
I think the first real disruption will happen with the people tinkering with CSS and variable fonts, who will be the ones to explore the possibilities a little more freely at first. Besides, the compact file sizes will make it much easier to deliver web fonts if they’re variable fonts, so we may finally see a broader range of styles on a page, and much more variety for sites with Chinese, Japanese, or Korean text.
First, though, there is still work to do at the infrastructural level: enabling support for variable fonts at the OS level, and then at the level of app-specific rasterizers, and then UI controls…
There’s a lot to unpack there that I want to get back to. But first let’s talk turkey: how will this change the way graphic and UX/UI designers use fonts? Will there be a learning curve?
There are already lots of fonts available, but not many ways to use them. The easiest way to play with variable fonts is to visit axis-praxis.org using Webkit or Chrome Canary in the latest version of macOS. It’s a test site using a couple of variable fonts already lurking in macOS, as well a number available through various Github projects.
Other than building web pages to test the fonts in experimental browser builds, there may not be many ways to work with variable fonts until other pieces are in place toward the end of this year, and then in 2018.
What are those major pieces? Is it principally support from the major operating systems that we’re talking about?
Exactly that. The specification for variable fonts is relatively fresh, so work on the rasterizers—the underlying support for interpreting the fonts—is underway, and then that has to be integrated into OS updates, so that browsers and other apps can draw upon that support.
Okay let’s suppose that that comes along next year. You said that we might see some applications present variable fonts simply, perhaps not exposing the full richness of what they can do, and maybe others will provide an interface that handles type very differently from what we’re accustomed to today. This sounds to me like a we could have a situation where there are a thousand different font selection user experiences, with each app handling it radically differently, emphasizing various aspects of the standard. Is that right? Or is there a possibility that we might get a single, uniform user experience for handling type? And I present these options without judgment—clearly there are pros and cons to both.
However, the likelihood of a variety of new approaches to application UI both excites me and scares me a little. It would be great for applications to rethink how they can help users improve their typography, but I know that people often resist change to controls that they’ve used for ages. I think a delicate balance is needed: an updated experience that leads people to understand the capabilities instead of just dumping new options in front of them. But I think there’s also a lot of potential in accessing some of the capabilities variable fonts under the hood—automating use of axes (when available) to control optical sizing, or weight and width depending on context in a document, maybe.
That’s an interesting possibility: the idea that all of these new options might actually impose a responsibility on application UI designers to help users improve their typography, not just simply give them access to fonts. Is there anything in the variable fonts spec that can alert a user when a particular variation has gotten too tall, or wide or just plain ugly? Should there be? I’m only half joking.
Sort of in-font alarm system? No, but there are built-in controls. Type designers need to specify minimum and maximum values for any design axis in a font, so some reasonable parameters are inherent in the font format. Anyone who uses the font, however, will still need to show some good judgement about what they use, within that “safe space” determined for each typeface. This increases the type designer’s responsibility, though, to make sure that they anticipate and resolve what can happen in the design space where different axes may intersect. That is, if there are axes for weight and width available to the user, then the designer needs to make sure all the possible outcomes that are allowed will work well.
How about the business implications? This is a lot more work for type designers. Are we going to see fonts get more expensive as a result?
Honestly, the business outcomes are still anyone’s guess. Every type designer I’ve talked to about this has had a different theory, but since we’re a ways off from widespread support, it will be a while before foundries test the waters and some patterns emerge.
Fair enough, but if you’ll allow me I’m going to press you a bit on the state of font licensing, which in my opinion serves neither type designers nor type customers particularly well. Most font licenses are restricted to a finite number of devices, which seems antiquated when fonts—like most contemporary software—should really be a kind of service, distributed on a usage basis via the cloud. Now we’ll soon have variable fonts, which sounds like a fundamental sea change in the mechanics of the technology—which in turn suggests a once in a generation opportunity to rethink the way fonts are licensed. Put another way: does it make sense to license next generation fonts with last generation terms?
That model of licensing is just one of many that are used now, though. Web fonts and apps, in particular, shook up a lot of the industry’s assumptions about licensing already, since actual font data is distributed. Subscription services—like Adobe Typekit, for instance—have already been gaining ground. Variable fonts will surely add another wrinkle to licensing models, since they have so much potential capability. So far, I’ve heard plenty of theories and ideas how they might be treated, but no consensus, since no one has yet to see what the value of the fonts in use will really be.
So a lot is up in the air, which is exciting but also somewhat fraught. Do you believe that the value that variable fonts brings to designers makes their success inevitable? How would you rate the chances?
The big players have plenty of reasons to want variable fonts for their compression and programmability, so one way or another they will at least find their niche. While there may also be lots of potential on the web, I worry that it will come down to how apps handle variable fonts for creative work that will make or break them as widely used commercial products. It will probably come down to UX decisions that make them vital design tools rather than a specialized technical solution.