Hi friends, it’s Matt.
At this point, it’s easy to take design systems for granted. For any interface spanning more than a few screens, creating a design system is usually on the checklist along with things like “test for bugs,” and “deploy to production.” But I think we’re coming to a moment where we’re going to have to re-think what a design system is and how it works. Some of the assumptions we make about design and the role of the designer just don’t work in a world where an app can show up on dozens of different types of devices, from postage-stamp-sized smart watches to the infinite expanse of the metaverse.
Needless to say, today’s essay is a long one. But before we get started, let’s turn on some tunes.
Skee Mask is the project of Bryan Müller, an electronic music producer from Munic. Müller already has a Best New Music under his belt — in the past 4 years he’s touched on a wide array of styles, from straight-laced trance tunes to ambient and atmospheric collages. “Dolan Tours,” from his latest release Pool, gives me some really good early 90s electro vibes; think Aphex Twin’s Selected Ambient Works, or Squarepusher’s “Theme From Ernest Borgnine.”
Ok, enough nostalgia. Here’s the essay! Since it’s a long one, you should read it on my website; I’ll kick it off here, but feel free to skip straight to the jump.
New kinds of computers and interfaces are becoming ubiquitous: in our cars, on our wrists, in our ears, at our desks at work, on our TVs at home. In response, we’ve seen innovations in the content, organization, and delivery of design systems: design tokens, interoperability, handoff automation, CSS-in-JS, and atomic styles, to name a few. But these may not be enough to match the pace at which new kinds of experiences are brought to market. As digital interfaces become more personal, more customizable, and more accessible, many design systems are being stretched thin, groaning under the weight of exponential possibilities.
That’s because, despite their name, most design systems aren’t all that much like systems. Granted, they are designed according to a system, and there’s a logical consistency to how their components and tokens are defined, but really, most design systems work like a dictionary: look up a component, get the instructions for using that component.
The most sophisticated design systems can provide instructions in multiple coding languages, with detailed documentation, with orderly, regular updates. But they’re still relatively static, needing a designer or engineer to interpret the documents and use them in the correct context. Today, when we put a design system to work, there’s no coupling; pushing one part doesn’t cause another to pull. There’s no interlocking or interconnection; no networks, feedback loops, or forces at play. And without those mechanisms, there’s no leverage, no simple machines multiplying force, no power in the output.
But there’s a glimmer of hope in design systems at the forefront of scale and complexity. It’s a new way of managing the growing diversity of users and interfaces. It has the potential to not only keep up with the current pace of innovation, but to enable new levels of customization and specificity.
In this essay, I’d like to analyze why the dictionary model of design systems is reaching its limit, then look ahead to imagine the tools and paradigms that are waiting for us just over the horizon.