Of the many passages in Ellen Ullman’s 1997 memoir Close to the Machine that stuck with me, one of the most memorable is about spreadsheets.
Ullman, a programmer, was describing her experiences working with end users in the early days of the web. “When I watch the users try the Internet,” she writes. “It slowly becomes clear to me that the Net represents the ultimate dumbing-down of the computer.” She describes the way users would click around, struggling to retrieve the information they needed—and when they failed to find it, rather than blame the tools, they blamed themselves.
When users showed her spreadsheets they’d created, however, she marveled at what she saw: complicated, ingenious constructions, with macros and links to databases, rich arrangements of data they’d designed to meet their own specific needs. “The spreadsheet presumes nothing,” she writes. “It is, literally, a blank sheet of paper with a notion of columns and rows—and everything held on that sheet is presumed to come not from the program but from the human user.” Where the web gave users narrow pathways, spreadsheets offered wide-open spaces. On the web, the user merely retrieved information; in the spreadsheet, they created it.
Spreadsheets obviously haven’t gone anywhere, and you could read these early web users’ frustrations as unfamiliarity with a wholly new tool. But there’s also something foundational in this passage: the idea that the average web user had to cede a sort of autonomy to use—and eventually to create—on the web. To get the answers they needed, they had to learn to ask a specific set of questions in a specific way; on platforms for content creation and social networking in the years that followed, technologists created specific sets of boxes for users to fill. The way these platforms were built inherently dictated the kinds of things users could create—and, perhaps, the things they even thought to create.
This isn’t new. The history of sharing both art and information has been defined by limitations, from ancient wax tablets to the invention of movable type to the development of audio and visual mass media. And there’s been plenty of research that supports the idea that creativity flourishes within limits—that the “blank sheet of paper” is often paralyzing in its wide-openness, and our brains really turn on when we’re given a small box to concisely fill. But on the web, content-creation platforms aren’t necessarily designed to foster creativity: Their limitations are often driving a specific set of metrics, from engagement to scalability to ad revenue, and these metrics sometimes directly work against the ability of users to make and share things safely and easily.
If we were to wipe the slate clean—no more platform-specific formats, no more slick UIs, no more engagement-capturing algorithms—would web users even know what to make online? The question has felt particularly acute these past few months, as Twitter users flounder to figure out where to go next, even as they still feel tethered to the increasingly broken platform. Setting aside the very real issue of building a critical mass of users on another site, the question of what to do on another site runs through many of these conversations. In an ideal world, what would a platform allow a user to do?
In fields like human-computer interaction, researchers use the term “affordances” to describe the properties of a platform. The concept has been used in the physical design world for decades: a coffee mug’s handle, for example, affords you to pick it up, and the shape of the mug’s body affords it to hold your coffee. Like their analog counterparts, the affordances of a digital platform are limited to the properties a user is aware of. “The key that a lot of people miss is that it has to be perceived,” says Michael Ann DeVito, a postdoctoral computing innovation fellow in the Department of Information Science at the University of Colorado Boulder. “So the system can do something, but if the user can't perceive it, it can’t.” Conversely, things that are not allowed by a platform—unintentionally or not—are “disaffordances”; things explicitly not allowed are often called “constraints.”
Creation on the web has always been about those constraints, whether technical limitations or the specific ways systems were designed. By the late '90s, the web had grown much more participatory than the one Ellen Ullman was writing about. With a little HTML and CSS, ordinary users could create all sorts of things on the proverbial blank page—so long it was mostly text, with maybe a few low-res images or the occasional sparkly animated gif. The first decade of the 2000s saw the rise of both social networking and blogging, but even as technical capabilities were rapidly expanding, for the average user it was far less of a free-for-all than the DIY spirit of the early years. The Web 2.0 shift to user-generated content centered the user—but it was on the platforms’ terms. And in an effort to make content creation as “user-friendly” as possible, platforms were once again, after the openness of the webring/Geocities era, building narrow pathways for users to take.
Tumblr was a pioneer in the sort of content-pathway fragmentation we see all over platforms today. While the big blogging sites of the early 2000s offered you a word-processor-like text box to which you could add other media elements, from the start, Tumblr signposted its affordances as it asked you to choose a direction: a text post, sure, but maybe an image instead, or a quote, or an audio file. Tumblr founder David Karp often spoke of how intimidating he found the standard blogging platforms. In 2011 he told one interviewer, “The longest one I ever kept up with was Blogger for I think three months, and what I always fell down at was that big empty text box that you got when you went to post. I just thought that’s such a burden, because I’m not a verbose person. I can write, but it’s work, and I don’t really enjoy it. What I really wanted to do with Tumblr was build a blogging platform that I could actually use. It was really about getting away from that big empty text box.”
The result was a platform that wasn't really about writing at all—anyone who’s spent any time on Tumblr could tell you that the site is an overwhelmingly visual one. This isn’t unique to Tumblr, of course. As smartphone use proliferated, and the web could more easily support images and videos, users were encouraged to think visually on all sorts of platforms, and over time the web’s omnipresence led to a broader cultural “pivot to video,” forcing users toward visual creation, regardless of whether that media made sense for their needs.
Of course, creativity often flourishes in digital platforms’ very specific formats. Vine’s six-second videos are perhaps the most famous example of innovation under extreme creative limits. In recent years, the quick technical uptake and endless experimentation of TikTok creators has shown that extraordinary things can be created with a relatively narrow set of features. And as for writing, maybe brevity really is key: Some research suggests we were all much better at Twitter in the 140-character days.
But constraints on the web today aren’t just about what our tools encourage us to do on a technical level—they’re also about what it’s like, more broadly, to use a platform. “On the old-school internet that I was on when I was a teenager, the constraints were the tools,” says DeVito. “Could you create a hit viral video in 1996? No, we did not have the technology and infrastructure to get that video distributed. For a one-minute video, you would spend two days uploading it, and nobody would have had the connection to download it. The systems didn't afford that kind of expression.”
But today, she explains, technical constraints are joined by constraints around things like moderation and audience. If you post something, will the platform allow it to stay up? And if it stays up, will that content open you up to harassment from other users? She gives the example of trans creators, whose art depicting themselves or their friends is often a particular target of both platform moderation tools and harassment from other users. “That starts to feel like this much bigger constraint,” she says. “Because you’ve got all these tools to build things with, and you have a system telling you, ‘Your expression is not welcome here.’ That’s not necessarily what they’re trying to say, but that is what it feels like every time.”
Online content creation today is inextricably entwined with these social components. DeVito talks about trans creators locking accounts or retreating to private digital spaces to share their work in a safer environment, which echoes behavior seen in many communities across the web in recent years, as users move from from big free-for-all platforms like Tumblr and Twitter to closed ones like Discord, or even leave the online world entirely. To DeVito, the question of whether current internet users would know what to do with wide-open spaces almost seems beside the point: “I think if Gen Z needed to go back to the old-school tools, they’d figure them out in less than a day and improve on them,” she says. “They’re clever.” But the current spaces, she explains, are known quantities: Flawed but clearly defined, users collectively share how to safely navigate them. “In that scenario, it wouldn’t be that we don’t know how to create,” she says. “It would be that we don’t know how to protect ourselves.”
The current moment feels like an inflection point for digital platforms across the web—far beyond Twitter’s woes, there’s a sense that people feel boxed in, even as they’re unsure what better spaces for creating and communicating might look like. Watching the discussions on any potential Twitter replacement, it’s easy to see competing—and sometimes wholly conflicting—desires and needs. Compare, for example, those who want the smaller, more controlled conversations of decentralized spaces to the creators who’ve built careers on scaled-up, engagement-driven sites. The technical points of friction joining a Mastodon instance are insurmountable barriers to some—and a central draw for others. Content policies on other proposed Twitter alternatives might limit hate speech but also punish people talking openly about gender and sexuality. No platform will solve everyone’s problems—but right now, it often feels like our current platforms aren’t solving anyone’s problems.
But can we solve our own problems? We do still create spreadsheets, after all. And the good parts of the social web—like being able to share and remix others’ work regardless of geography—still sit alongside the bad. The ideal solution likely lies in multiplicity: no massively scaled platform can do everything, so why continue trying to make one size only sort of fit all? Fragmenting our social and creative platforms wouldn’t just expand the ways we could share things with the world; a greater variety of affordances—and yes, constraints as well—would give us a greater range of pathways into creativity. As the current big platforms rush to copy each other (or, more to the point, copy TikTok), the idea of smaller, more varied platforms might feel antithetical; so, too, might the idea that the tech industry would be willing to invest in something that won’t endlessly grow. But the current platform malaise won’t be solved by scale and brute force. Users have many different needs, and in the next era of the web, they should be offered many different solutions.
I think back to Ullman’s end users, clicking around fruitlessly and then blaming themselves when they didn’t get the results they needed. In three decades of the web, we’ve all gotten good at working within limitations, but maybe more importantly, we’ve gotten better at recognizing when the problem is not the users, but the tools. It’s a lesson we can take into the next era of digital platforms—whatever the size and shape of the boxes we get to fill.