Analysis Example: Infinite Scroll
I wrote this after preparing for our lecture on data, information, and knowledge. I was thinking a lot about the process of consumption on the internet, the way it addicts us, and the way that design choices that for-profit organizations make end up shaping what we consume.
I remember the world without the internet. It was a quiet place; my information landscape was my friends, the monthly magazines I subscribed to, and our local library. Sometimes I'd find a book of my mom's and read it, not because I was interested in it, but because I just wanted something to read, and there was nothing else. These pre-internet days afforded a kind of situated serendipity — a collection of information curated by the people in my lives, my local community, and my own purposeful parent-mediated decisions about what media to access. These data-information-knowledge pathways, where highly curated, finite, and slow.
I remember the world with the nascent internet. It was slow too, and felt bounded. The Yahoo! and AOL landing pages were browseable indices of a fixed number of articles. I would read the ones that were interesting, kind of like a daily newspaper, and when I ran out of content, I would go ride my bike, play basketball with friends, draw something, or talk about the latest episode of Animaniacs. And at times, the 56 kilobytes per second that traveled through my phone lines to my modem to my computer took just long enough to trigger boredom, and I would just lose interest. This was a new serendipity to be sure, as it connected me to the curators at Silicon Valley content teams — but it was still finite, and still slow.
I remember when Google was new. There were many search engines in the 1990's that did what Yahoo! and AOL did, curating content for visitors, while providing some basic search functionality for discovering new content. Lycos was the most advanced, indexing most of the public internet. But finding relevant content meant digging 10 pages deep, since so much of the public internet was garbage. But Google, building upon a century of bibliometrics research, had found a better metric for relevance: citations. Suddenly search was a single box, a single page, and maybe the first few results on that page, nothing more. My information landscape suddenly shifted from slow, curated, and situated, to fast, global, and context free.
Change was swift. Searches went from taking one second to one tenth of a second. My deliberation over my query was replaced with query completion. My search results were replaced with summaries of pages, extracted images, derived "facts" from a web of "knowledge". The next button at the bottom of the page that allowed me to dig deeper into search results was replaced with infinite scroll. These design changes had two clear goals: get me to content faster, and show me as many ads as possible along the way.
But of all the changes, infinite scroll was perhaps the most influential. Google did not invent it. One of the earliest implementations emerged in the 1990's, when researchers across the U.S. were experimenting with new graphical user interface paradigms. One, for example, was by Wayne Christopher, a researcher at UC Berkeley, funded by a U.S. Department of Defense DARPA project. He worked on a system called Ensemble-C (Christopher, 1990), which used an infinite scroll model to browse documents. Another, by George Furnas at the University of Michigan, explored infinite canvases of information, and the many challenges of their interaction design and implementation (Furnas, 2000). Aza Raskin, son of a human-computer interaction expert who conceived of the original Apple Macintosh, was hanging in these research spaces, watching these innovations unfold. This ultimately led him to adapt these ideas for the web while at a small user interface startup in 2006 (Atul, 2006). It caught on, Google implemented it for search, and then every other content platform implemented for their seemingly endless collections of content. Infinite scroll was born, and the web was hooked.
But this feature was the one that most broke my brain, and others'. The problems were numerous:
- It's addictive. It makes it much harder to know how much content one has consumed, and removes an opportunity to pause, reflect, and self-regulate one's engagement. This, in turn, leads to more consumption.
- It deters mindfulness. No matter how good someone is at setting limits, it replaces the decision to get more content with the implicit navigational gesture of scrolling, increasing the need for users to impose their own structure to regulate their attention, since the interface is no longer doing it for them.
- It's inaccessible. For anyone who relies on a screen reader, it hides footers, it makes it hard to navigate to the end of the page, it constantly changes the structures of pages, and there's no easy to return to results previously seen.
These flaws, of course, come with one advantage: profit. With most web companies' revenue being tied to attention, the more ads can be served to visitors, and the more money companies make. This, therefore, like so many things, is a story of the cost of capitalism. More attention, more content, more ads, more money, all without any attention to the consequences on human wellbeing or societies capacity to know.
So what of my capacity to know? Am I better off in a world with infinite content, one that draws me into its depths, with a promise and illusion of boundless knowledge? Or would I be better off with that quiet place where I began my childhood, where I may have known less, but had control over my attention, and a life full of connecting in spaces and places, not around content, but around the shared interests of my friends, family, and community? I'm not willing to say that today's world is categorically worse — I've certainly benefited from finding community I never could have found as child — but at least for me, it is mostly worse.
References
Atul (2006). Reading, Humanized. https://web.archive.org/web/20060518084033/http://www.humanized.com/weblog/2006/04/28/reading_humanized/, retrieved September 26th, 2023.
Christopher, W. A. (1990). Constraint-based document presentation. University of California, Berkeley, Computer Science Division.
Furnas, G. W., & Zhang, X. (2000). Illusions of infinity: feedback for infinite worlds. In Proceedings of the 13th annual ACM symposium on User interface software and technology (pp. 237-238).