Two Arrows and a Handshake: The Webmasters Still Running the Internet's Smallest Neighborhoods

At the bottom of a page you will likely never visit sits a small graphic of two arrows and the word next. No advertisement, no recommendation engine, no autoplay teaser. Just one site pointing at another because a person decided they should be neighbors.

Webrings are the part of the old web that never quite filed for bankruptcy. In the mid-1990s they were the connective tissue of the personal web — enormous directories of independent sites sorted into topical loops, each one carrying a little navigation bar that sent you forward, backward, or to a random member. Yahoo! bought the largest of them in 1998, which tells you how much they mattered once. Then search got good, blogs got feeds, social platforms got frictionless, and the rings went quiet.

Except a few dozen of them didn't. On hand-coded sites, Neocities pages, and small servers humming in somebody's spare room, the circles are still closing. You just have to know where the entrance is.

The mechanics of a circle

A webring is almost insultingly simple. Every member site carries the same snippet of HTML: a link to the previous site, a link to the next, a link to the ring's index, and often a link that flings you to a random member. That's the whole architecture. There is no feed, no ranking, no engagement metric. The ring is a list, and the list is the product.

The interesting part is what happens when a list gets old. Members disappear. Domains lapse. Someone's Angelfire page gets eaten by a hosting migration and the next-link on the site before it starts pointing into a 404. So every ring has a keeper — a ringmaster, in the old vocabulary — whose actual job is link hygiene. They email dormant members. They prune the dead. They renumber the circle so the arrows still connect end to end.

One keeper of a forty-one-site ring, who asked to be identified only by her handle, described the work as 'changing the oil on a car nobody drives.' She has been maintaining the ring since 2003, through four hosts and two laptops. The ring's topic is embroidery. She estimates that maybe six of the forty-one sites are updated more than once a year, and she does not consider that a problem.

The software question

Modern rings mostly run on two things: a scrap of Perl-era ring management software that still works fine, or a plain text file that the ringmaster edits by hand and pushes to a static page. A few have moved to GitHub-hosted JSON and a little JavaScript that injects the nav bar, which is about as much infrastructure as anyone wants.

The tech stack is not the point, and the people running rings will tell you so within about ninety seconds of meeting you. What they care about is the policy. Do you accept anyone who applies? Do you read the site first? Do you require that the member link back before they're added? Rings that accept everyone turn into link farms. Rings that read every application stay small, stay legible, and stay alive.

That review process is the entire difference between a webring and a directory. One is a circle of people. The other is a spreadsheet with delusions.

What a random button is actually for

Ask ring members why they bother and you get a version of the same answer: discovery should cost something. Algorithms are extraordinary at giving you more of what you already like. They are terrible at giving you the thing you didn't know to want — the personal site about municipal water infrastructure run by a retiree in Ohio, the hand-drawn comic that updates twice a decade, the fan page so specific it has a page dedicated to one character's eyebrows.

A webring is a curated random walk. You land somewhere with no idea why, and you have to actually look at it. There's no infinite scroll to rescue you. The page either earns your attention in ten seconds or it doesn't, and either way you're the one deciding.

'It's slower on purpose,' said one ring administrator who runs a small circle of synth-music hobbyist sites. 'Nobody's optimizing anything. There's no funnel. It's just a hallway with doors.'

The maintenance nobody claps for

There is no growth story here. Nobody is scaling. The rings that survive tend to hover between ten and sixty members, and the ones that get big tend to die young, because the keeper burns out and the links rot in a batch.

What keeps them going is the low cost of quitting being only slightly lower than the low cost of continuing. A ring's upkeep is maybe an hour a month. A domain is fifteen dollars a year. The bar for doing something thankless on the internet has always been low, and some people clear it anyway.

Which is, honestly, the whole ethos. The old web was not a product. It was a pile of people doing small things for each other with no expectation of return. The webring is that idea reduced to its simplest possible form: one link, pointing at a stranger, because you checked their work and decided they were good.

The next button as a vote

If you want to find one, the entrances are not hidden, just unfashionable. Search for the ring's index page, not the sites. Look at the bottom of hobbyist pages that look like they were built by a person instead of a template. Follow a footer. Follow a footer at the bottom of that footer.

What you'll find is a functioning parallel to the feed — smaller, weirder, and a lot more manual. It won't recommend anything you'd have found on your own. That's the feature. Somewhere out there a keeper is changing the oil on a car nobody drives, and the door is unlocked.

Signal from the static, as always.