Wikipedia Deep Dive

Usenet

13 min read

The Ghost Network That Built the Internet

Before there were upvotes or likes, before anyone had a Twitter handle or a Reddit username, before Facebook existed or Mark Zuckerberg learned to walk—there was Usenet. And here's the strange thing: it's still running.

Usenet launched in 1980, more than a decade before most people had ever heard the word "internet." It pioneered concepts we now take for granted: threaded discussions, newsgroups organized by topic, the ability for strangers across the world to argue about wine yeast or debate the existence of God. It also gave us some of the internet's most enduring vocabulary. The word "spam" for unwanted messages? Usenet. "Flame" for an angry online attack? Usenet. "Frequently Asked Questions," now abbreviated universally as "FAQ"? That's Usenet too.

The system works on a principle that sounds almost quaint today: there is no central server. No company owns Usenet. No single administrator controls it. Instead, thousands of independent servers around the world store copies of messages and share them with each other, like a vast game of telephone played by machines. When you post a message to a Usenet newsgroup, your local server passes it to its neighbors, who pass it to their neighbors, and within hours your thoughts have propagated across the planet.

How It Actually Works

Imagine you run a small coffee shop with a bulletin board by the door. Customers pin up notes—job listings, apartment rentals, lost cat flyers. Now imagine that every coffee shop in your city has a bulletin board, and every night, the shop owners meet to share copies of each other's notes. Soon enough, every bulletin board has the same content.

That's essentially Usenet.

The technical term for this is a "flooding algorithm." Each server automatically forwards new messages to every connected server that hasn't seen them yet. The servers are called "news servers" (a name that has nothing to do with journalism—"news" here just means new posts). They communicate using something called the Network News Transfer Protocol, usually abbreviated as NNTP.

The messages themselves are called "articles," and they're organized into categories called "newsgroups." If you want to discuss science fiction literature, you might join rec.arts.sf.written. Computer programming in Python? comp.lang.python. The naming convention follows a hierarchy: the first part indicates the broad category, and subsequent parts narrow it down.

The major categories—called the "Big Eight"—cover most of human interest:

comp.* for computer-related discussions
humanities.* for arts, literature, and philosophy
misc.* for everything that doesn't fit elsewhere
news.* for discussions about Usenet itself
rec.* for hobbies, recreation, and entertainment
sci.* for scientific topics
soc.* for social and cultural discussions
talk.* for controversial topics like religion and politics

Then there's the wild card: alt.*

The "alt" hierarchy—short for "alternative"—operates outside the formal rules governing the Big Eight. Anyone can create an alt newsgroup, which means it contains both hyperspecific communities of genuine enthusiasts and complete chaos. Alt.binaries, a subsection dedicated to sharing files rather than text, grew so massive that it now accounts for over ninety-nine percent of all data flowing through Usenet.

The Birth of Digital Community

In 1979, two graduate students at Duke University—Tom Truscott and Jim Ellis—had a problem. They wanted to share information with colleagues at the University of North Carolina at Chapel Hill, but the ARPANET (the Pentagon-funded precursor to the internet) wasn't available to them. They had access to something far more modest: UUCP, which stood for Unix-to-Unix Copy Protocol. It was basically a system for computers to call each other over phone lines and exchange files.

So they built something on top of it. They called it Usenet, a portmanteau of "User's Network," with an optimistic hope that the USENIX organization—a Unix users' group—would eventually take over its operation.

The system was designed for a world where networks were slow, expensive, and unreliable. Many early Usenet sites connected only once or twice per day, usually at night when long-distance phone rates were cheaper. Messages didn't arrive instantly—they propagated gradually across the network over hours or days. But that was fine. People weren't expecting real-time chat. They were expecting something more like a global bulletin board that updated itself.

The first newsgroup was called NET.general, quickly renamed to net.general. Discussions were organized into threads: someone would post a question, others would reply, and those replies could spawn sub-discussions that branched off like tributaries from a river. Sound familiar? That's because every modern forum, from Reddit to Stack Overflow, borrowed this structure from Usenet.

Anarchy by Design

Here's what makes Usenet philosophically interesting: it was built to be uncontrollable.

When you post on Twitter, your message lives on Twitter's servers, subject to Twitter's rules. When you post on Reddit, a moderator can delete your comment. But when you post on Usenet, your message replicates itself across independent servers around the world. No single administrator can remove it everywhere. No company can shut it down.

This was intentional. The creators wanted a system where information could flow freely, without gatekeepers. Individual server operators could choose what to carry—a server in Germany might refuse to host hate speech, a university server might skip the binaries to save storage space—but users could simply connect to a different server that carried what they wanted.

The result was a kind of information ecosystem governed by natural selection rather than central planning. Content that people wanted to read proliferated. Servers that provided good access attracted more users. The whole thing held together through mutual self-interest and informal cooperation.

This architecture had consequences, both wonderful and terrible.

The Good, the Bad, and the Spam

Usenet in its heyday was remarkable. Imagine a world where academics, hobbyists, programmers, and random curious people could find each other and discuss anything, completely free of charge, with no advertising, no algorithm pushing engagement, and no corporate overlord deciding what you should see.

The discussions could be extraordinary. Experts in narrow fields would patiently answer questions from beginners. Flame wars over programming languages would rage for weeks with a passion that made clear how much the participants cared. Somewhere, right now, there are probably still archived debates about the merits of Emacs versus vi that date back forty years.

But the same openness that enabled great discussions also enabled abuse.

In 1994, two immigration lawyers named Laurence Canter and Martha Siegel became infamous by posting an advertisement for green card services to every newsgroup they could find. This was the first commercial spam on Usenet. The reaction was furious—users flooded their inbox with complaints, their service provider kicked them off—but the precedent was set. Spam worked, and spammers followed.

The "sockpuppet" problem emerged early too. That's the term—coined on Usenet—for a fake identity used to give the impression of grassroots support or to evade bans. When your identity is just an email address, and email addresses are cheap, nothing stops you from pretending to be multiple people.

Why Usenet Matters to History

To understand why Usenet is historically significant, you need to understand what came before it: essentially nothing.

Before Usenet, if you wanted to discuss a topic with people outside your immediate physical community, your options were limited. You could write letters. You could make expensive long-distance phone calls. You could maybe find a local club or organization. But connecting with strangers who shared your obscure interest in, say, science fiction fandom or Unix system administration or winemaking? Practically impossible.

Usenet changed that. For the first time, geographic isolation didn't matter. You could live in rural Montana and participate in discussions with programmers in California, academics in Boston, and hobbyists in London. The only barrier was access to a computer with a modem and a connection to a Usenet server.

In the late 1980s and early 1990s, before the World Wide Web existed and before internet access became affordable for regular people, Usenet was often the only way to reach a global audience. Some users connected through FidoNet, a network of dial-up bulletin board systems that exchanged Usenet messages. They weren't on the internet at all—they were using phone lines and cooperative forwarding—but they could still participate in worldwide discussions.

This created a culture. Usenet users developed norms around quoting previous messages (using the ">" character), around writing clear subject lines, around not feeding trolls. They invented the FAQ document as a way to prevent the same basic questions from being asked endlessly. They created netiquette, the informal rules of online behavior that persist in mutated forms across the modern internet.

The Technical Details, Made Somewhat Comprehensible

If you've ever used email, you already understand most of how Usenet works technically. Articles on Usenet look almost identical to email messages: they have headers (From, Subject, Date, and so on) and a body containing the actual content. The difference is distribution. An email goes to specific recipients. A Usenet article goes to a newsgroup, which means it's stored on servers and made available to anyone who subscribes.

Servers exchange articles using the Network News Transfer Protocol. Standard connections use TCP port 119. Encrypted connections use port 563. When you post an article, your local server assigns it a unique identifier and begins sharing it with other servers. The flooding algorithm ensures it eventually reaches everywhere—though "eventually" might mean minutes or hours, depending on how many hops away a server is and how frequently servers synchronize.

There's a formal specification for how this all works, documented in technical standards called RFCs (Request for Comments, the internet's equivalent of official documentation). RFC 850 was the first Usenet specification, later replaced by RFC 1036, and now superseded by RFC 5536 and RFC 5537.

One curious feature of Usenet is the "cancel message"—a special article type that tells servers to delete a previously posted message. In theory, this lets you unsend a post. In practice, it's rarely used because anyone could forge a cancel message for anyone else's post. Many servers simply ignore cancels now to prevent abuse.

Moderation, or the Lack Thereof

Most Usenet newsgroups are unmoderated. Post something, and it propagates immediately to the world. There's no approval process, no waiting period, no human gatekeeper.

The advantage of this is obvious: speed and freedom. The disadvantage is equally obvious: spam, abuse, and off-topic noise.

Some newsgroups are moderated, though. In a moderated group, posts don't go directly to Usenet—they're emailed to a moderator who reviews them first. If approved, the moderator adds a special header and injects the article into the system for propagation. This slows things down but keeps discussions focused.

Creating a new newsgroup in the Big Eight hierarchy requires a formal proposal process. You submit a Request for Discussion, specifying the group's name, whether it will be moderated, and ideally a charter explaining its purpose. The Big-8 Management Board reviews these proposals and votes on whether to approve them.

The alt.* hierarchy has no such process. If you want alt.fan.your-favorite-obscure-topic to exist, you just create it.

The Slow Decline

Usenet is still running. You can still post to newsgroups today. But let's be honest: its cultural relevance peaked decades ago.

What happened? The World Wide Web happened.

When web forums emerged in the mid-1990s, they offered something Usenet couldn't: a graphical interface that didn't require special software. You could discuss topics in your web browser, the same application you used to read news and shop online. No need to configure a newsreader, find a news server, or understand arcane protocols.

Then came social media. Facebook, Twitter, Reddit—these platforms made online discussion trivially easy. They also introduced features Usenet never had: user profiles with photos, like buttons, algorithmic feeds that surfaced popular content.

Internet service providers, once reliable sources of Usenet access, gradually stopped offering it. Running a news server is expensive—the sheer volume of data, particularly in the alt.binaries hierarchy, requires massive storage. And for most users, the effort wasn't worth it when alternatives existed.

Today, Usenet survives largely in two forms. Text-based discussion groups still exist, with small but dedicated communities. And alt.binaries remains active for file sharing—a legacy use case that has outlasted most others.

Google Groups provides a web interface to Usenet, making it easier to access for casual users. But the culture has changed. The long-time Usenet community often views web-based gateways with suspicion, sometimes filtering messages that come through them.

What Usenet Taught Us

Looking back, Usenet was an experiment in decentralized communication that mostly worked. For over a decade, it was the primary way that geographically distributed communities formed online. It proved that people wanted to connect over shared interests, that they would participate in thoughtful discussions without being paid, and that a global communication network could operate without central control.

It also taught lessons about the dark side of openness. Spam emerged on Usenet before it emerged in email. Flame wars, trolling, and sockpuppetry were Usenet phenomena before they were social media problems. The tension between free expression and community standards—the same tension that defines content moderation debates today—played out on Usenet first.

The vocabulary Usenet created persists everywhere. When you see "FAQ" on a website, that's Usenet's legacy. When someone describes an email as "spam," they're using a term that was popularized on Usenet (borrowed from a Monty Python sketch about the canned meat, applied to unwanted posts that flooded newsgroups). The threading model that every forum uses—reply to a post, and your reply appears under it—was standard on Usenet before the web existed.

Perhaps most importantly, Usenet demonstrated that the internet wasn't just a tool for governments and universities. It could be a medium for human connection, argument, creativity, and chaos. Everything that followed—Geocities, LiveJournal, MySpace, Facebook, Twitter, Discord, Reddit—inherited something from those early newsgroups where strangers first learned to talk to each other online.

A Strange Kind of Immortality

The really remarkable thing about Usenet is that it hasn't died. It's diminished, certainly. It's no longer culturally central. Most internet users have never heard of it. But the servers still run. The messages still propagate. The protocols still work.

In 1980, two graduate students rigged up a way for computers to share messages over phone lines. More than four decades later, that system is still functioning, still carrying discussions, still quietly demonstrating that sometimes the old ways persist even as the world moves on.

Whether that's admirable or just stubborn is perhaps a matter of perspective. But the next time you post a comment on Reddit, argue with a stranger on Twitter, or browse through a FAQ document, you're participating in traditions that trace back to those early newsgroups. The ghost network is still with us, even if we've forgotten its name.