Store, Forward, Repeat: The FidoNet Relays That Still Dial After Midnight
Photo by Photo by Navy Medicine on Unsplash on Unsplash
At 2 a.m., a modem in a spare bedroom in Iowa dials a number in Illinois. The call lasts less than a minute. A batch of messages moves one way, then another. The modem hangs up. Somewhere else, a node in Pennsylvania is waiting for its turn. No internet connection is involved. This is FidoNet, or at least the part of it that still answers the phone in 2026.
FidoNet was built to solve a simple problem: bulletin boards were islands. If you wanted to send a message from one BBS to another, someone had to carry it. FidoNet turned that into a routine. Systems called each other overnight, exchanged mail packets, and passed messages along a chain of volunteers. It was store-and-forward networking, and it ran on the public phone network.
A Network Made of Bedrooms
FidoNet started in 1984 with two systems and a lot of ambition. Tom Jennings, a programmer in San Francisco, wrote the software that let BBSs exchange mail automatically. The network grew into a worldwide hierarchy of zones, nets, hubs, and nodes. Zone 1 covered North America. Under it were regional coordinators, net coordinators, and individual sysops who agreed to follow a common addressing scheme and a common schedule.
The phone company was the backbone. Long-distance rates shaped the topology. A board in a small town might call a hub in the nearest city because that was cheaper than calling across the country. A hub might collect mail from a dozen downstream systems and forward it in one batch. The whole thing was a routing problem solved by people with phone bills and patience.
By the mid-1990s, FidoNet was enormous. It was also being overtaken by the internet. Once cheap, always-on connections arrived, the idea of waiting until midnight to exchange messages started to feel like a chore. Most nodes went dark. A few stayed. In 2026, the network is small, but it is not gone.
How Echomail Moves Without a Cloud
The modern FidoNet node still uses the same basic tools. A mailer handles the phone call. A tosser sorts incoming packets into the right message areas. Echomail is the public message traffic, the stuff that gets passed around in conferences. Netmail is private person-to-person mail. The node list is the directory that tells each system who to call and how to reach them.
A typical exchange goes like this: a node collects outbound messages, compresses them into a packet, and waits for its scheduled window. When the window opens, it dials its hub. The two systems authenticate, exchange packets, and disconnect. The hub later forwards the mail upstream or passes it to other downstream nodes. The message might take a day to cross the country. It might take three. Nobody expects instant delivery.
That delay is not a bug. It is the architecture. Store-and-forward means each system only has to be available for a short time. It does not need a permanent connection. It does not need a static IP address. It does not need a cloud provider. It needs a phone line, a modem, and someone willing to keep the schedule.
The Node List Is the Map
The node list is the part that makes FidoNet feel like a real network rather than a collection of hobbyists. It assigns every system a number: zone, net, node, and sometimes a point. It lists the phone number, the sysop name, the flags, and the routing. Coordinators maintain it. In 2026, those coordinators are volunteers, often the same people who have been doing it for decades.
The list is smaller than it used to be, but it still matters. If your node is not in the list, you are not really part of the network. You are just a board with a modem. The discipline of keeping the list accurate is part of the appeal. It is a small civic duty, and it keeps the whole thing from collapsing into a pile of misdialed numbers.
The hobbyists who run these nodes are not trying to bring back the 1990s. They are maintaining a piece of infrastructure that still works. One operator in the Midwest describes it as a way to keep his hand in. Another says he likes that the network has a clear boundary. You are either in the node list or you are not. There is no feed. There is no recommendation engine. There is just mail, moving on a schedule.
Why Bother in 2026
There are easier ways to send a message. That is obvious. But FidoNet offers something the easier ways do not. It is offline by default. It is small enough that bad behavior gets noticed. It has no ads, no algorithmic ranking, and no engagement metrics. It also has a kind of ritual that appeals to people who like their technology with a little friction.
The overnight call is part of that ritual. You set up the schedule. You watch the log. You see the packet arrive. You know exactly how the message got from there to here. There is no mystery about the cloud because there is no cloud. The network is visible, and it is held together by people who answer the phone.
That does not make it superior. It makes it different. FidoNet is not a replacement for the internet. It is a reminder that the internet is not the only way to connect. A phone line and a shared convention can still do the job, as long as someone is willing to keep the line open.
The Last Mile Is a Phone Cord
The last mile of FidoNet is not fiber or coax. It is a phone cord. In 2026, that cord connects a small number of hobbyists who still believe in store-and-forward networking. They are not many, and they are not loud. But they are still dialing after midnight, moving mail one batch at a time, and proving that a network is ultimately a promise that someone will pick up.
