For most people, the internet is the web. You open Firefox, Chrome or whatever browser you happen to use, type something into the address bar and off you go. Websites, Google, YouTube, online shops, news sites, blogs, social media. It all happens in a browser, so internet and web have gradually become almost interchangeable words.
Except that they aren’t. The web is just one way of using the internet.
Once you start looking beyond the browser, you quickly discover that there are surprisingly many other ways to use this enormous network we call the internet. Some are decades old. Others are relatively new. Some have tiny communities compared with the web. But that doesn’t make them less interesting.
Before everything became a website
Take Gopher. Gopher dates from 1991, just before the World Wide Web really started taking off. Instead of websites filled with graphics, menus, advertisements, JavaScript and cookie pop-ups, Gopher mainly gives you simple hierarchical menus containing documents and links. It is very basic. Which, oddly enough, is precisely what makes it interesting today.
There is still an active Gopher world — usually called Gopherspace — containing personal pages, archives, articles, software and all sorts of other material. You obviously don’t browse those with Chrome. You use a Gopher client. I use Lagrange (https://github.com/skyjake/lagrange), for example, which makes exploring Gopher remarkably easy. Lagrange is actually useful for another alternative protocol as well, but more about that in a moment.
If you want to start exploring Gopher, Gopherproject at https://gopherproject.org/ is a useful starting point.
After using Gopher for a while, something interesting happens. You stop wondering where all the graphics are. You simply start reading.
And then there is Gemini
Perhaps even more interesting is that people aren’t only keeping old internet technologies alive. They are also inventing new ones. One of my favourites is Gemini. Gemini appeared around 2019 and deliberately takes a very different approach from the modern web. A Gemini page — usually called a capsule rather than a website — is mostly text.
The protocol and its document format are intentionally simple and leave very little room for the sort of complexity that has accumulated around the modern web. That means no giant JavaScript frameworks. No advertising networks following you around. No autoplaying video appearing halfway through an article. No newsletter window suddenly covering the text you were trying to read. Mostly, you get the information you came for.
Gemini is tiny compared with the web, obviously. But there are thousands of Gemini capsules containing hundreds of thousands of pages. That is more than enough to spend quite a bit of time exploring. You can use Lagrange to do that. It supports both Gemini and Gopher, which makes it a convenient way to start exploring this alternative internet without having to install several different clients.
But the numbers aren’t really the interesting part. What makes Gemini fascinating is the idea behind it. What if we looked at what the web has become and started again? And what if, this time, we deliberately kept things simple?
Spartan goes even further
There is also Spartan, a newer and even simpler protocol inspired by ideas from Gemini, Gopher and HTTP. The philosophy is much the same: keep the protocol extremely small and avoid gradually rebuilding all the complexity of the modern web.
Again, Spartan isn’t going to replace the web. That isn’t really the point. For decades, computing has largely worked from the assumption that more powerful computers and faster networks give us room to build increasingly complicated software. Faster processor? Great. Let’s add more code. More memory? Excellent. Let’s use that too. Faster internet connection? Even better. Now the page can download fifty additional scripts before showing me an article consisting of 800 words.
Projects such as Gemini and Spartan ask almost the opposite question. What could we build if we deliberately didn’t use all that power?
Or build a different network altogether
And then things get even more interesting. Because Gopher and Gemini still operate using much of the internet infrastructure we already know. But you can go further. There are projects that rethink parts of the network itself. Take Yggdrasil (https://yggdrasilnetwork.org/). Yggdrasil is an experimental decentralised routing system that can create end-to-end encrypted IPv6 networks. Computers running Yggdrasil can connect directly or establish peerings across existing IP networks. The result is something that sits on top of, or can operate alongside, the conventional internet infrastructure. That already feels very different from simply opening another website.
Then there is I2P or the Invisible Internet Project (https://www.i2p.net/). I2P is a decentralised, privacy-oriented network designed to allow people to communicate and publish information while making it much harder to determine where that information originates or where it is going.
There is also Hyphanet (https://www.hyphanet.org/), formerly known as Freenet. It is built around decentralised data storage and publishing and can be used to publish and access information without depending on a conventional central web server.
And those aren’t the only experiments. GNUnet (https://www.gnunet.org/en/) describes itself as a network protocol stack for building secure, distributed and privacy-preserving applications. Lokinet (https://lokinet.org/) takes yet another approach and uses onion routing across a decentralised network. It can be used both to reach the conventional internet and to access services that exist inside Lokinet itself.
None of these projects are simply “another browser”. They are attempts to rethink how computers find each other, communicate, exchange information and establish trust. And that makes them much more interesting than their relatively small user numbers might suggest.
Then there is Reticulum
An intriguing project is Reticulum (https://reticulum.network/). Reticulum is not really one alternative internet. It is a networking stack that can be used to build networks. And those networks don’t necessarily need a conventional internet connection. Reticulum can work across Wi-Fi and normal internet connections, but also across technologies such as LoRa and packet radio. It is specifically designed to keep working with very high latency and extremely limited bandwidth.
That changes the way you think about what a network actually needs. We have become accustomed to connections measured in hundreds of megabits or even gigabits per second. Modern websites are built with those connections in mind. Reticulum starts somewhere completely different. What if the connection is terrible? What if there isn’t a conventional internet connection at all? What if two computers can exchange only tiny amounts of data? Can we still build a useful network?
You don’t even need to leave the web
There is another possibility, though. Keep using the web, but strip it down. Install a text browser such as Lynx (https://lynx.invisible-island.net/), w3m (https://w3m.sourceforge.net/) or ELinks (http://www.elinks.or.cz/) and suddenly the web looks very different. Websites that depend heavily on JavaScript may become unusable, but plenty of other sites work surprisingly well. And you quickly discover something else: a lot of what we consider essential parts of a website aren’t actually necessary if all you want to do is read the information.
There are even search engines designed around this simpler web. Marginalia Search (https://marginalia-search.com/) deliberately prioritises non-commercial content and tries to help you discover websites you probably wouldn’t find through a conventional search engine. Then there is Wiby (https://www.wiby.me/), a search engine aimed at what it calls the classic web: simpler websites that feel much more like the web of years ago.
That means you don’t necessarily have to abandon the web to get away from the modern web. You can simply use a different corner of it.
Don’t forget RSS
Then there is RSS. RSS is a particularly good example because it is something many people used years ago and then somehow forgot about. I still use it every day to follow huge numbers of blogs and websites. It facilitates actually one of my (many) news feeds. Instead of repeatedly visiting twenty websites to see whether they have published something new, you subscribe to their feeds. Your RSS reader checks for updates. New article? It appears in your reader. No algorithm decides whether that article is interesting enough to show you. No social network inserts other content between the things you actually asked to follow. Use something like Feedly (https://feedly.com/) if you prefer a graphical interface. One of my favorite tools is Newsboat (https://newsboat.org/) that gives you access to RSS and Atom feeds using the terminal of your Linux laptop.
A completely different daily internet
Put a few of these tools together and something interesting starts to happen. Use Lagrange for Gemini and Gopher. Use Newsboat for blogs and news feeds. Use Lynx when you need the web. Perhaps experiment with Yggdrasil, I2P or Reticulum when you want to see what happens when people start questioning the architecture of the network itself. Suddenly, a surprisingly large part of your daily internet use can take place without a conventional graphical browser.
You could even do much of this on an old laptop running a very minimal Linux installation. Which, as regular readers of this blog can probably guess, sounds rather appealing to me.
It is not about replacing the web
None of this means I am going to stop using the web. Why would I? Just Me itself is a website. And there are countless things the modern web does exceptionally well that Gemini, Gopher or a terminal application couldn’t sensibly replace.
That isn’t the point. The important part is realising that the internet does not have to look like the web. It can be a Gopher menu. It can be a Gemini capsule. It can be an RSS feed displayed in a terminal. It can even be a plain-text document retrieved with curl. It can be an encrypted network running across Yggdrasil. It can be a decentralised service living inside I2P or Hyphanet. And it can even be a Reticulum network exchanging tiny amounts of data over LoRa.
Once you realise that, the internet suddenly becomes much bigger again. Not necessarily bigger in terms of the amount of information available. The web obviously wins that contest by several orders of magnitude. But bigger in terms of possibilities.