When you want to use Linux, you do so in two ways. You either use the command-line interface and type out every single command, or you use the graphical user interface, which is much more user-friendly (on average). However, there is a third way, and it's not talked about as much because it never had enough time to pick up steam.
"What 'third way'?"
Think about it: What gives you the benefit of a visual interface, without the bloat that comes with Wayland or X11? What mode of computing seemed to have come and gone so fast, that modern applications in its form seem noteworthy and bizarre because of the form factor alone? The Third Way is the text user interface, a.k.a., the TUI.
The True Middle Ground
The TUI experience is text-only, but is more approachable than the command line, despite technically being run inside of a CLI session. And because it doesn't require services like X11 or Wayland on the lower level, it frees up even more RAM for the system to use, while still giving an accessible UX.
You typically see it used in isolation: htop/btop, Midnight Commander, and distro installers like archinstall and slackinstall. What we've never seen is a unifying system that treated the TUI itself as the primary user experience. I think, in 2026, it's time to consider making the TUI a first-class UX for *nix OSs.
Modern-Day Considerations
I don't know if you've checked RAM prices lately, but they're at a premium. Users are now urging devs to optimize their applications and are largely working with what they have, because you never know when your memory could start to fail. Add in other factors, like the shift away from bloated frameworks whenever possible, and people are starting to realize something that a few of us already knew: Obsolete doesn't automatically mean bad.
Linux repos still contain GUI-based desktop environments and window managers from almost 30 years ago, yet there have been no notable TUI-based equivalents to them from any point in time. Why? Personally, I think it's because the concept was considered obsolete before it could gain traction at a time when "obsolete" meant "not worth navigating". GUI toolkits had evolved and been standardized early on, so all of the attention had been placed on them. Approaching the TUI in new ways was never considered, so because of that, it has a generational gap that stops us from using our devices in a different way entirely.
"Doesn't a tiling WM like i3 or sway accomplish this concept for the most part?"
Yes. However, they are still GUIs, not true text mode. They still have the same heavy graphical dependencies as any GUI, with a UX and aesthetic that is explicitly tied to the GUI, regardless of how hard you rice your setup. No matter how much you cut down a tiling WM, it's still not going to be as light as a pure TUI.
Obvious Flaws
Of course, there are drawbacks. A/V and graphics editing are out of scope, but modern-day web browsing could still be done... so long as you're a fan of the IndieWeb, since Electron apps are also out.
Gaming is mostly stuck to emulation. Both RetroArch and EmulationStation could be run without the need for either X11 or Wayland, though doing so does mostly break the TUI UX concept.
But the limitations are what make this "The Third Way": Of course the actual experience is going to have its differences, and is obviously more geared toward uses that don't involve graphics.
How To Accomplish?
Something like TUIOS. It's a great showcase for the concept, but still has visual dependencies. Keeping it in TTY, without the assumption of a mouse, and with "windows" taking up the full screen instead of a small "windowed" section would be how I would envision it. You'd have the main window, with ASCII panels at the top and bottom giving up-to-date system info. And to switch between running apps, just Alt+Tab like you would any other WM.
Linux is not short of TUI tools. They are pretty common and mature. What's missing is a unifying environment to bring them all together in a cohesive way, and hopefully inspire more TUI-native development. Then over time, the actual UX could evolve on its own, ideally without just becoming another GUI.
Could it be done? Absolutely. Could I do it? ...Check back with me in a few years.
Leave a comment, thanks for reading and stay tuned for more bullshit.
[ ~/third-is-the-way ]