Tiling window managers have a reputation: hardcore, fiddly, the domain of people who compile their own kernels for fun. That reputation is mostly wrong, and it gets in the way of a simple claim. For focused work a keyboard-driven tiler beats a mouse-driven desktop almost every time, and the reason has nothing to do with looking clever.
The mouse is a tax you pay on every context switch
Floating windows feel natural because they mimic paper on a desk. The desk metaphor is the problem. Real desks get cluttered, papers overlap, and you spend energy shuffling them. On a screen that energy is the trip your hand makes to the mouse, the hunt for a window edge, the drag, the release, the correction.
Individually those are nothing. Across a day of moving between an editor, a terminal, a log tail and a browser they add up to real friction, and each one arrives at exactly the moment you were about to think about something else. The cost is not the second it takes. It is where the second lands.
A tiling window manager removes the negotiation entirely. Windows do not overlap, they partition the screen. When you open something it takes its share of the space automatically, and you move between regions with a keystroke. Your hands stay on the home row.
The goal is not to look like a hacker movie. It is to make window management stop being a decision you make hundreds of times a day.
Pixels are not free, and neither is your attention
A floating setup quietly accumulates two kinds of waste.
- Wasted pixels: title bars, window chrome, the gaps between overlapping windows, and the wallpaper you can see but never use.
- Wasted attention: every window you can partially see but are not using is a small, persistent draw on your focus.
Tiling addresses both. Every pixel is allocated to a window that is doing work, there is no chrome you did not ask for, and because layouts are explicit you decide what is visible: a clean split between code and a test runner, with nothing peeking out from behind.
This is the contrarian part. More visual minimalism gives you more usable screen, not less. A 14-inch laptop running a tiler often has more effective working area than a large monitor buried under stacked floating windows, because the large monitor is spending its area on overlap.
i3 is a contract, not a toy
What makes i3 worth learning is that it is deterministic. You state the rules once, in plain text, and it obeys them. No drifting window positions, no wondering where a panel went. The config is the behaviour, and you can read the config.
Here is the slice that matters when you are starting out:
# the modifier every binding hangs off. Mod4 is the Super key
set $mod Mod4
# a terminal and an application launcher
bindsym $mod+Return exec --no-startup-id i3-sensible-terminal
bindsym $mod+d exec --no-startup-id rofi -show drun
# move focus with the home row, vim style
bindsym $mod+h focus left
bindsym $mod+j focus down
bindsym $mod+k focus up
bindsym $mod+l focus right
# split orientation: does the next window open beside, or below
bindsym $mod+v split v
bindsym $mod+b split h
# a resize mode you enter and leave deliberately, so h and l can mean
# something different for as long as you are inside it
mode "resize" {
bindsym h resize shrink width 5 px or 5 ppt
bindsym l resize grow width 5 px or 5 ppt
bindsym Return mode "default"
bindsym Escape mode "default"
}
bindsym $mod+r mode "resize"Read that and you understand the system. bindsym maps a key to an action, exec launches a program, and mode carves out a temporary keymap so a single key can mean "shrink this window" while you are resizing and "focus left" the rest of the time. There is no hidden state and no settings daemon holding an opinion you cannot see.
Where people quit, and why they should not
The honest objection: the first two days hurt. Muscle memory built over a decade of clicking does not transfer, and you will fumble. That is the part people quit on, and it is exactly the wrong place to quit.
The curve is front-loaded and short, and the payoff compounds for years. You are trading two uncomfortable days against a workflow that never asks you to reach for the mouse to arrange a window again. The comfortable option is not free either. It just bills you in smaller amounts, every day, indefinitely.
Defaults are destiny
The reason tiling stays niche is not that it is hard. It is that on most systems it is something you have to install, configure and debug before it does anything useful, and most people never get through that. A workflow gated behind an afternoon of setup is a workflow almost nobody adopts, however good it is on the other side.
The design position for 0dyssey is that i3 should be the default rather than a power-user opt-in, and that the unglamorous work is the entire point: a keymap that is sane before you have touched it, a config that stays a single readable file you can open and edit, a discoverable list of the bindings so the first hour is not guesswork, and floating on demand for the few applications that genuinely need it (image editors, video calls) so nobody has to fight the paradigm to do ordinary work.
None of that is running software you can go and use today. The operating system is in development and there is nothing to download, so read the paragraph above as a design position rather than a feature list: make the correct thing the easy thing, and lower the cost of the hard path until taking it is obvious.