I feel like my eyes can only look at one thing at a time. I just have shortcuts to switch between programs.
Why do you prefer using a tiling WM and how do you use the tiling functionality in your workflow?
Main reason for me is because I don’t have to manage the position or size of apps I open.
Web dev here.
- editor + web browser + devtools when working on frontend (workspace 3)
- editor + tests + another terminal for whatever when working backend (workspace 2)
- server terminal + lazydocker for both (workspace 1)
- web browser with work related tabs + todo list + notes app on workspace 9
- chat apps and email on workspace 10
- long-running jobs and performance monitors on workspace 8
- 4-7 are used for whatever
- music on scratchpad
Tiling (as well as stacking) make it manageable to have a bunch of windows open with a minimum of fuckery. Sure I can only read one at a time, but when coding for example I’m rapidly switching between the code and the result. I can have a text editor, browser, and devtools accessible as fast as I can think, and I spend very little effort arranging windows.
Also, a good tiling window manager replaces the need to learn a bunch of windowing features for other apps. My devtools open in a new window, I don’t use tmux or my terminal’s split features, and I generally have a bunch of browser instances open because my window management handles it all, better.
I’m also a web dev and I have a similar workflow where each workspace has its purpose. Except I never tile anything. I do have shortcuts that switch to specific windows but I never tile anything. That way my eyes are always on the center of the screen.
I very rarely need to tile windows next to eachother. So rarely that I just don’t see the point in making it the main feature of my WM/DE.
Very interesting to read though. Thanks for the thorough example.
¯\_(ツ)_/¯ whatever works for you, but that sounds painful to me. Why is only using the center of the screen so important?
I’ll clarify that I only use the more complicated layouts on my big monitor at home; when I am on a laptop it’s single window or side-by-side for the most part.
I am just a regular user but on a smaller screen ie. 13 inch 1440x900 I just have a single maximized window visible even if I have multiple apps open like web browser because I can’t see anything if I have multiple windows.
Do you have a small monitor?
In my opinion, on a >32 4k or 1440p display, the full screen is just way too big for a single window. Which isn’t a problem, because as easy as it is to switch between two windows, it’s even easier not to. Especially for things like having a web browser and dev tools, switching back and forth every time I tweak a CSS rule would be agonizing.
I like having a big window open. What annoys me with tiles is that the center of my screen which is the most natural to look at is now just a intersection of the other screens so I have to move my eyes over it just feels uncomfortable.
Sounds like your screen is too close to your face.
Don’t need the mouse. Nothing ever gets tucked behind anything else. dwm allows you to put one window on multiple tags (desktops) in multiple window arrangements. Also it’s fully scriptable. Why NOT use a tiling WM?
You really hit the nail on the head here. Never having to take take your hands off the keyboard, while always having windows take up exactly the right amount of room, is the main reason I hate having to use non-tiling WM.
And your other point is spot on, too. Any workflow that you use in a standard WM you can also do in a tiling WM, except (imo) more easily. And there are lots of workflows that are agonizing without tiling functionality.
I want to read this book full screen. Hang on, didn’t that other book say something different about this? I want to open it. This is complex, I want to compare side-by-side. Oh, I get it, I should take notes on both of these. Hang on, I need to look at both books while taking notes. Okay I’m done with the second book but I still want to take notes on the first.
Slogging a mouse around to click, drag, click, drag, double click, drag, all while repositioning your hands to type, sucks so bad.
The case is even more clear when you consider that the concept of tiling WMs is just an extension of the game-changing paradigm behind terminal multiplexers and IDE splits.
It’s just better. There’s probably a bit of an adjustment when you’re first adapting to it, especially if they’re really used to a mouse-centric, window-draggy workflow, which is likely the only reason that people think they don’t like them.
You don’t really get my point though. I am fully on board with using keybinds and using the mouse as little as possible. I just mean why tile windows at all. I just can’t focus on anything other than one window if I need the other program I have a shortcut that brings it forward. I keep my eyes mostly in the same position this way.
Anyway this is just a matter of workflow that’s what it comes down to. I may just have to accept that it’s not for me.
Yeah, definitely a matter of workflow and personal preference. Nobody wants to convert anyone else, you just ask why people use tiling WM, and people are answering.
why tile windows at all
I can answer that pretty comfortably. There are two main reasons, the first is that it’s very common to have to look at two things at once. If I’m taking notes while reading something complicated, or writing some complex code while referencing the documentation, or tweaking CSS rules while looking at the page I’m working on, it’s just way too disruptive to constantly have to switch windows.
The second main reason (for me) is that a lot of the time, the content of a single window is too small to make use of the space on your monitor. In those cases, if I have something else I’m working on and it’s also small, I’ll tile them. It might be easy to toggle between windows with a hotkey, but it’s strictly easier to not have to toggle, and just move your eyes over. Peripheral vision means that you don’t entirely lose the context of either window. When you’re ready to switch back to the one you just left, you don’t have to touch anything, and you don’t have to wait for the window to render to visually locate where you left off.
I prefer a tiling WM for programming work. On my personal PC I actually use GNOME or KDE or something. But on my old study machine and my current work machines it’s i3 all the way. Being able to quickly tile 4 terminals together makes my work much easier. Often I have many terminals open, each with a bash history specific to what I’m doing there. Workspaces then act as a sort of bundle of applications with the same general purpose. For instance, one workspace for installing and copying stuff to a machine, another for VNC related stuff, etc. If I’d have to alt+tab between 8 terminals I’d never know which one is which. But now I can remember them by location which is way easier mentally. Similarly I sometimes have multiple projects open in an IDE, and I usually remember which workspace is for which project.
It’s even better on 4K monitors, where having 4 1080p terminals open is amazing. I can see everything and I only need to move my eyes. No keypress to switch terminal, everything is right there.
The main reason I use a tiling WM is because I grew tired of having to drag my mouse all over the screen to switch maximised windows, cycling through them with hotkeys or, even worse, spending a fat minute resizing windows with surgical precision in order to have them both visible.
At first I used KDE’s ability to make them transparent, which was ok enough until I tried experimenting with Sway; now I have the habit of splitting the workspace in two, and swiftly resizing the window I want my focus on.
In certain situations floating windows are more convenient, so I just Meta+F, make it a bit transparent, then drag it around.If I really do not need nor want anything else on screen, Alt+Enter forces the window to its size, and if I want to look at the time or smth I have 9 other workspaces to switch to without any delay.
The downside of tiling WMs is that no desktop PC software developer considers their existence, and most applications don’t like being forcefully resized.
Also, popups often take half the screen - I can’t even blame anyone, portable graphical libraries and frameworks do not expect that popups need special treatment for the WM to display them correctly.I’ve configured my entire interface to provide the most amount of input bandwidth for me so that I never have to wait for the computer to do what I want. I3 vim and vimmium solve my problems
I like vimium, but qutebrowser is way faster for me. It’s my go-to for research or reading documentation.
Removed by mod
Even on Windows I’ve always used a “split” desktop. Windows has pretty okay tiling features (for the drag-and-drop folks). I found it pretty efficient being able to look at two browser tab windows or having a PDF file on one side and Word/Docs on the other.
When I was venturing into Linux and found out about i3wm, I pretty much fell in love. It does the same tiling thing on Windows but better. Now I can have four windows without it feeling too cramped and it’s reallly easy to move around with workspaces. I think it’s really great for students and researchers.
I tend to use floating or fullscreen for general browsing but often you have to type something while frequently referring back to something else - for example when programming I will be looking at the documentation. Or maybe debugging something on the command line while looking at your code to see what’s going on. In those circumstances tilling is perfect.
Usually it’s just one program per virtual desktop, and maybe a second (briefly) for one-off terminal commands, etc.
The whole point for me is to avoid wasting time moving a mouse around or manually manipulating anything.
It’s basically the same as a stacking WM, except you can’t lose windows under others. And it automates the window handling with freely sizeable ones like terminals, not hiding them on top of each other, while eg. Steam can get its own Workspace.
Not needing to search through layers when I’m multitasking is the big thing. Having a file manager, ide and documentation all on one monitor makes things much easier.
I’m with you. My eyes can only look one place at a time, I don’t even understand the draw of multi monitor. Alt-tab is your friend.
The only objections to that which I find convincing is the difficulty of managing switching tasks between complex sets of windows and tabs that way. But that just tells me that someone needs to invest in better controls for managing a full-screen switching workflow as a third alternative to tiling vs overlapping.
If you’re only actively using one window at a time, that makes sense, but alt+tabbing through a stack of 8 open applications to go back and forth between something you’re working on and something you’re closely referencing sucks. If your primary workflow for a computer involves that, I honestly don’t understand how someone can live without tiling.
That’s why I said “someone needs to invest in …”. It’s not ideal, and besides alt tab, I do select from taskbar icons (often cascading) which is not ideal either. But I do want most of these windows maximized so tiling is not really the right solution.
I really think there is something that both the tiling people and the overlapping people are missing. There is perhaps something basic in the windowing paradigm that none of us can see past to be able to get to something better.
BTW though, it’s not as bad as you seem to imply with the stack of 8 windows because alt tab goes to the last used window, at least in the Plasma desktop I use. It’s really only the more complicated sequences which get awkward (which for me is pretty common anyway).
Someone has invested, the solution is tiling window managers.
As 217 people have told you in this thread, tiling window managers allow you keep all your windows full screen if you want.
217 people have certainly not told me anything. Maybe you’re confusing me with OP, I think you’re the only one who has replied to my 2 comments.
However I just looked at the rest of the comments to see if I was missing something, but no, no one has addressed what I’m saying. Maybe there is some property of a tiling wm that I don’t get, but to me if a window is maximized, that means it occupies the whole screen and there are no other tiles visible. Whether the other non visible windows are tiled or layered is moot. I think what I want is a way to organize and select windows that has nothing to do with how they are layered in the Z axis, or tiled in X and Y. It’s a logical problem, not a physical space problem.
Again, I’m selecting between a bunch of maximized windows 95% of the time. I don’t deny the use cases for wanting multiple windows to be visible at once, and tiling is a good solution for that, but those use cases are rare for me. I spend a trivial amount of time rearranging and resizing windows. This is the only thing I hear people say tiling solves. This is a non problem for me.
However I’ve never used a timing wm. So I’m all ears if there is something I’m missing.
Oh yikes sorry for the hostility, I definitely did mix you up with OP.
I am always creating content on one half of my screen and browsing through documentation or specifications or chat logs on the other half. So it makes sense, and saves me lots of time, if I setup my computer to automatically place certain windows such that they fill the whole left of the screen and other windows so they fill the whole right of the screen. And this is precisely what tiling window managers are designed to do – especially ones that let you define your own rules about what windows go where.
The ‘tiling’ in tiling window-managers is only half-truth. It distinguishes them from stacking window-mamagers, which would be equal to only looking at one thing at the time.
They all can also do stacking and tabbing, the term means they can do tiling as well. Most users some form of stacking even more than tiling in itself.
Most also can do floating windows, usually on per-app basis. This is achieving what most fully fledged desktop users do: have one fullscreen window per workspace, and have small things like pacucontrol as small floating window.
In no way are you limited to tiling. If you were, tiling window managers wouldn’t be very popular. They’d be like stacking window managers are today.
My most often use half-and-half layout: browser and emacs, emacs and console, emacs and emacs, and so on. If I want to have two consoles and emacs, I instead of tiling, make a stack for the two consoles. This way I always have emacs showing, and switch between the consoles, no matter how many there are. This is kind of like a “sub-workspace”.
The main advantage is ease of configuration, assuming familiarity with config-files. That enables quick, keyboard based navigation, in a very personal and fine-tuned manner. Modern tiling window-managers can also be configured on trackpad or touchscreen gestures, and work intuitively with mouse pointer. So while many users do specifically keyboard centric configuration, the key point in my opinion is that you as a user need, and get to, choose.
