There is a general rule that the more time people spend futzing with some particular aspect of their shit, the more they fetishise it. It's a kind of cognitive sunk-costs fallacy. I've used i3, and it's great. I wish something like it were available on Windows or macOS. But it's just a trivial convenience; hardly something to develop an emotional attachment to. Let's face it, humans are odd creatures and will attach to almost anything.

IKEA effect! An example of which is those prepared doughs where you have to add eggs... the egg could (and has been at some point) be included but marketing discovered that requiring the add-your-own-egg step made people love the end product more because they somehow made it themselves.

https://en.m.wikipedia.org/wiki/IKEA_effect

I've heard of that before, but it always made me wonder - are we just supposed to accept unquestioningly that powdered eggs are indistinguishable from fresh? Sure, there's no question you can make something edible with dehydrated eggs, but that isn't proof there's absolutely no difference.

In general, the existence of a bias is insufficient proof that you should reverse your conclusion.

Nice - I hadn't come across that. Spot on.

i3 allows the user to maximize screen real estate, and also to minimize time spent adjusting window borders and navigating with a mouse. It’s a huge time saver for anyone who learns a few simple controls. It’s like tmux on steroids.

Sure it’s a convenience, but so is basically everything we do to alter a computer after unboxing it.

I've never spent any appreciable time 'navigating with a mouse' and 'adjusting window borders' on other OSs. Sounds like those thundering imprecations that we all must use vim or emacs because IDEs require mouse use (which they don't).

I agreed it's nice - but essential enough to merit the time sink of running linux? Not for me.

I have to agree. I don’t use a tiling window manager and I don’t consider myself an average user. The comment above was pretty insulting.

Maybe I’m just a Luddite that doesn’t know any better, but I’ve been quite happy without tiling windows for many years. Then again, I spend most of my time in a terminal in a tmux session. Maybe that counts.

A common thread here is - your use case is not my use case. Many people are assuming that their workflow is the ‘one true way’. That not a healthy way to have a discussion.

It might've been, but I agree with the core point: make window tiling an option. For the amount of time that it would take to implement, and the non-zero number of people that would benefit from it, it seems a glaring omission. Every OS should let you line up two windows side-by-side out of the box: this should be a minimum requirement of any window manager that lets you have windows at all.

Full screen mode is macOS native tiling mode, you can put two windows side by side and use as much real estate as possible (no menu bar nor title bars)

Eh, that’s pretty darned far from an i3 or Xmonad experience. It’s roughly as wrong as saying that the Gnome frameworks are just as good as Cocoa :-)

I have used i3, awesomewm and Xmonad extensively for years so rest assured I wasn't implying anything like that ;)

I was more like pointing out that macOS does allow for the following in some way out of the box:

> Every OS should let you line up two windows side-by-side out of the box

Sorry, that really was only an example. I also didn’t say, or mean to imply, full-screen only. Every os should also let you line up windows top-to-bottom out of the box. Etc. Macos' approach of allowing one very specific combination of window arrangement, then trying to make a feature of it by calling it 'fullscreen' or 'side-by-side' (or whatever they call it) just doesn’t really cut it.

The problem is that there is no good solution for tiling window management on OS X, even with third-party software. The parent may not have stated it that way, but you’re being over sensitive if you’re actually offended by their comment.

Personally I feel like I’m using a toy when I have to use a non-tiling window manager. It may not be for everyone, but the display server or desktop environment should at least provide an interface to allow someone else to implement it.