Rethinking Window Management on GNOME – Space and Meaning

One important missing piece is having information on the maximum desired size of a window. This is the size beyond which the window content stops looking good. Not having this information is one of the reasons that traditional tiling window managers have issues, especially on larger screens.

I have been upset over losing this functionality from the classic Macintosh days for decades now. This was built into the Macintosh OS going back to at least System 7. Clicking the "expand" button in the title bar would expand (or shrink) a window to its optimum size. For Finder windows, that meant the smallest size that could display all the files without scrolling (if possible).

Developers had to implement logic to make it work. For the most part, they did, and it worked very well up through OS 9.

Then came OS X. The green button, at first, worked exactly the same way it did in OS 9. The problem was that even Apple didn't give a damn to write any logic for it into their apps. It might as well have been a "randomize" button. Users got frustrated. Windows converts wondered why there was no "maximize" button and blamed the very concept of expanding rather than Apple's now-piss-poor implementation. Longtime Mac users wondered why we effectively lost a very useful feature.

Over the years, Apple continued to neglect the function of the green button, and third-party devs largely followed suit. Eventually Apple changed the default behavior of the green button to go into full-screen mode, hiding the original (still mostly broken) "zoom" functionality behind an Option-click. At this point, the the difference is effectively "full screen" vs "windowed full screen". RIP the classic Mac OS expand function.

A good start would be to implement quarter tiling by dragging window to screen corner, like half tiling is done by dragging to screen edge.

I have a 3840x2160 monitor specifically so that I can have four windows open at the best size for their content (email, document, web browse, and terminal) and can avoid the use of workspaces and see everything at once. Having to manually resize and place windows is a pain.

Moving maximized windows to their own workspace seems like a really cool idea. Workspace management is one of the things I struggle with so I usually just end up with way too many tiled windows on a single screen when they could be moved around more efficiently.

This is looking really great, and I want it now. Wish I didn't know about it until just before I could use it :P

Looks really interesting, hopefully this can be a step forward for window management as a whole

I've been using paperwm on gnome for a couple years now, it's my preferred paradigm for tiling. This looks like it has a lot of the same influences, so I'm interested in seeing where it goes

I love the direction this is going, I've been using i3/sway for years and gnome apps recently became awesome in tiling mode because of their responsiveness. If this is implemented this could definitely get me back on gnome 👍

Looks awesome, I can't wait.

When I was using Gnome on a laptop, I really enjoyed the PaperWM tiling manager extension. It's not exactly something that can be used with a mouse, but it's a really pleasant touchpad/touch first multitasking interface, where instead of having traditional workspaces that are constrained to the size of your monitor, you basically get infinite horizontally scrollable workspaces that are a joy to navigate with a touchpad.

Are they going to rethink putting thumbnails in the file selection dialog or many of their other insane decisions?

Gnome seems like they want to take the Apple approach to UI design without the attention to detail that Apple's UX has.

This looks cool. I might consider trying GNOME again if it gets implemented.

As long as the windows are being maximised into a workspace you can jump to with SUPER+[0-9], this seems interesting, if overly reliant on the mouse.

You can only focus your eyes on one window at a time. Alt-tab swapping between windows covers the vast majority of my use cases. Arranging windows next to each other to view them at the same time is a waste of time, other than a few rare use cases for me. What would be an improvement would be better ways to manage sets of windows to alt-tab through.

If this becomes a Wayland protocol, then I'd love to see other desktops adopt it as well.

I could see a few classic TWMs use those hints, or at least expose them for users to script functionality.

Went tentative, left excited. Honestly it sums up my hopes for float/tiling window future.

Looks well thought out, I'm interested! I think the biggest problem will be leaving the user enough control to modify the layouts after they've launched and not downsize apps after the user has upsized them and the like. As long as that is overcome this would be a fantastic way to use the desktop.

So I've used the Pop Shell extension. It's really neat when you have a bunch of little windows like terminals and file browsers open. 95% of the time it's actively annoying though. I appreciate that it's on a toggle so I can use it when I want it. The proposed mosaic mode doesn't seem terribly different, and has the same problem where it just randomly moves things around breaking my association of "where I put that". Most of the time I really need the spatial aspect, and am willing to manage a few windows by hand to get it.

Also: Joining half screen windows into a single unit?! Please don't do it! D: Augh! Apple did that on OS X about the time I left and I absolutely hated it. It was so actively bad. :(

That would be really awesome! Especially if it could learn your preferences automatically by how you change the default arrangement.