Is Linux distancing itself from its enthusiast/power user roots?

Some developments in Linux have left me wondering if Linux is turning away from its enthusiast/power user origin. For me, it is not the increasing standardisation aspect which for cross platform development is helpful, but the choices being made, for example Wayland focusing on preventing developer placement of windows and the rise of immutable distros which deliberately discourage tinkering with system packages.

If Linux continues down this path and locks everything down, other than being open source and more secure than the alternatives, why would someone want to use it instead of Windows or MacOS?

5 Likes

I used it because it worked, and because I didn’t like being forced to pay to continue using my computer after replacing the motherboard/cpu.

I also use it because, unlike Windows, it never managed to destroy my data (I lost about 3/4 of the images I took with my first digital camera back in the early 2010’s along with a good many photos from my SLR at that time).

1 Like

I don’t think “Linux” is one single entity that has any ubiquitous directions.

Enthusiast Linux continues to exist. There are many distros and softwares that appeal to power users. They can co-exist. I think if anything, the scope of desktop Linux is growing more broad to support all types of users, not shifting to one particular type.

Wayland I think is not so much about enthusiast vs novice (most novice users would not even know the difference) but more about the features and capabilities that Wayland provides over Xorg.

I am a developer and enthusiast and I am quite excited about the idea of immutable distros. I’m currently running KDE Linux and I am enjoying it very much.

One of my favorite things about Linux is that there are lots of choices for each person to find what suits them :smile:

14 Likes

The idea of being able to use the computer as a device to be turned on and used is indeed a return to the days when we used pocket calculators, and those early Nokia phones which recieved occasional flash updates, but were immutable and again - reliable.

Phil (Manjaro) has also stated his opinion about the Arch model now becoming an outdated concept, and he’s also getting involved in the ‘new generation’ of operating systems.

Many of the ‘features’ of X11 were actually bugs… the biggest loss for me is the demise of ‘mouse gestures’ or easily defined mouse actions.

1 Like

I second this sentiment.

These “immutable” distros discourage tinkering? Maybe in the very narrow sense you are thinking here, but as an enthusiast and tinkerer, I don’t feel anywhere near as restricted as I thought I’d be when I started using some to see how they worked, and daily drive one on my laptop. Plus “enthusiasts” will always have Arch, Gentoo, on top of the other “normal” distros, and will continue to as long as there is interest. Which there will be. You don’t see the gaming focused things like Bazzite as NOT being for enthusiasts and tinkerers?

I see these atomic distros as being a supplement, not a replacement.

2 Likes

From what I can tell, with immutable distros, you are in luck if they support all your hardware well and have all the system utilities you need, but if you need to compile a driver from source, patch some system software or you want a specific system level utility, for example to remap inputs, then it becomes more challenging.

As for X11, it had many weaknesses, but they could have been overcome without the tight restrictions that Wayland imposed.

For me, the ability to set things up the way I want and install whatever I wish is the biggest draw of Linux. Reducing this ability diminishes a key differentiator that sets Linux apart from Windows and macOS.

2 Likes

And yet the X11 experts did not do as you suggest, because it was not possible inside the design limits of X11.

5 Likes

I didn’t mean by fixing X11. I meant by designing Wayland without making preventing apps and games from positioning their windows a primary goal.

1 Like

Agreed, it is sad that it’s taken so long to allow this feature to work again.

That is a common misunderstanding or misconception.

Wayland’s protocol stack was designed around the concept of aggregation rather than subtraction.

Meaning that at its core it only deals with the essential operations needed by all compositors and their clients, e.g. communicating window content.

For each type of usage scenario, e.g. mobile, embedded, desktop, etc., there can then be additions that add features needed but also possible.

One of the things that is not universally possible is a global 2D coordinate system.

For example in a 3D scene (VR, XR, AR) there can be many 2D surfaces and each could have a 2D coordinate system, however it would be very hard, if not impossible, to define one that is valid across all of them.

This absence of global coordinates is often misunderstood as preventing window placement because global coordinates can be used for that.
However, there are many other options for placement, some of which being available on compositors that support multiple windows, e.g. desktop.

For example one of the most common placement tasks is to position a popup, e.g. context menu near the mouse clock position, normal menu close to the menu entry, choice dropdown below the choice input field.

Desktop compositors provide the features for that through an interface called the xdg_shell protocol. Essentially the application will ask the compositor “I have this popup, this is my reference position/rectangle and I want the direction of popup to be down”.

There are lots of other possible requests, e.g. “on the left hand side of my main window”, etc.

One of the primary ideas behind this is to add context in the communication between app and window manager.

A tiling window manager, for example, can then decide to show certain sub windows in the same tile as the main window while showing other types in their own tiles.
If it just had a pair off coordinates and maybe some window flags, it would need to guess and possibly get it wrong.

Another important aspect is that any compositor can offer special features to its clients (e.g. a global coordinate system) and they can react on that instead of always seeing a capability but never being sure it is actually supported (a tiling window manager on X11 will often just ignore positioning requests)

11 Likes

I’ll be blunt and that is to say I smell pure bs in the first post. No real need for discussion.

3 Likes

All of these are possible on KDE Linux, which is an “immutable” OS. The way you do them differs from the way you do it on a “mutable and made of packages” distros, but you can still do them. Especially if you can compile software, you can do almost anything you want.

Input remapping is something that really should be a fully-integrated 1st-party feature, though, and we’re working towards that. Waaaaaay better UX than fiddling with config files for system deamons and stuff.

5 Likes

Would this be implemented at the level of KDE Linux, or Plasma? It would be great for Plasma users on all distros tbh.

Plasma, of course!

4 Likes

Fantastic to hear! :smiley:

1 Like

Input remapping in Plasma would be fantastic especially if it can cover mouse, trackball, gamepad and joystick as well as keyboard so that I don’t need Input Remapper or AntiMicroX. Would that then mean a Flatpak or AppImage game that uses something like SDL would receive the remapped keys, buttons and axes?

I’m also heartened to hear that it is possible to install custom drivers in KDE Linux. in particular, I’ve only been able to get my fingerprint sensor to work with a specific version of a driver (as per this Debian guide). If I were to switch to an immutable distro like KDE Linux, then I would need to figure out how to replicate that setup.

1 Like

Thanks for your helpful explanation. I have been working on reviving a game that was first started in 2006 (Oolite). The game which uses SDL 1.2 has a splash screen and a main window. On Windows (MSYS2 UCRT64), macOS (Intel-based) and X11, the splash screen and main window are centred. On XWayland, the splash screen is centred, while the main window is offset - it does not seem to respect SDL_putenv (“SDL_VIDEO_WINDOW_POS=center”);.

What I ended up doing for Wayland was a hack that launches a separate splash screen executable from the initial shell script before launching the game itself (with game’s splash screen disabled).

I am not afraid about immutable distros (I belief some distros will always stay mutable for good reasons), but more about “modernization” of DEs. Especially the development of file managers of GNOME and KDE scares me as nothing else right now (the huge square that contains icon and label). Such “modernizations” have not even to do with “being enthusiast/power user”, but with a design language that tries to do something better while ignoring the huge usability downgrades. I know, devs try to find a good middle ground, but I have enough experience in UX design to know it will be worse than it was for 30 years. Some designs made decades ago have a really good reason to be that way.

But I actually like Wayland and the possibilities it gives. It just needs to be “fixed” for multi-window applications and better implementations for portals/global shortcuts, which I guess are on the way.

So no, it is not distancing from enthusiast or power users root. It is just changing. Most times in a good direction, sometimes in a worse one.

This is being worked on: Gestures in Graz, and beyond - KDE Blogs

3 Likes

I hope so, and I’m also excited to see the prospect of ‘long press’ keyboard shortcuts being a possiblity.