One decade later, GNOME still sucks

It has been more than a decade of GNOME 3’s initial release and GNOME still sucks. Two of my most popular posts have been about GNOME 3 (#3 and #5), and in 2023 people still keep referencing them. GNOME 3 is still discussed today as a major blunder, similar to what happened with New Coke, and the complaints still remain as valid as before.

In 2007 GNOME was the undisputed king of the Linux desktop, with 60% of installations having GNOME (according to Debian statistics), since then GNOME has continued to decline, and other desktop environments like KDE continued to increase. It is safe to say GNOME will never be as popular as it once was, and it has only GNOME to blame.

gnome-session vs. kde-plasma-desktop,kwin,xfce4,mate-desktop,lxde,cinnamon

The mistakes made launching GNOME 3 are numerous, but they can all be summarized in one: the complete and utter disregard of the needs of their users. It seems pretty clear GNOME developers do not understand what the purpose of GNOME even is–or the purpose of all software for that matter: to be useful to the users.

No project is more important than the users of the project.

Linus Torvalds

GNOME developers have always had a disdain for their geek users, and have always desired more normie users, like Apple has. That’s why GNOME shell has always been a macOS copycat. The GNOME project even had the delusional objective of gaining 10% of the global desktop market share by by 2010: 10×10.

In the pursuit of this delusional goal of attracting more normie users, they threw under the bus their existing geek users. Superficially the goal made sense, since there’s many more normie users, but unfortunately for GNOME: normie users don’t care about Linux.

By the time GNOME developers could realize their mistake, it was too late: several forks of GNOME had been already been created, and other GNOME-like environments were created, the users who left were never going to come back: the trust had been broken. Not that it would matter, because GNOME developers never accept their mistakes.

This is why you don’t stab your users in the back.

New nonsense

If I tried to list the number of mistakes, this post would become a book, but I want to explore one at dept: the alt-tab behavior nonsense.

The GNOME project listed a bunch of references to justify their redesign, but it’s clear that they didn’t even read them. One related to workspaces was titled: Rooms: the use of multiple virtual workspaces to reduce space contention in a window-based graphical user interface. If the title doesn’t give the purpose away, here’s a sentence in the abstract:

Rooms is a window manager that overcomes small screen size by exploiting the statistics of window access, dividing the user’s workspace into a suite of virtual workspaces with transitions among them.

The purpose of workspaces is to divide windows, this should be obvious, but if it wasn’t obvious, this paper explained the reasons why in detail.

GNOME developers did the opposite of what their own reference suggested: they made alt-tab cycle through all the windows without any separation.

Not only did the new behavior go against their own references, it was different from macOS’s behavior, Windows’ behavior, every other Linux window manager, GNOME 2, but more importantly… common sense.

If they just made a mistake it would be understandable, they could just fix it, but they refused to accept it was a mistake. I debated this point with GNOME developers, and they all denied it was a problem.

alt-tab: works like OSX and Windows, I’m sorry. you’d probably need to check out other operating systems before making sweeping statements.

Emmanuele Bassi

I explained to Emmanuele that this just wasn’t true (works different in Windows and macOS), did he bother to reply back? Of course not.

This is ultimately the problem with GNOME: the attitude of developers. It is extremely trivial to fix these software issues, I could have easily written a patch to fix their buggy code, but I cannot write a patch to fix brains. Not even Linus Torvalds could get through their thick skulls, his perfectly valid patches were dismissed without any rational thought.

Here’s another lovely example from a gnome-terminal bug:

No.

Christian Persch

Does this look like GNOME developers care about their users? In fact, when I argued that users should be allowed to vote in bugzilla, I got permanently banned with no explanation. When tried to present my case they told me they weren’t interested in hearing it. GNOME has in fact an abusive relationship with their users.

It wasn’t until August of 2020 that one developer merged a patch that just made it work correctly: Cycle only through windows of a workspace when using “Switch windows of an app directly” keyboard shortcut (cycle-group).

So nine years later it now finally works correctly, but why couldn’t they just listen to me back in 2011? Because they are proud idiots who couldn’t accept they were wrong. That’s the sad truth.

Lazy

The excuses they gave as to why they wouldn’t implement something were endless, but the main one was: too much work.

This excuse never made any sense to me. The only way to get more developers is to get more users: obviously nobody is going to program on something they don’t use. And the only way to reliably get more users is to add features, and in order to not lose users, features shouldn’t be removed. It’s not rocket science.

It takes work to add new features without removing or affecting existing ones, sure, but that work is worth it so your existing users don’t get negatively affected.

Linus Torvalds explained how this can be done, and how this is done in Linux to Lennart Poettering (who has a GNOME-like mentality of screwing your users). Of course as is typical Lennart did not want to understand.

Weirdly enough in this very talk Alan Cox used GNOME 3 as an example of what not to do:

If you want to understand the importance of not suddenly changing your users’ experience, I would go and take a look at GNOME 3.0.

Irrespective of whether it’s a good user interface design or not, it’s a demonstration of why you don’t suddenly change everything on people who rely on what you were doing.

Alan Cox

It didn’t matter in how many ways Torvalds and Cox explained how it’s possible to add new features without destroying your users’ expectations, Lennart just didn’t want to understand.

Because the Linux project is willing to do the extra work to keep their users happy and productive, they continuously gain more and more users, and therefore more developers, with zero forks. GNOME on the other hand lost their existing users because it was “too hard” to keep them happy and productive, and as a result now they have less developers.

Gnomesplaining

As big as these issues were, they don’t come close to the biggest blunder: pretending they knew better than their users.

Any time any user presented a problem with GNOME 3, developers just dismissed the complaints explaining that users don’t actually know what they want. In typical cargo cult fashion, GNOME developers believed they could follow Apple success and divinate what their users truly desired.

It’s true that users probably couldn’t tell you that they wanted an iPhone, or a new kind of tomato sauce (as explained by Malcolm Gladwell), but GNOME developers’ mistake was the failure to consider that users are perfectly capable of telling you what they don’t like.

The automatic response to most complaints was “people don’t like new things”. Sure, that’s true for some people, but all of them? It’s a hasty generalization fallacy to think that if some users don’t like a specific feature because they don’t like new things, that all users do the same for all features.

It’s just obtuse and uninformed venting which I won’t dignify with a comment.

Emmanuele Bassi

Bruh. GNOME developers always know better than you.

Actually, ten years later it turns out that new generations also don’t like GNOME, so their fallacy was actually a fallacy. Some people did not like GNOME 3 for valid reasons, namely: it sucks.

Systemic dismissivism

When I complain about GNOME, one common retort is that not all developers are like what I describe. Some are truly hard working and care about the users. Sure, some are like that, but so what?

When discussing about a broken institution, it doesn’t really matter how many personal exceptions you can find, the institution itself is still broken. I’m sure there are some good Mexican politicians, but in general the system is still corrupt. It’s a similar argument with USA cops: not all of them are bad, but enough are in order to cause concern (or so they say).

Studies show that even if only 10% of a population believes something, if they do it fervently enough they can affect the whole population. By that metric even if only 10% of GNOME developers thought it’s OK to disregard the opinions of users, if they believe it strongly enough, or worse: if they are in the positions of power, that belief is still effective.

It doesn’t matter how many good GNOME developers you know, the system is still broken.

And I bet it’s more than 10%. Even if the group is not completely homogeneous, it’s still pretty homogeneous. The class “GNOME dev” has definite type, and I’m not that type, neither is Linus Torvalds, or Zed A. Shaw. That type does not include people who rock the boat or complain about the system, and it includes people who like to be in echo chambers.

Shame

It is a shame because GNOME 2 was amazing. If only GNOME developers were willing to do the necessary work to keep their existing users happy and productive, none of the forks would have been needed.

They could have kept GNOME 2 as a classic mode, and GNOME shell as a new alternative mode. But they decided to throw away their existing real user base in favor of a new hypothetical one, ignoring all their very real and valid complaints.

We all know how it went. GNOME lost the biggest Linux desktop market share ever achieved by any desktop, and they will never gain it back.

And for what? Just so developers could pat each other in the back.

GNOME could still turn this ship around, they could today decide that they are going to create the one desktop environment to rule them all, the one meant for all users, completely configurable and themeable. A truly inclusive endeavor.

I wouldn’t bet on that though.