Gnome Mess Is Not an Accident | Hacker News Reader

Gnome Mess Is Not an Accident

medium.com

78 pointsrc0053 comments

The fact that on KDE you're able to achieve a consistent look across Qt and GTK2/3 apps ootb, but not on GNOME, should tell you a lot.

And then GTK4 comes along, and you basically can't make it look consistent. Not even their titlebars, since they're now CSD.

All of this is sad, I genuinely think GNOME has something good in there. I try every new release. And, there is a pretty good work flow there. But, I always hit some sort of a wall, impeding my enjoyment. For example:

- The inability to drag and drop files from the file roller to nautilus (as said on the article)

- Alt-tabbing groups apps and requires you to use your mouse or the arrow keys (of all things) to specify which window to change to

- Full screen video still leads to strange frame pacing and/or tearing (due to unredirect fullscreen windows)

- Still no good way to interact with background apps that use a tray icon by default (I know they now allow background apps to be interacted with, via the power menu, but it's simply a worse experience. Also, app developers need to account for this and develop this feature specifically for GNOME, instead of just doing it the standard way everyone else does it)

- Touchpad multi-touch gestures aren't supported in X11, even though other projects (Touchégg) demonstrated that it's possible

- Having to search for a newly installed app 2 times before it appearing in your apps list (Xfce, KDE, Cinnamon work the first time. Is it that difficult to refresh the cached apps list?)

You can always install extensions to make the experience better, but they almost always need to be updated to account for new releases that have breaking changes.

(Edit: formatting)

> Alt-tabbing groups apps and requires you to use your mouse or the arrow keys (of all things) to specify which window to change to

For what it’s worth: you can use alt+` (the key above tab on my keyboard) to cycle between windows of the same application. You can immediately use it without releasing the alt key, or use it directly if the application is already in focus.

I did not know about that! Thanks!

Gnome after version 1, definitely after version 2 is a symptom of the "software is never done" fallacy. After building a solid, performant and working desktop environment, instead of leaving it alone and start making other applications, they decided to just keep doing stuff.. and how do you continue improving something that's already finished ? You don't, because it's f*cking done! So what's else to do than just randomly screw with things. This is also seen in Firefox, Thunderbird, and the Windows operating systems.

I switched to Mate when they made the first major screwup, but eventually Mate got bloated too for no good reason, back to XFCE, but they also can't keep their hands off of it.

So here I am in i3, which is not really what I'd prefer, but at least they don't seem to have the same itch to constantly change stuff just to change it.

Here's how software _SHOULD_ work: Develop it until it's _DONE_ and _STOP_. Now, you ONLY submit code that makes it run faster and use less memory. Otherwise, just stop, move on and make something else.

Have you tried out Enlightenment? I haven't noticed much changing in a negative way.

https://www.enlightenment.org/

That was actually one of my very first encounters with Linux, it was a debian based distro called "storm linux" that had it as one of the default options.

I've gotten back to it over the years, but it seems.. odd somehow, slightly on the janky side and a bit slow too.

The "Cascade of Attention-Deficit Teenagers" model, or "CADT" for short.

https://archive.is/t5m32

  >Gnome after version 1, definitely after version 2 is a symptom of the 
  >"software is never done" fallacy. After building a solid, performant and 
  >working desktop environment, instead of leaving it alone and start making 
  >other applications, they decided to just keep doing stuff.. and how do you 
  >continue improving something that's already finished ? You don't, because it's 
  >f\*cking done!

It was bad even then, because it imitated some elements of MS windows, some elements of OS/2, some elements of mac os, it had messy APIs, but we didn't care because most of us have used Windows as first GUI DE and it was performant and mostly cohesive, yeah, so all in all was OK.

  >So what's else to do than just randomly screw with things. This is also seen in Firefox, Thunderbird, and the Windows operating systems.

And with GNOME we witness the same fate of all three, being a commercial software passing for a software you can use for free. It will infuriate some people forever.

> > I understand that this is not your use case, but it doesn't apply to everybody, right? Please notice that this is not a feature request, but a regression report. Something in 1.17.1 broke a useful feature, but no changelog explaining why has been provided (AFAIK), and when someone reports an issue you close it because you don't agree with the use case.

> > This issue can be replicated using pkexec. And, yes, I know about admin://, but it doesn't address many use cases that have been working for years. For example, when in an admin instance of nautilus if the user:

> > opens a terminal from inside it, this terminal won't be loaded in admin mode wants to execute a pkexec operation it will ask for super user password again > > right clicks on a left panel folder and select to open a new tab/window it will not be loaded in admin mode believe me, there are many more examples

> > I appreciate your work and I understand the time constraints. I'm also an open source developer and have been contributing to many projects -- actually, that's the reason why I'm reporting this since I don't use this project. Not sure if you guys see this, but it seems this behavior of yours is one of the reasons why so many people get sad about GNOME and its side projects. As Linus Torvalds once said, the whole we know best thing can be toxic.

> > No hard feelings. I hope you understand :)

> This comment was marked as abuse

Hilarious.

This article sums up what I have been complaining about for nearly a decade.

Yes, it's the GNOME mentality / disease, that has been metastasizing to other areas of the Linux ecosystem such as Freedesktop.org.

The same arrogance can be seen in systemd developers as well (another freedesktop.org / Red Hat project).

I'm often the first to attack Gnome for doing something counter-user but on TFA's opening charge…

> the most limited and bloated desktop environment

… I struggle because for all its developer-instigated limitations, Gnome performs really well. Usually that translates to Mutter is really good at getting out of the way of the rendering pipeline, but to users it means the same thing.

It also gets a ton of community support to resurrect a lot of the butchered features through extensions. The JS API is pretty decent and as it improves —and we stop expecting Gnome to look after anything but itself— that Gnome could be the most flexible desktop.

They do a lot of things wrong at Gnome —stemming from a plethora of individuals having too much unchecked design power— but the display server stuff isn't bad at all.

Most of the issues pointed out in part 1 of this series of articles were resolved by part 2 and many of the ones in part 2 were resolved by part 3.

But the fundamental problem with the rant is that it doesn’t attempt to understand what Gnome is.

Gnome is not intended to be a final end user DE. There’s a reason the Gnome devs have absolutely refused to release an distro (besides one intended only for Gnome testing and development). It’s because Gnome is intended to provide a baseline DE that can be used by several distros allowing them to build on top of it, select sensible defaults for their distro, be able to distinguish their distro DE from Gnome and other Gnome based distros, while still providing great familiarity to users across Linux distros.

So, for example, Canonical could create Unity, and System77 PopOS! which are both based off Gnome, provide significant differentiation, but at the same time a user going from PopOS! to Unity to vanilla Gnome and any other path would barely miss a beat (other than the added features missing altogether).

Further, Gnome’s focus for the past several years, which was much overdue (and something most other DEs haven’t even attempted) has been on creating a bunch of cohesive apps based on a single modern framework (GTK 4+ in the case of Gnome). Much effort has been spent rewriting apps in GTK4 and/or creating new app replacements even if they might be under featured for now, because if a distro feels the features in the previous app was more important it can continue using it.

This isn’t to say there aren’t a bunch of problems with Gnome, much like there are a bunch of problems with both closed and open projects of all sorts. However, any criticism will be more useful if you understand what Gnome actually is before criticizing it for being something it isn’t.

I don't understand how it isn't intended to be an end user DE, they've been adamant that they hate how some distros change the base GNOME DE. Which I do agree with them to some extent.

Nevertheless, even if that was the case, the default UX should be better.

Changing base and creating a fork are vastly different things.

As someone that used to be an active Gtkmm/Gnomemm user, it certainly wasn't their goal during the first decade of this century, and Canonical only came up with Unity due to disagreements of what the user experience should be all about.

If it was so good at then, then why is System76 investing so many resources in replacing GNOME with their own Cosmic system.

Where is this nuance articulated on their site? All I see is mention that it can be used as a ‘fully featured stack’ or ‘pick and choose’?

It's funny because none of the article resonates with me, except Gnome Software. That somehow still manages to be the worst application that requires you to kill it's background process to make it restart and only then will it work. Until you leave it for a bit and come back.

I tracked some issues raised when Zoom was trying to get screenshare working with Gnome and Wayland. It was pretty clear the Gnome devs had no interest in UX feedback for video-conferencing apps. And people wonder why those apps suck on Linux.

This is what I don't understand about GNOME and their sponsors. Who are they even trying to serve? Because I could understand a project having goals that are misaligned with common users because their sponsors want them to do something else, but with GNOME they just seem to striving to produce a DE that nobody wants to use. And seemingly denying that any use case for a computer exists except staring at the desktop and doing nothing. Why are Red Hat et al. pouring money into a project that doesn't think videoconferencing is an important use case??

>Simply opening a folder with 40 images in Files (nautilus) takes 4.6 seconds to load all thumbnails, while in thunar (Xfce 4.18) the same thumbnails load in 2.6 seconds. This is a local file browsing experience on a real machine (not VM):

Neither of these speeds are acceptable for 2023, 40 images should feel instant. On MacOs folders with 100s of images don’t even visibly load in you open it and the thumbs are there.

This is what happens when you have UI implementors who reject the last 20 years of UI progress.

I would say the opposite. If GNOME was a clone of OS X Panther it would be fine. Instead they've introduced a bunch of UI "innovations" that most people don't like.

I'm talking more about methodology. There are a lot of "forgot to set the dirty bit" type bugs in GNOME-based applications. UI moved on to declarative rendering ages ago, and GNOME never got the memo. When you are writing Flutter or React it's virtually impossible to make those mistakes.

It is not Gnome fault. It is more related to design of GTK API

Gnome can choose the tech. They can also build a reactive layer on top of gtk to eliminate a class of issues. Or adopt Grex https://refi64.dev/posts/introducing-grex.html Or...

If Gnome keeps rewriting apps and making mistakes, that's very much Gnome's fault.

GTK is a core gnome project. It's entirely gnome's fault.

It keeps getting worse with each version, and they drag a big part of the open source community with them, because so many apps depend on GTK. Especially important ones like Firefox and Chromium. It absolutely sucks when Gnome controls such a large chunk of the open-source ecosystem.

In fact, it was Gnome that dragged Debian into the systemd debacle. Lennart Poettering managed to convince the Gnome developers to introduce a hard dependency [1] on systemd.

1. https://news.ycombinator.com/item?id=31993161

100% agree - there is so much similarity between those projects. They will be defended to the hilt but you only start to see the tangled vines when you try to build something without them. If you try to pull them out all other software goes too in a big clump. Then their developers ‘own Linux’.

But then again, it is Gnome's fault to pick the wrong tool for the job.

Who maintains that again? :)

This opinion is reflected by many RHEL users, anecdotally.

Such as...?

Middle click to close a window

Cargo-cult enshittification, where they blindly follow the corporate UX design trends, which are usually dictated by the marketing department, and done to maximize revenue instead of making a usable and productive interface.

Such as following trendy fashionable fads such as low contrast text, excessive whitespace and overuse of animations. Also to make the entire interface a part of the brand identity. In my opinion, these concepts have no place in open source software.

Also: if you don't like the new interface they will say you have problems accepting change, despite it objectively being inferior to the old versions.

KDE has done a lot better than Gnome in this area, by the way.

GNOME has an interesting progress history.

- Initially, Miguel De Icaza started GNOME in 1997, he also created the Gnumeric spreadsheet for GNOME, and when .NET was released, he together with Nat Friedman (CEO and Co-Founder of Github), started Helix Code which later become Ximian/Xamarin and Mono. Miguel De Icaza abandoned GNOME even before 2013, and finally left Linux in 2013.

- Several default apps in GNOME were written in C# Mono, namely F-Spot and Tomboy Notes.

- Nat Friedman also created the Beagle Desktop, which is a search-and-indexing tool for GNOME back in 2006.

- Ximian Desktop developed the Evolution Groupware Suite, to allow GNOME to be used in office environments. Evolution had become GNOME's standard email client.

- Eric Sink, which led the SpyGlass browser team that licensed the Mosaic browser, which had become the codebase for Internet Explorer, developed AbiWord, which is GNOME's standard word-processor.

- In 1999, Andy Hertzfeld from Apple together with other Apple engineers, started Eazel, which created the Nautilus File Manager, which is had become the standard file manager for GNOME.

- Novell Desktop introduces the Compiz compositor in 1999, also SLED10 has a new GNOME Main Menu, which become a standard menu for SuSE.

- Novell also spearheaded the Tango Desktop project, which provides theming consistency for GNOME and it's apps. Later, Tango Desktop was adopted by other projects as well.

- RedHat developed the NetworkManager layer, also, Robert Love from Xamarin (Now Google) contributed heavily to the project. NetworkManager, makes it easier to configure and switch WIFI networks in Gnome.

Several underlying tech that build GNOME comes from:

- GTK+ toolkit, that is the basis GNOME comes from GIMP Image Editor.

- GNOME has several WM throughout the years, starting with Sawfish, then MetaCity and then Mutter.

- Ubuntu having invested heavily on Compiz, tries to keep Compiz and GNOME 3, AFAIR, they have initially opposed the GNOME 3 direction of using Mutter and several GNOME 3 design principles, hence they have developed the Unity Desktop. I am not sure what mainline contribution that comes out from Unity, as I didn't follow after Unity.

GNOME was actually a fun project, several changes had happened to GNOME, but IMHO, it's a masterpiece out of orchestrated chaos.

never been happier in my life since i stopped caring about gnome, i wrote a couple of gnome extensions to make it usable (for me) and having to deal with all the breakages was really getting on my nerves.. switched to use i3 with a bunch of shell/python scripts and can go on with my life

I agree Gnome could use more polish and customizability. But it is my favorite desktop by far. His opiniated design is both a blessing and a curse, but like Apple, I accept the tradeoff: the UX is too good compared to the other players.

Product management is hard, and even harder in an open source realm.

Yes, there are some things to fix and make it better, it is impossible? I don´t think so. I've found GNOME to be simple and cleaner than other DEs, yes, you can go full tiled or even no DE at all, its a personal decision and its great that you have the change to do it.

But hey, if you paid for that product and is not working, you should get your money back.

Maybe getting people to agree with you in a more constructive way can have a better response than ranting around (ranting is an old trend, we should stop doing it).

I used to be a GNOME fanboy during the early GNOME vs KDE stuff, dabbled in Gtkmm, even wrote a couple of articles about it as kind of advocacy.

Even if I not a big C fan, the moment they decided JavaScript was a great way to program basic functionally, which had to get as 3rd party extensions, I was out.

First Unity, tried GNOME again after Ubuntu dropped Unity only to realise things have hardly changed, afterwards happily in XFCE.

Although, the way Gtk application development is going, with Glade's deprecation, with a Web based tool as replacement, still WIP, isn't much better either.

Nah, XFCE is absolute trash. The failure after decades to fix the window grab handles just shows that it really doesn't care about it's users.

In comparisasion with GNOME, it is wonderful.

I never had an issue with windows, by the way.

quite happy with Xfce, but yes it has some problems. I am using it in old school windows 98 style and would be interested in similar lightweight desktop environments. Any suggestions

That's just it, for some reason XFCE ties the window grab areas with the theme so the solution is "don't use the default theme".

Mad that the best way of using XFCE is to not use the defaults (last time I used it I also had to change the whisker menu because the default one was Win95 era), whilst people in the same breath complain that you have to change the Gnome defaults.

I'm fine with Sway, the nwg-piotr project adds some interesting panels if you want it.

Sorry for the hyperbolic initial response but the hypocrisy with the alternatives which have the same sorts of problems is annoying.

The default theme in Xubuntu works just fine, maybe try another distribution.

XFCE disagrees with you.

https://forum.xfce.org/viewtopic.php?id=10432

Just Google and there are hundreds of hits with people struggling to resize windows due to the tiny boarders.

My laptop disagrees with you. So is the curse of Desktop Linux.

I like when people whine on the internet, but do not do anything to improve. Go ahead and help them. It's an open source project in which everyone can participate.

As the article explains, Gnome developers are often hostile to even acknowledge a problem exists, even when offered patches, and instead just search for ways to deny that the use case could ever exist

Have you read the entire article? A large part of it is about bugs, features that are not acknowledged and contributions that are rejected on the spot.

Why would you spend time making someone else’s collapsing project better for free while they get paid while you could just work on your own vision

He literally submitted bug reports and patches that were rejected

Come on, we all know the GNOME project does not work that way.

Reply on news.ycombinator.com