Is Cinnamon Dying?

If you know about desktop Linux, chances are you know of Linux Mint. Linux Mint is seen as an easy to use Linux distro for Linux users switching from Windows. And for a while, it was. It used GNOME 2 on a solid Ubuntu base, but it added things such as non-free codecs and Nvidia proprietary GPU drivers and making the experience easier than on most Linux distros. It also put a focus on graphical apps rather than having users enter commands in the terminal, which isn’t very user-friendly for new users.

Linux Mint also sometimes did things different from upstream Ubuntu that benefited users.

In 2019, Canonical decided to switch the Chromium web browser in future Ubuntu releases from an APT package to a Snap, most notably with Ubuntu 20.04. In 2020, around the time of Ubuntu 20.04’s release, Linux Mint project lead Clement Lefebvre announced that he was against Snaps due to a conflict of interest as while the front-end was open-source under the GPLv3, the back-end was (and still is) proprietary to Canonical. This is in stark contrast to Flatpaks, which were pushed by other distros like Fedora, and not only were fully open-source, meaning anybody could set up a third-party Flatpak remote, but it also did not suffer from some of the other problems of Snaps, such as longer start times. For those reasons and others, Lefebvre announced that Linux Mint would block snaps from being installed by APT (the package manager of both Debian & Ubuntu-based distros) as was done on Ubuntu but users could remove this restriction manually. (Source: https://blog.linuxmint.com/?p=3906)

Since then, in Ubuntu 22.04, Ubuntu did the same thing as with Chromium to Firefox, and made it available only (officially) as a snap package. This meant that users of Ubuntu who did not want to use snaps and distros based on Ubuntu who wanted to remove snaps now had to find alternative sources for Firefox or had to maintain their own versions of Chromium and Firefox.

This has been seen as a good move by Linux Mint, as many users preferred Flatpaks over Snaps for reasons dicussed earlier. In fact, as of January 2023, only the following distros use Snaps out of the box:

  • Ubuntu (and all official derivatives, such as Kubuntu or Xubuntu, although Ubuntu MATE also packages Flatpaks out of the box as well)
    • Zorin OS (which is based on Ubuntu)
    • KDE Neon (which is based on Ubuntu)
  • Manjaro
  • Solus

So yeah, not that many apart from Ubuntu and its flavors and several derivatives. Not to mention Manjaro’s constant issues, and Solus seems to be in decline.

A (Condensed) History of Linux Mint

In 2011, when GNOME released GNOME 3 with its massive design, UI, and UX changes, the Linux Mint team tried to compensate for these changes by adding GNOME extensions to try and focus the desktop environment back towards a desktop metaphor. However, when this became unworkable and unfeasible, they did two things:

First, they forked many GNOME 3 components and released the Cinnamon desktop environment in 2012, a desktop environment based GNOME 3 technologies but with a desktop-oriented focus that they could control.

Second, they added resources for a team to fork GNOME 2’s final version, GNOME 2.32, and its applications, to keep its traditional desktop metaphor but allow it to be updated to feature newer technologies, such as GTK 3, HiDPI screens, and even preliminary support for Wayland, the next big development for the Linux desktop. (Keep this last item in mind for later.)

But Linux Mint wasn’t just a distro that used GNOME 2 / MATE as well as Cinnamon. In fact, the first version of Linux Mint was actually based on Kubuntu 6.06, with KDE 3 as the desktop environment. In addtion, Xfce was also an option starting from Linux Mint 3 in 2007, based on Xubuntu 7.04.

While there were the occasional support for other window managers and desktop environments, such as Fluxbox & LXDE, by Linux Mint 13 in Ubuntu 12.04, there were 4 offically supported desktop environments:

  • Cinnamon (the flagship desktop environment of Linux Mint)
  • MATE
  • Xfce
  • KDE Plasma (officially supported from 1 to 18.3)

And at that time, Linux Mint seemed to be doing quite well. It became very popular in the Linux and FOSS world for its ease of use and familiarity to Windows users. But things would not stay this way forever. An issue that would be even more pressing that any other one prior slowly reared its gorgeous head. And it involves the heart of their flagship desktop environment Cinnamon…

The Big Mistake

Wayland is awesome. It allows screens to support fractional scaling, vastly improves multi-monitor setups’ refresh rates when they’re different, and makes the compositor’s codebase much leaner than with X11.

Work has been done to support the new display protocol by GNOME and KDE Plasma. MATE, Xfce, and even Unity (yes, Unity is alive and well) have preliminary plans for Wayland support. Heck, Xfce will likely have it in time for 4.20.

Cinnamon?

🦗🦗🦗🦗

This is a problem, because X11 is not seeing any new development as its developers move on to Wayland, which is clearly the future for desktop Linux. Not even attempting to try to start Wayland support is a pretty large red flag for not just the Cinnamon desktop environment, but also Linux Mint as a whole.

This development likely takes time, and not even starting work or plans for Wayland is not looking like Linux Mint planned this out very well.

With GTK 5 rumored to remove X11 support, Linux Mint has a choice to make. And so far, it seems to be the wrong one, and one that will drive them into the abandonware bin more and more over time, along with LXDE, Crunchbang Linux, and the Jingpad.

Here’s a quote from July 2022 about Wayland support in Cinnamon (and this is the latest feature request) when asked if there was an issue for tracking Wayland support progress for Cinnamon:

Not really, no.

We still don’t have any specific plans for implementing Wayland support. Theoretically, we are a lot closer to achieving it now that muffin has been rebased on a newer version of mutter, but I don’t know how much work would still need to be done to make it work, and we’re not in a hurry to do so. It’s not that we’re just dragging our heels either – there are some major downsides still to using wayland. One in particular that we discussed just a couple of weeks ago is that wayland doesn’t run in a separate process the way x11 does, so if cinnamon crashes (which can be caused by any number of different things, including a buggy applet), or potentially even if you just restart cinnamon, the entire session is gone, along with all of your open apps (and any unsaved progress). This is a huge regression for us, and while there may be a way to work around this, I suspect that it would require huge architectural changes throughout the project in order to do it, and even then, it may not even be feasible.

Then there’s problems like Nvidia support still being subpar (from what I hear), and utilities like redshift and screen recording apps wont work out of the box.

TLDR Wayland isn’t ready, and I personally wouldn’t expect significant motion on this in the near future.

Stephen Collins (collinss) on GitHub

If you ask me, this argument seems to fall apart like a house of cards. Here is my dissection of Collins’s argument and why each point is invalid in 2023:

  1. Nvidia support is subpar. Yes, in 2021 (and definitely before that), Nvidia was dragging the Linux Wayland experience back. However, things have gotten so much better in 2022 that this comment feels like Collins and/or the rest of the team didn’t do their research. In 2022, Fedora, Debian, and even Ubuntu have all moved to Wayland (after testing the waters in 17.10, reverting to X11 in 18.04, and then back to Wayland again for all users in 22.04)
  2. Screen recording apps won’t work out of the box. Have you heard of OBS Studio? It’s a FOSS app and currently works on Wayland! (Again, the Cinnamon developers not doing their homework.)
  3. “Wayland isn’t ready, and I personally wouldn’t expect significant motion on this in the near future.” Wow, how to spot a liar in action 101. Seriously, when GNOME & KDE Plasma have fully working Wayland sessions and MATE, Xfce, and even Unity, a desktop environment developed by an even smaller community as of 2017, have plans for supporting Wayland, with Xfce preparing version 4.20 to have full Wayland support, well, how can you be so ignorant of basic user trends?
  4. Wayland doesn’t run in a separate process and if the DE crashes, the entire session crashes. While there is some truth to this, for most people who are building a “stable” and “Windows-like” distro, this shouldn’t be an major issue! I mean, it’s been over 10 years since GNOME started adding Wayland support to Mutter. Your window manager, Muffin, is literally based on Mutter! Is there anybody on the Cinnamon team that knows even a shred of information about Wayland!?
  5. Collins suspects that it would require huge architectural changes throughout the Cinnamon project in order to add Wayland support, and even then, Wayland may not be feasible for Cinnamon. I’ll discuss this in the next section, but if this is true, then Cinnamon is in danger.

The Solution

As you can probably tell, Linux Mint has some issues. On the other hand, Cinnamon is practically a zombie. With the team betting on the wrong horses, kind of like how Ubuntu still embraces desktop Snaps when all other distros seem to be collapsing on Flatpaks, things don’t look to good for the future of Cinnamon.

Fortunately, there are several (practical) options for Linux Mint to deal with the Wayland issue and survive (and possibly continue to thrive) as a Linux distro:

  1. Drop Cinnamon and move to Xfce or MATE by default, dropping down the supported desktop environments to just 2. I mean, Xfce appears to be moving quite nicely on Wayland support; why not use them!?
  2. Drop Cinnamon for KDE Plasma, which has had Wayland support for a while now, and if need be, add extensions to customize it. (You know, like they used to!) Obviously, this would be easier said than done given KDE Plasma uses Qt rather than GTK of the current DEs Linux Mint uses, but porting Linux Mint’s apps over to Wayland may be easier than porting over Cinnamon proper to Wayland.
  3. Put in the time and work to port Cinnamon to Wayland. Again,, this would be the best choice for users, but is it going to be done and does the team have the manpower or will to do it?

Whatever option Linux Mint chooses, they need to communicate their plans soon. I cannot and will not recommend Linux Mint Cinnamon or Cinnamon on any other desktop environment right now until they reveal their final plans for Wayland. But I do hope that they don’t stick their heads in the sand and ignore this pressing issue. Because in the Linux space, you don’t want to bite the hand that feeds you; in this case, your users. And look, X11 still works for now. But will that be the case in 5 years time or whenever GTK 5 comes out? Will other DEs get features Cinnamon developers can only dream of? Only time will tell.

Reputation and nostalgia kept Ubuntu alive (as well as business customers). But will that be enough for a community-focused distro like Linux Mint?

Thank you Linux Mint for your service with Cinnamon. But maybe its time to right the ship or leave the ship behind.

I know I did.


Posted

in

,

by

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *

en_USEnglish