Plan "b" for when ubuntu is no longer a viable/desirable/etc base for kde neon?

On the front page of the website of KDE Neon, it says : The latest and greatest of KDE community software packaged on a rock-solid base.

Which is precisely the point of Flatpak, a rock-solid base and the latest of software.

… Arch snapshots, immutable and with a bit of extra testing, have the potential to work for end users where LTS-based distros have always failed. In addition to being free from corporate meddling and GNOME-centric decision-making.

I agree with the release cycle sentiments.

Packaging Question:
Would an Arch base underneath KDE Neon imply anything re: Flatpak vs PkgMgmt ? In KDE Neon right now, I find the duplication of many packages to be confusing… and Discover would be better with Flatpaks, without the duplicate distro packages.

Yes agreed to what some said. KDE Neon / Kubuntu+Backports is already very weird but Debian often is said to not workout well.

It is an arbitraty separation between the desktop and the rest OS. My personal Fedora Kinoite experience was very nice.

I have limited experience with regular Fedora KDE as that broke pretty quickly too so I switched to Kinoite.

Had some issues but nothing big, and Fedora is not opinionated I would say. Using the Ublue framework you can build any variant easily, which just works, like Secureblue (“kinoite laptop userns hardened” is what I use).

There also already is a beta Kinoite image for testing KDE 6.

I would just push this up a bit, as KDE Neon seems to be a mess again.

KDE without BTRFS snapshots or even an atomic system seems to be total chaos.

Personally I think openSUSE would be the better choice if it ever came to that. openSUSE did KDE Neon before it was cool :sunglasses: (Neon User is pretty much equivalent to openSUSE Argon, Neon Unstable is very close to openSUSE Krypton).

Plus a huge advantage: thanks to zypper’s repository management ability, you can easily switch back and forth between Leap ↔ Argon and between Tumbleweed ↔ Krypton.

That’s not even mentioning btrfs snapshots, and we could rely on OpenQA. We rely on openSUSE builds for our CI too.

6 Likes

The real problem is that Neon is mixing a desktop that is moving fast and pushed in a rolling release model and a distribution that holds packages for stability.

Even Arch Linux is still holding Plasma6 in testing repo, while Neon advertises itself as complete. Also, Kubuntu team didn’t ship Plasma6 until it gets enough testing.

I think Neon should be labelled as testing/developer distribution to not give people bad reputation of Plasma.

Plasma should add another testing period for its final version, for example, it would be best if it’s released to Neon then wait 4 or 6 weeks to launch its official release.

Arch Linux did have KDE-Unstable repo for testers, but even after the official release Plasma6 was pushed to testing repo, and it did not get yet to stable repos.

I think the most challenging problem for Plasma6 is to handle bugs coming from gamers using dual GPUs on Wayland, especially that cursed NVIDIA.

5 Likes

Exactly! :+1:
In bold letters on the website and download page.

That’s what I have been thinking for years.

3 Likes

Fedora as the base i feel may not be the best choice, I feel its far too opinionated on things vs other distro bases. While Fedora is fantastic and stable that opinionated nature may be a maintenance burden.

Arch is a great base BUT that will often scare people who think its way more complicated than it is. Its community driven nature, great packaging system, and being less opinionated make maintaining less complicated.

Opensuse may end up being the best option even though Im personally not a fan of the its unique bits. Its a more middle ground between Fedora and Arch though the media codecs decision may make things complicated for some people looking for an OOTB ready to go experience.

1 Like

The best choice is to avoid any distribution backed by any commercial company like Red Hat, Canonical, SUSE… Those kinds of distributions always have a governing board consisting of people appointed directly by those companies via sketchy votes, so the distribution decisions are always aligned to follow money, like what happened with the great and loved CentOS distribution.

This is the main reason why Valve decided to use Arch Linux on their Steam Deck, because it’s currently the only Linux variation guided by a real open community.

3 Likes

That would only leave Arch and Debian out of “the big five”.
Personally I would prefer openSUSE for the aforementioned reasons - but I am neither a KDE developer, nor a packager, nor… And those are the people who should and must decide what is best for them!

6 Likes

Personally I like that it’s ubuntu/debian based.

For software, when searching the web most often than not there’s always, and sometimes only or in priority, “ubuntu”/debian based solutions/how-to.

For hardware, if there’s even a mention of linux, it’s easily ubuntu/debian derivatives.
Many times (if not always) there’s no mention of linux or we get a typical and infuriating “we don’t support linux”.
Happens with motherboards, ssd, trackball mouse, monitor software, racing sim steering wheels/pedals, flight sim gear, and more.

Going with linux distributions less heard-of by the commoners/outsiders of the linux world doesn’t seem like a good idea. It’s like going “niche” of a “niche” o.s.

I’m thinking something like a distribution without the whim of corporate/political decisions, one step (not a full marathon :slightly_smiling_face: ) removed from the bleeding edge.

Kind of debian (sid or testing) + rolling kde software.

1 Like

Why not? As far as I understand its purpose KDE neon is not meant for them anyway…

Here’s my two cents: especially after the recent fiasco it is very clear to me that y’all should be prioritizing stability over shiny new updates. There are actual end users like myself who download this “not-a-distro” thinking that it will be the most stable and highest quality permutation of KDE products, and end up getting burned by it. There already have a lot of users who want stability, and you can’t just turn this into a broken Gnome OS analogue, it’s wayyyy too late for that. Neon works because it has a stable base combined with a nice shiny layer of updated KDE products. Turning Neon into an Arch-based OS will simply drive users to Kubuntu, which is inferior, and will give Plasma devs a lower ability to showcase the direct vision for Plasma. One of my most common complaints about Gnome is that they try to dictate to distros what to do instead of showing what they want and letting people be.

Fedora simply has a lack of non-FOSS software availability for anything beyond Flatpak and rarely do I see RPM releases from official software downloads. PPA’s rule the Linux world as well, and Ubuntu PPA’s aren’t compatible with Debian PPA’s, so it’s just not the same. Many .debs wouldn’t work either. Switching to Arch would be disastrous, the Arch -->USER<-- Repository is not a viable substitute for official releases off apps for anyone who isn’t skilled at using Linux, I don’t care how good it is it’s not an official release source and not centrally managed and should not be substituted for that. Ubuntu has a large package base anyways with just as much availability of software on a more stable base. If you don’t like aspects of Ubuntu just cut out the bad bits like Mint, Pop, Elementary, and others do.

The only reason Arch was viable for the Steam Deck is because Steam OS nukes the Root File System every update. Most Steam Deck users are only sticking with Flatpak+a few scripts, and the ones who don’t end up with a horrible experience if they’re not careful. You have to set up a root password on your own. Why do you think Valve used Debian before Arch? They wanted a stable user experience that was idiot proof. Steam OS is idiot-proofed too much to be a great desktop for only pc usage, and Arch is too complex for regular users.

How are you going to update to the new base? Leave Neon-Ubuntu users stranded? Put in all that work to move everything over for no tangible reason besides slogans, simply so you can have your playground?

2 Likes

Neon is supposed to be the flagship OS for KDE, quite frankly it needs to be the highest quality KDE distro. Let’s hold off making the root file system unchangeable when vital packages are mispackaged and the only fix is adding files to that root file system.

1 Like

+1 opensuse Tumbleweed

1 Like

Also not a dev, but as a user, I personally found opensuse very frustrating. I’ve used a lot of distros and most are very same-y when it comes to how to config stuff, etc. But opensuse has big not-invented-here-syndrome.

In my experience some things were broken out of the box and then when trying to fix those none of the standard solutions worked and it was a nightmare.

If you want the latest desktop to be truly stable and provides maximum performance on new hardware, then you should bring also all the latest required low level libraries, GPU drivers, kernels…

Only true rolling release distribution like Arch can provide this kind of beast for Valve.

Putting the latest Plasma on top of old libraries is never going to work.

openSUSE certainly has its quirks, but perhaps I got used to them in the past more than 20 years on- and off-relationships with S.u.S.E./openSUSE
And the pre-configured snapper/btrfs is a godsend for trying out things.

But we can discuss all we want about the advantages and disadvantages of a distribution base for a KDE Plasma showcase - finally it is and (rightfully) should be the decision of the people who will use it for developing (and showing) KDE stuff…
At least that’s how I have understood so far why KDE neon exists, maybe I’m wrong.

Is it a distro? Not quite, it’s a package archive with the latest KDE software on top of a stable base. While we have installable images, unlike full Linux distributions we’re only interested in KDE software.” (from the KDE neon FAQ)

1 Like

I don’t really care about it being Ubuntu based. I just don’t feel like reinstalling my system all over again. I haven’t reinstalled in over 5 years with Neon.

This is a bit of a misunderstanding that many people seem to have about software development in general. People seem to be under the impression that software development is steered with all hands going one direction which simply isnt the case. The KDE team doesnt just all do one thing at a time there are a number of people and there is always someone working on big fixes. Plasma 6 is the first major QT version change in 10yrs so there are bound to be a few pains with it and a huge chunk of the commits over these last few months have been bug fixes and stability improvements for both Plasma 5 and Plasma 6.

Did anyone at KDE actually make this claim? Can you get mad at someone for an assumption that “you” made? The only claims ive seen are of the ubuntu LTS base being stable, which is true.

EDIT: I want to make clear, Its perfectly fine to be upset about the state of Neon at the moment as Neon currently seems to be the release of Plasma 6 with the most issues even though its supposed to be the KDE showcase. The language on what Neon is has been inconsistent but i think thats less of the overall issue with this release and more a problem of “what the heck happened with packaging?”

The AUR isnt even considered the best idea on Arch, there are plenty who use it but when it comes to stability, security, and maintenance the wisdom is dont use it. This really has nothing to do with KDE ever possibly using Arch as a base and even on the AUR page it says “DISCLAIMER: AUR packages are user produced content. Any use of the provided files is at your own risk.” as it is meant for those who can deal with the issues it may cause. Im not sure why its being mentioned tbh as the Arch folks dont consider it an official release source and never have.

The issue isnt the Ubuntu bits but having the fact that ubuntu LTS (and debian stable ) tends to become very stale and having to deal with sometimes very old packages along side very new KDE can be problematic.

It does have a large package base, but the issue isnt the size of the package base which i dont think anyone was arguing against.

Valve went with Debian because that is what the scout runtime was based on and at the time it seemed like a good idea to keep the system fairly static. There is also the fact Arch was in the middle of a transition to Systemd and the package management lacked signature verification at the time Valve was making SteamOS. They moved to Arch specifically for the reasons that people even are suggesting it here. Debian is freaking ancient most of the time, which is fine in some cases but not in others. It became very clear that debian or ubuntu lts simply dont cut it as the base for a gaming centric platform as theyre just too old.

Arch isnt complex and its in their very motto not to be, the hardest part of arch is the install which has already been solved by a number of other distros and the fact that they simply dont support GUI package managers. Valve didnt make it immutable because Arch is too complicated, they did it because they needed a reproducible system that should be exactly the same across every deck. Most of the time immutable distros are about being 100% reproducible and little to do with protecting the system. They went with Arch because it reduced the maintenance burden for keeping packages updated because theyre 9 times outta 10 gonna be at most a couple months old or brand new (LLVM tends to lag a bit on Arch).

This is a valid question, what would happen to existing installs? Thats not an easy question to answer in this case.

1 Like