Thoughts on dropping Debian to make PureOS an Arch-based distribution

Do PureOS developers have any thoughts on dropping Debian to make PureOS an Arch-based distribution?

Both Debian and Arch have their strengths. However Debian’s conservative approach, while precious on webservers, is not considered a value for Desktop PCs, and in this case Arch is generally considered to be a better option with its rolling release model.

On the mobile side, distributions like Manjaro have shown the power of distributing Phosh in a environment that has pacman as package manager. Furthermore Valve is switching from Debian to Arch for its Steam Deck portable console.

Installing experimental new packages in Arch is way simpler than in Debian. And a 100% free/libre distribution based on Arch already exists and could constitute the basis for an Arch-based PureOS: Parabola GNU/Linux-libre.

Is it not worth wondering whether PureOS become too an Arch-based GNU/Linux distribution?


Speak for yourself? As much as I love to geek out, I also value not having to geek out i.e. things that work and stay working.

At the end of the day, is the case for change compelling enough?

For sure there would be costs in changing the distro that PureOS is based on. I know what I would rather Purism be working on.

Worse still, there would have to be some migration approach. You can’t expect all existing customers to blow away their system, install from scratch and then re-establish their setups - well, not without a really really strong reason (such as distro going away or too poorly supported).

I also value the large market share of the Debian family, which means that if I am having some problem that I can’t work out, someone else on the internet is more likely already to have had the same problem and come up with a solution. (I don’t want to start a distro-war argument here about which distro has the largest market share etc.)

Another angle to consider is

Nothing stopping you installing Parabola on Purism hardware? Purism goes out of its way to market that you own your hardware and you choose what software to put on it.

At the very least, have you tried to install Parabola on Purism hardware? (To make it easier, let’s say a laptop or the Mini, rather than on the Librem 5.)

Final thought for now: If there were a single killer feature in Arch that Purism particularly liked, maybe they would be better putting the effort towards upstreaming that feature into Debian. Of course, in the open source world, features can flow in all directions. :wink:


Debian has experimental packages as well.

This sounds quite like a myth. I used Ubuntu for years starting from 2008, and I remember that with every distro “upgrade” (that thing that they do every six months) I needed to pray the gods that my computer wouldn’t break down in a mess that would force me to re-install everything from scratch. I switched to Arch and… Wow! No more distro “upgrades”! Everything self-sustaining for years. Maybe once every two years you are required to launch a couple of shell commands as a due manual intervention, but besides that the system remains consistent and always (very) up-to-date.

Migrations are already happening as long as you stick to Debian. The next one will be the migration from Amber to Byzantium. These “migrations” are literally what made me abandon Ubuntu.

Arch community and Arch wiki are generally considered the most pro-active and problem-solving community and documentation. But I agree, we shouldn’t distro-war about that.

There is at least one killer feature for me: Arch’s package manager, pacman.

1 Like

I am obviously opinionated about this. But comparing any distro’s experimental packages with the simplicity whereby any user can upload a package to AUR is a lost battle. AUR is literally one of Arch’s workhorses. It reaches the point of making Flatpak appear as a useless complication, created for making life in distributions that do not have pacman just a little bit more bearable.

1 Like

What does Ubuntu stability have to do with Debian?

1 Like

The distribution model is the same, unless you switch to Debian unstable, which is… well… unstable.

A way to think of Arch is “a stable version of Debian unstable”, with added facilities for creating new packages from source.

Hate is a very strong word. As a fellow penguin you should remember what the Community Code of Ethics and Professional Conduct are

  • Protection of the public interest
  • Demonstration of personal integrity
  • Maintenance of confidentiality and privacy
  • Showing respect for colleagues and collaborators

Withdrawing my previous post. All good.

Switching to Arch may have been a good decision for Valve, but I don’t see any advantages for Purism. Valve needs to support the latest hardware, since it is a platform for gamers, and gamers often live on the cutting edge. Purism on the other hand is always behind the tech curve, since it takes a while for FOSS drivers to be created for the latest hardware and Coreboot to support new generations of Intel processors. Then, it takes even longer for Purism to make a custom designed laptop based on what is supported by 100% FOSS. The i7-10710U was launched in August 2019, but the Librem 14 with that processor didn’t ship until April 2021.

Arch is not a good platform for corporate customers who care about stability and long term support. Most commercial Linux software supports Debian (or Ubuntu LTS), but not Arch, and Purism is trying to reach the corporate market. For example, my job is offering independent consulting for ProcessMaker. There are no instructions for how to install ProcessMaker on Arch, but there is for Debian, and it is the same situation for almost all the commercial Linux software. Arch also isn’t a good choice for people who care about security, which is one of Purism’s chief marketing points, because it changes too often and doesn’t have enough review before new packages are introduced.

Purism is trying to reach nontechnical users as well, and frankly Arch is not a good choice for that market segment. Another issue is that Debian main offers better support for people who want to use 100% free software than Parabola. It there is a technical problem with Debian main, you have an organization of 3000 developers that wants to get that problem solved. With Parabola, you only have a handful of developers, and the wider Arch community is not particularly concerned if their system needs a bit of proprietary code to function, whereas Debian is far more concerned about excluding proprietary code from main. Also, Debian has its roots in the Free Software Foundation, and most FSF diehards are still using Debian or one of its derivatives, whereas fewer of them use Arch. Based on the number of page hits it got over the last 12 months, DistroWatch ranks PureOS as #46 and Trisquel (an Ubuntu derivative) as #83. In contrast, the two 100% free software distros based on Arch are Parabola #146 and Hyperbola #184.

If we can believe DistroWatch, there are 3.4 times more users of the Debian family than the Arch family, so there is a higher probability that a buyer of Librem products will prefer that PureOS be based on Debian rather than Arch.

Distro family # distros Hits per day % of total
Debian 126 26376 53.54%
Arch 23 7799 15.83%
Red Hat 24 2845 5.78%
Slackware 10 1071 2.17%
Gentoo 11 920 1.87%
SUSE 5 1225 2.49%
Mandriva 4 868 1.76%
Puppy 3 829 1.68%
LFS 4 299 0.61%
CRUX 4 238 0.48%
OpenELEC 2 165 0.33%
Void 2 330 0.67%
Recalbox 2 81 0.16%
Pardus 1 47 0.10%
ROCK 1 104 0.21%
Android 1 230 0.47%
Independent 29 3370 6.84%
Non-Linux distros:
BSD 15 1694 3.44%
Solaris 4 291 0.59%
ReactOS 1 200 0.41%
RISC OS 1 58 0.12%
MenuetOS 1 58 0.12%
Haiku 1 166 0.34%
Total 276 49264 100.00%’s unscientific poll finds similar results, with 2.7 times more users of the Debian family than the Arch family:

Distro Distro family Votes % of total
Mint Debian 1784 10.85%
Debian Debian 1073 6.53%
Manjaro Arch 1011 6.15%
Ubuntu Debian 2599 15.81%
Antergos Arch 262 1.59%
openSUSE SUSE 568 3.45%
Solus Independent 1274 7.75%
Fedora Red Hat 1441 8.76%
elementary Debian 927 5.64%
Zorin Debian 62 0.38%
deepin Debian 120 0.73%
TrueOS BSD 19 0.12%
CentOS Red Hat 288 1.75%
Arch Arch 1120 6.81%
PCLinuxOS Independent 2852 17.35%
Other 1041 6.33%
Total 16441 100.00%
Total Debian 6565 39.93%
Total Arch 2393 14.56%
Total Red Hat 1729 10.52%

At any rate, Purism is now offering both Stable and Testing (Byzantium) versions of PureOS, so Purism has a solution for people who want a rolling release. Testing is usually only a month behind Arch in terms of the software that it offers, except when Debian is doing a freeze in preparation to launch a new Stable version once every 2 years.

Finally, think about the fact that Purism’s Matthias Klumpp has spent years developed Laniakea, so it takes much less manual labor to maintain a Debian derivative. To develop an equivalent system for Arch would take a long time, and distract Purism from the critical dev work that it is doing on the Phosh mobile environment.


Yes but Debian is still a great tool.

1 Like

Well, it could take less if the FOSS drivers ship faster into the official repositories, which is what regularly happens with Arch – although in this case there would not be probably any difference, since Purism would take care of uploading their drivers of interest whatever package manager they use.

You might be wrong about this. There are no instructions for how to install ProcessMaker on Arch because ProcessMaker is already in the Arch User Repository (obviously someone uploaded it there). I have used both Debian and Arch, and the situation is often the following concerning commercial (non-official) software:

  • Debian: search for the install instructions, or possibly the .deb package of the commercial software
  • Arch: in 99% of the cases someone has already uploaded the commercial software on AUR; in the 1% of cases where nobody did, I will do that (although I will try not to use commercial software when possible)

Manjaro (based on Arch) is often considered way more beginner-friendly than Debian.

That might be true if we focus on history (historically Arch was born as a crossover between GNU and BSD), but, seriously, today the difference between the two is only the package management. Developers close to the FSF might indifferently use Arch or Debian, and the FSF does not endorse either Debian or Arch anyway (but Trisquel and Parabola instead).

We are really talking about the two most widespread user bases, and between the two we should also distinguish the qualities of such user bases.

That would be Manjaro’s approach (six weeks behind Arch).

There might be some truth here, although there would also be a huge help from the community of Arch enthusiasts, which would equalize the effort. Theoretically Purism’s approach could be that of simply forking Manjaro and removing all its non-free software.

FWIW I would love PureOS to be based on Arch. I’ve had equal problems with each distro, I’m just end up stuck with them forever in Debian. The only remaining reason I use Debian on my work laptop is because I dual boot with Windows 10 and don’t want to run without Secure Boot enabled on a work machine. Debian has a legit cert that works out of the box, where as Arch you have to mess around with shim and all that to get secure boot to load. I use PureOS on my personal boxes now as I want to stick close to Purism’s work and vision of their ecosystem, though sometimes I get tempted because I like running the newest software. :slight_smile:

1 Like

There are multiple ways to express the same sentiment, but this is one of the least constructive ones. Please refrain from doing that.


I would say that one does not simply fork a distro :slight_smile: It costs, and you never know how much it will cost until the fork is done.

1 Like

Not going to happen.

BTW. I use Arch.


Screenshot taken for posterity.

I can imagine it might cost… But what if…

…What if Purism started a parallel project, let’s call it PureOSNext, investing as few as possible and filtering out at the beginning Manjaro’s packages that do not appear in PureOS, simply for facilitating the experimentation process of new software. Slowly PureOSNext would look more and more similar to PureOS, only it will have more recent packages. It would work as a testing ground for PureOS.

One day in the future Purism might decide that PureOSNext (an Arch-based distribution) does well its job as a test distro and must remain as such. Or it might decide that it should replace the old PureOS.

What if…

After all Valve did it.

Maybe some arguments about why not?

First one: Purism invested and keeps investing in tooling for maintaining Debian derivatives called Laniakea, there’s no reason to let that work go to waste.

Other one: lots of people working on PureOS have Debian background; we have Debian Maintainers and even Debian Developers around. Having a good relationship with the base distro helps a lot.

Another one: there’s hardly any reason to switch away from Debian in the first place. We’re not looking for another base distro to use, Debian serves us well.

Yet another one: this team is small enough to make maintaining single and rather slowly moving distro pretty challenging already. Creating “PureOSNext” on the side just because someone asked “why not” would be just insane :stuck_out_tongue: