How do people have the time to administer free distros?
- Anmelden oder Registrieren um Kommentare zu schreiben
Distros like Debian and Fedora probably have a million developers, so none of them has to do an insane amount of work.
But small distros like the libre ones, run by a few (I assume) people: Are these people retired, workaholics, independently wealthy? How does work on the distro not take over their lives?
Good example of why Hyperbola's development is currently stalling so bad.
They need funds SO badly...
Comparing against Hyperbola, Trisquel has survived the 20-year test specifically because it keeps its scope tight. Trisquel focuses solely on enforcing software freedom, keeping changes small, and automating as much as possible, rather than reinventing the wheel. It's all about resource management. Linux-libre uses the same strategy.
I agree with you on the points after software freedom for sure. The one before? mmm... not as much. but that being said, I consider software like dbus, systemd, pulseaudio, and other similar redhat stuff to be non-free due to it being deliberately made to break backwards compatibility and force adoption.
The thought of keeping changes small is for this reason not good enough for software freedom in my opinion.
This being said, Trisquel is definitely an improvement over ubuntu. Due to them, ubuntu devs, trying to put everything in snaps and other alternative package managers. Maybe flatpak too? I don't really know for sure.
I will admit Hyperbola devs are going beyond the mission I thought they had set out for and are removing some software that doesn't break backwards compatibility and adopt services due to it being a non-community type development around it.
I disagree on that though.
And there are other little things too of course. But the long and short of it is, you do have a point on them changing too much.
However trisquel is keeping changes too small for my tastes as is devuan and debian and largely most distros.
So... I am stuck in a weird place right now.
I am currently using Devuan with linux-libre on my main comp till HyperbolaBSD is out and gets updated more, etc...
It's important not to conflate software architecture with software freedom. By the FSF's definition, software such as systemd and dbus are unequivocally Free Software.
The Free Software definition on gnu.org is there to check, and things like "retain backward compatibility" aren't in there. It's better to use a different term for architectural issues rather than conflating them or trying to re-define the term by using using some different, other definition because free software and non-free software are already very well defined terms with specific meanings. You're perfectly free to criticize any software you want, of course. However, calling it non-free factually misrepresents the definition of Free Software.
Yeah... that kind of stuff puzzles me, truly.
Because while they break backwards compatibility, they are actively trying to restrict freedom by limiting other options. For example, Gnome devs are currently saying their next gnome version will only work with systemd. They plan to make that restriction even tighter.
Why does a desktop environment need an init system?
Mate, LXDE, XFCE4, LXQT, for all their annoying faults don't require this.
Redhat has some software that is similar in that way. So i guess i don't understand the FSF's position.
"they are actively trying to restrict freedom by limiting other options."
If you really want to understand the FSF's position, you have to evaluate it from their point of view, not your own - which means not thinking like that. Otherwise, it will probably be very difficult to understand indeed.
The disconnect here is that the FSF is talking about 4 very specific freedoms, not some overall concept of freedom or about "freedom of choice." If this were simply about having choices, then having 10 different software options that are all entirely proprietary would be considered a wonderful thing. But that scenario is the exact opposite of software freedom.
Richard Stallman addressed this exact mindset perfectly when he said, "To have the choice between proprietary software packages is being able to choose your master. Freedom means not *having* a master."
So it's not about choice itself as a value.
When GNOME developers tightly couple their desktop environment with systemd, they reduce your out-of-the-box choices, but they do not restrict your software freedom because you still have the legal right to take their code, rip out the parts you hate, and build something else. That legal right is exactly how projects like Devuan are born. It's how MATE got started.
They aren't taking away your software freedom; they're just making architectural decisions you don't like. Software freedom means you have 4 specific freedoms, which include the right to fork and change the software yourself without asking permission, not that the original developers have to design it the way you want. But nor do you have to put up with their design decisions either - Invoke freedoms #1 and #3. I recently did that with another program: https://jxself.org/narrowed-estraier.shtml
Okay I guess? but why a desktop environment needs that is beyond me.
It feels like they are trying to give the middle finger to people who don't use systemd if nothing else.
And for that reason they, GNOME and others likewise, have my contempt and bewilderment.
From a developer's perspective, systemd solves several complex plumbing problems that a modern desktop environment needs to function, which are very hard to replicate otherwise.
In the past, desktop environments had to use sketchy workarounds (like running X11 as the root user) just so your desktop could talk to your graphics card, mouse, and keyboard. systemd-logind manages this and safely hands permissions for these devices over to your specific user account when you log in, making the system more secure.
A modern desktop isn't just one program; it's a web of interconnected ones: audio servers, file indexers, network managers, notification daemons, etc. systemd-- user acts as a traffic cop for all of this. It ensures programs start in the exact right order, wait for each other, and automatically restart if they crash.
Handling exactly what happens when you close your laptop lid, press the power button, or leave the computer idle requires deeply integrating with the kernel. systemd provides a standardized way to do this, so GNOME doesn't have to write its own custom power-management daemon. Before systemd, GNOME had to maintain its own background plumbing (like ConsoleKit) to handle all this. By outsourcing plumbing to systemd, the GNOME team frees up thousands of hours of developer time to work on the desktop itself rather than reinventing the wheel.
It's a software engineering tradeoff, a decision born out of resource management and technical convenience, not spite. This also circles back to the very topic that started this thread: How do people deal with limited development resources? That problem isn't limited only to distro maintainers. In this thread we've seen how Trisquel and Linux-libre try to handle it by minimizing scope and using automation. GNOME went a different direction with outsourcing. What other ways can people think of?
Why does a desktop environment need an init system?
Systemd is not only an init system. GNOME uses components of systemd to have several graphical sessions (some being remote, through RDP) and to start and manage services inside a session, because they are better than what is the barely-maintained gnome-session, as https://blogs.gnome.org/adrianvovk/2025/06/10/gnome-systemd-dependencies/ explains.
> Are these people retired, workaholics, independently wealthy?
No | presumably | I wish, but sadly no.
> How does work on the distro not take over their lives?
I blame magic, but the power off switch works too ;)
I am a member!
This is something I worry about. I worry that small operating systems don't have a massive pool of labor to scale well. I'll still use Trisquel, but it's something I worry about.
This is a logical fallacy: making an argument unfalsifiable by constantly moving the goalposts. Trisquel started in 2004, so it's 22 years old. If someone worries about a project failing for 22 years, and it eventually sunsets in year 25, they haven't correctly predicted a collapse - they've just outwaited a normal software lifecycle.
Trisquel's older than the iPhone. Billion-dollar companies kill popular things all the time (the "Google Graveyard" is full of them). Corporate-backed distros like CentOS can have their entire structure and purpose upended overnight by a boardroom decision - and have.
At some point, one has to conclude that the distro has proved itself.