With surprise I read Marks article when I woke up last Friday. And surely I was shocked.
It makes me sad to see Upstart being abandoned (at least for Ubuntu) and that we are forced to use systemd after 14.04 LTS.
It took me a while to swallow this, and this post is a try to express my concerns. This can be controversial for some people, so if you are not
in the mood of reading a controversial post, please stop here and press the 'Back' button of your browser or scroll up or down.
Don't get me wrong, I appreciate the work of all contributors of systemd, but I fear right now we are depending on
one single entity named RH. Just because they are paying the main drivers of systemd.
So, what happens when you remove one important part of a nicely build machine? ... It will break.
This can happen to systemd as well.
Honestly, I don't want to disgrace the committers and contributors to systemd, but I have the feeling, that
the main drivers of this project did build up a great knowledge
of all the needed parts of a working Linux boot up process.
Removing this knowledge slows down the process of development and improvements.
RH is a business driven company, so if it doesn't make sense (anymore) for RH to invest
into a certain project, dedicating some expensive human resources, then they will remove
those assets and/or money investments. Means, removing the main drivers.
So what happens than?
Especially now, in the "OpenSource" business, we have at least 3 competitors, and at least 2 of them are fighting for "World Domination".
And we already know, that sometimes the collaboration between two commercial entities can be pretty cumbersome, especially when their business models are diverting/clashing and/or their ideas and visions are not really aligning.
I am asking myself:
- How can we make sure, that a crucial part of a distro is not in the hands of one commercial entity?
- How can we make sure, that systemd will be a project, which is not depending on only one company?
Moving the whole systemd project (and all payed human resources involved) into a non-profit organization would make sense to me.
Companies can become Members of this non-profit, can invest money into this entity, the non-profit can pay human resources, and the people working for the non-profit will deliver the software we all need.
People working for this non-profit are not allowed to work directly with the members, they need to be independent. The board of the non-profit can receive ideas and wishlists from members, and use those ideas/improvements as backlog list for the project.
Does this sound insane? Or could it be a way to "streamline" the collaboration process between all parties?
What about the other (non-commercial) distros out there?
So, let's say Fedora, RHEL 7, SuSE, Debian and Ubuntu are all using systemd.
Most likely all derivatives of those distros will use systemd as well.
But what about distros like Gentoo? Right now, Gentoo is using OpenRC as their main init system. And from my experience during the last weeks, when I dived back into Gentoo, it's a real cumbersome process substituting OpenRC with systemd (mostly when you want to run a gnome 3 desktop, and you already have a running system).
For me, it would make sense when Gentoo is also changing their init system to systemd, to be more streamlined with the other major distros.
But really, it needs to be discussed and a decision needs to be made. But this time, it shouldn't be handled like the discussion we saw in Debian-land.
I know there is already a debate going on inside the Gentoo Community, but to make my point here, while OpenRC is a good and reliable init system, Gentoo should know now, that they are also losing a battle. Gnome 3 as as desktop depends on some parts of systemd, and breaking all those parts into separate projects, doesn't make sense anymore, because the majority of the mainstream distros decided to go with systemd.
I think this is true for all the other distros out there which are not really backed by commercial sponsors.
Now, most mainstream distros are running/will run systemd as their main init system.
With this in mind, there is now a monopoly of one project, which is mainly backed by one commercial entity.
A project which is one of the important parts of a running Linux System.
My fear is that this monopoly will be used (one way or the other) to hurt other distros which are not under the roof of this commercial entity.
Especially those distros which are under another roof of another commercial entity.
The idea of moving this project into a separate non-profit organization would make sense, to prevent this from happening.
Right, it could just be me who thinks like this, it could be that I already saw
so many barking dogs biting, while the society says, barking dogs don't bite.
But honestly, for me this situation feels wrong.