A while ago someone I didn’t know contacted me in reddit about the list of distributions without systemd, suggesting I should add ataraxia linux. Ataraxia Linux was very similar to kiss-linux, source based, musl, but featuring an array of init and service supervision systems available to choose. A few weeks ago, someone (turns out to be the same person) added a comment to this blog to this more precise list of “linux distributions without systemd” and suggested I should add Ataraxia to the list, and I did.
Then a comment comes in by someone who is really going through the list of distributions and also has been reviewing the narrow list of distributions, or systems, built on musl instead of glibc. He alerted me to the fact that Ataraxia is now using systemd as its default init system. It didn’t strike me as odd in the beginning, even though I had recently given it a try. I was under the impression, never tried, that systemd couldn’t be built on musl as it wasn’t even written in C. So I asked that same guy, who turns out to be the distro owner of ataraxia, how has he achieved this, and I got a single word response: “patching”.Ok, so Ataraxia is dropped from both lists now, but this is not all. It is a trend among the distributions that do not use systmed as init to be open to either parts of systemd, as elogind, in order to support complex popular desktops and the software that are developed within those platforms. The same distributions will also have as default audio management pulseaudio, attempt to have gnome and its associated desktop variations functional, etc. etc. This can be perceived as cheating in a way. It is almost like an effort to circumnavigate around a legal clause. Like if there was a law around this community of distros not to use systemd as init, but you can use it for everything else it does. First offender that comes to mind is MX linux and its relatively recent popularity according to Distrowatch statistics. MX uses antix scripts to start the system but has the entirety of systemd installed and used by software instead of substitution libraries. Systemd is not running but it could if you chose to, you just “turn it on”.
The point I am trying to make is we should be differentiating the choice of alternatives on technical merit (simpler, smaller, more controllable, faster/more responsive, non-redundant, easier to audit and troubleshoot, etc.) and the political matter behind its use. Systemd clearly and since its inception was financed and promoted with money from the giants of the computing industry. RedHat at its start was the joke of the computing industry, but it made an entry with a big bang. Especially in the US, apart from the frenzy of AOL (an internet provider for the computer illiterate) the entire country was flooded with disks that supposedly substitutes windows with linux. It really didn’t, and in the few occasions that it did boot it was so dysfunctional and frail most people gave up trying. But it did flood the world with free CDs to install redhat. That is a large funding start not given to just anyone without a goal to match the financial backing. In other words, someone co-signed those loans to make redhat what it is. Well, years passed and IBM came out of the “blue” and bought RedHat, or should I say made their relationship known with a public engagement. If Red-Hat had failed and created too much negative turmoil, IBM would have kept its distance.
For every public voice that came out in the past 10 years in any public forum to criticize systemd in technical or other merit, 10 more voices would “instantly” pop-up to discredit not the content of the criticism but the person voicing this criticism. In the best case scenarios where real known experts voiced their concern, the discussion was diverted to a “manageable” topic: systemd vs sysvinit. Since systemd is better than sysvinit, and it is newer, modern, it is better. Hence there is no need for more discussion, let the best system prevail as a monopoly.
Many other corporations diverted tremendous funding as to be present and part of the booming open and free software industry. Why? Because if you can’t beat your challengers, you might as well join them as to dominate them with financial power from within. Google, Microsoft, Oracle, HP, IBM, Cisco, Motorolla, have all contributed open and free code. Lately Facebook has also joined the circus and before you know it the most popular distros are incorporating Facebook’s code into their package building and management.
There is a big difference between someone writing code on their free time and without direction and control, willing to share it with others, allow others to contribute, help, criticize, locate bugs and assist development, without direct financial interest, and large corporations hiring people to code under their direction with a specific “financial motive and goal”, to dominate and increase profits. Directly or indirectly. The first group may indirectly expect some user support and financial help to keep doing what they have started, the later group writes code with specifically dictated goals, guidelines, and are only doing it because they are paid to do it. When the mafia stopped developing consolekit nobody really stepped up to the plate to continue the labyrinth development consolekit had become. ConsoleKit2 did receive some attention but then it fell victim to the orphanage of linux packages.
I am asking this question, is the aim of not using systemd solely on technical merit, a better, simpler, faster, more manageable system is here and available, or is it an attempt to block and cleanse linux from mega-Corporate influx and control? In the second case any code contributed under any license by any large corporation, written by replaceable employees, should be resisted and not utilized.
Can we really have such a system?