04.27.20
Gemini version available ♊︎Running GNU/Linux With a Top Process Controlled by Microsoft and Far Too Many Lines of Code
“In February 2014, musls Rich Felker opined that PID 1 is too special, should be 10 lines of code and not require reboot on upgrade. PID 1 should only start the real init script, and reap zombie processes. All the functionality of systemd then can be provided by the init script and programs run from it. PID 1 so has only a small attack surface, and user level programs can evolve diversly.” –Wikipedia on “Broken by design: systemd”
Summary: Bloated programs may be very feature-rich (plenty of decent features); but what happens when systems that are supposedly Free (libre) become far too complicated to break apart and study, modify, and fork? Exercising one’s software freedom certainly becomes harder and it’s a support contract lock-in (high exit barriers due to “features” or complexity creep)
THE systemd project, controlled primarily by IBM and hosted by/developed on a Microsoft proprietary software trap, is something that merits at least scepticism if not criticism and condemnation, regardless of whom (person/s and company/ies) it was developed by.
“Systemd is well over a million lines of code! Watch the file count: over 36,000!”Nothing should be above criticism, we’re not a cult and Free software is not a religion (it was founded by an Atheist). This morning I decided to check just how massive systemd had gotten, knowing that it adds far more lines than it removes (many also get removed, which makes keeping abreast of this project close to impossible and studying security impact truly impractical). This is why Techrights is moving away from systemd and has partly done that already.
Based on this tool (hijacked by Microsoft of course, seeing a monopoly that wasn’t its own), this is how bloated systemd became:
Oh, look… it’s so bloated that the program gets stuck and never executes until exhaustion (except of resources/timeout). It didn’t expect projects to become so massive. Systemd is well over a million lines of code! Watch the file count: over 36,000! Here’s the output as an attachment (this list alone is a 1.3 MB text file).
As we wrote this morning, eventually we hope to use no virtual machines and no systemd, either. We have our own, self-hosted Git server (not GitHub) and code we’ll eventually share publicly. The code we have is short and functional; it does the job with not even a thousand lines of code. Simple programs are generally superior and are considered more elegant than messy programs with endless flaws and massive technical debt. That’s what many things became, perhaps because it’s conveniently assumed that security doesn’t really matter and old computers should just be retired. This is increasingly the mentality in “Linux”, unlike say in OpenBSD. █
ᶃ Gemini Space
This post is also available in 
Content is available under CC-BY-SA