First of all, don't read the argument section below. Read OT: Open
letter to the Linux World by Christopher Barry instead. He's actually
worded it right, and it's much more intelligent than my document.
After that, read Broken by design: systemd
from EWONTFIX.
And then read Participating
in the [systemd] Debate by Steve Litt.
Then, if you want to catch up to the rest of the debate, read Systemd: The Biggest Fallacies by Jude C. Nelson.
This argument is to be done question/response style using the HTML
dictionary term/definition tags. The systemd proponent goes first and the
systemd opponent goes second. Reason will be explained below.
- What's wrong with systemd?
- What's right with it? (The burden of proof lies with the person proposing
the change.)
- It has this feature which people think they want or need, faster boot times,
socket activation, boot time analysis, or other feature.
- This is important to know about software: Features are the least important
aspect of software. The most important aspects are design, implementation
(bug-freeness), documentation, compatibility, and maintainership. systemd
fails all of those. If you rely on systemd it is time to make other plans.
- What's wrong with the design of systemd?
- What is right with it? Intelligence is required to design software, not
just being clever or smart. systemd appears to have no design except to be
"The glue between
applications and the kernel" (page 5 of 85). I think this document was
written by systemd's primary programmer, Lennart Poettering. "The glue between
the applications and the kernel" means that systemd takes over Linux.
Home, Technology, Linux
Contact