Hacker News Reader

Comment by thenewnewguy | Hacker News Reader

I see we've reached the point in any systemd-related debate where someone tries to claim systemd is bloated by listing off projects that are not part of the systemd init system and just happen to share a common name and dev team.

View on HN

The problem isn't the common name, it's the tight coupling between components. This is antithetical to the Linux philosophy, which emphasizes modularity, interchangeability, and configurability.

That is not the Linux philosophy. According to a usenet post made by Linus Torvalds in 1996 [0], the Linux philosophy is "Do it yourself."

[0] https://groups.google.com/groups?&selm=Pine.LNX.3.91.9610161...

When you wish to cite Usenet articles, please cite the message ID.

(I also think that those who use mailing lists should just use NNTP instead, possibly having NNTP as an alternative interface to the same data.)

In this case, the Message-ID seems to be <Pine.LNX.3.91.961016155929.27735D-100000@linux.cs.Helsinki.FI>.

I assume you mean the UNIX philosophy? Regardless, almost all of said components are optional, systemd doesn't in any way force you to use them. As a matter of fact many of the components you listed have alternatives that are more popular than the systemd tool (NetworkManager over systemd-networkd, GRUB2 over systemd-boot, etc).

Most yes, not all of them. For example logind. Is there any reason why this should be coupled with systemd? Apparently not as has been proved by elogind project.

Reply on news.ycombinator.com