The Delusions of Debian

Some of these issues are well known in Debian and being addressed; https://salsa.debian.org/debian/grow-your-ideas/-/issues

If your marriage is ending over a toilet paper roll... it's not actually about the toilet paper roll.

It's the same deal with Debian and systemd. It's not the init system. It's the thing it represents. Having to either adopt systemd or run GNOME in an unsupported configuration seems like a clear-cut choice (which is why systemd is now the Debian default), but having a major upstream force you to significantly rearchitect the distribution seems like a pretty significant loss of control. This is at the same time as other upstreams have gradually ripped control out of the hands of Debian developers in other ways: Firefox, librsvg, and python-cryptography adopting Rust suddenly made it a lot harder to support a bunch of niche CPU architectures. Speaking of Rust, and also Go, they use static linking and have their own library package managers, which both makes it harder for Debian to package and simultaneously easier for users to install a binary directly provided by upstream. And languages that don't have static linking support can always use Docker.

Honesty, I wouldn't want to be a Debian developer right now. Red Hat has tried to reinvent themselves in a world of containers, but since Debian is a volunteer organization and not a corporation, it's a lot harder to pivot (attempting to pivot would probably cause all their existing volunteers to quit, while failing to attract new ones). What sort of future does Debian even have, if they have no decision-making power over the core OS, and all the applications route around them?

I may be naive, but don't we have automation handling a lot of the maintenance of packages and scanning for vulnerabilities these days? Of the 90k packages mentioned there's probably only a handful that have known CVE's and when they come up they probably just bump the version number in their CICD if a patch is available from the primary code maintainer and if not it's reasonable 32 people could manage that many packages depending on their CICD structure. Asking more than making an assertion btw.

Debian project has been a blessing to me personally.

For a daily driver always buy a ThinkPad, a year old model when new stable comes out, and use it as long as possible.

For servers, run the latest stable with minimal packages from main and occasionally few from backports.

Build from source/ download from upstream where Debian packages are available (specially browsers).

Reliable kernel, coreutils, vim and gcc from Debian stable, git + ooenssh from backports, tmux built from source, golang and "go get xyz" is mostly what I need. (I understand this can be too minimalist for many, but works for me).

I have a hobby website using fossil, running on 8year old SBC running bullseye.

At work, almost all bullseye docker images on minimal Bullseye running Dockerd. PostgreSQL, Redis, and few more central piece of software work fine with Debian.

Low maintenance, reliable systems that I seek, and Debian has never failed me.

Ahh and yes, when I need bleeding age, Arch is what I go to, but won't be running a daily driver machine or server on it. It's too fast for me to keep pace with.

I can safely say I have more peace of mind with free Debian than any paid OS I used (taken together with application ecosystem). All systems have their own pros and cons. I just picked what worked best for me, I am thankful for that and wish Debian project never runs out of great people.

I would downvote you because:

1. "GNOME in unsupported configuration" is, with due respect, FUD.

2. There was no dichotomous choice to make. Debian could have let you decide whether or not to install systemd. Instead, the project decided to essentially force the use of systemd, necessitating a fork.

3. "suddenly made it a lot harder to support a bunch of niche CPU architectures" 4. Language-related package manager do make things difficult for an OS distribution, at least somewhat - but that's not particular to Debian.

However - I actually agree with your main point. I'm sure Debian maintainers/developers have it hard.

I really wish Debian would:

1. Admit the Devuan people were right and re-merge the distributions.

2. Recruit, offering a self-training track for aspiring package maintainers.

3. Modernize some of their tooling (as people seem to be complaining about that)

4. Fundraise effectively, to finance the above.