The Problems Debian Is Facing In 2020

[waxhead]

As so many have pointed out already , stuff like BLM is completely irrelevant for Debian. What matters for Debian is to communicate technical excellence properly. If the distro's desktop and/or website is pretty enough is utterly nonsense and completely the wrong focus. It is compact and functional - no need to change that.

IMHO Debian would benefit most if the testing repo was either renamed to rolling or if there was a new repo in addition to stable, for example something like "rolling" or "prestable" or "nextstable" where testing packages are migrated just like packages migrate to testing now. If for nothing else, a new name would help people understand that testing is a lot more stable than what "testing" may seem.

Also a repo where you could pull slightly experimental debian kernels e.g. a mix between testing and latest vanilla kernel would be useful. Documentation would also be more useful if they could group oldstable information somewhere special. The bugreport tool could also be improved it it kept track of what bugs you actually reported so you could revisit them easily.

Luckily there will be a new project leader in a little while

[pininety]

• Stability for Debian means "there might be bugs, but there won't be _new_ bugs".

It also means "it might not work at all but at least does that reliably". Especially troublesome if you have a package that depends on an external service that might get updated making the unupdated package unfunctional. Like the dyndns1 protocol deprecation.
It also means much more work for upstream because they are hit with a ton of support requests for very old software because many users just go directly to upstream to report instead of the horrible Debian issue tracker.

At some point, I just started switching everything over to Arch Linux, had less issues with that on multiple machines that just having debian on a signle router for 2 years. Arch just worked, the few times I had an issue after an update, I normally just had to wait for a few hours as the bug was always already reported and worked upon and the mirror servers just hadn't picked it up while any issue I had with debian was "Yeah, that was fixed upstream x years ago but we refuse to update, just run unstable!"
which kind of makes the hole point about Debian being so stable mood.

[creoflux]

The idea that people are rushing to Debian project to get their take on a issue not related to their software is a little bit odd. I somewhat cringe at the idea that they formulate positions on these things. The world is a lot larger than the US and ignoring any issue would open you to criticism.

Did they release a statement about wildfires in Australia? What about their position on Uyghurs in China? Women's rights in Saudi? Boko Haram? AfCFTA?

[birdie]

It's not too hard, you can pull a newer kernel from backports. It's true, you have to know what you're doing, but nobody claims that Debian is a distro for newcomers. You could say the same for Arch, or Gentoo.

The Problem of Debian is well presented in this comment section. People (that don't really use it) blame it for old software (which isn't fair since there is Testing/Sid with fancy new software), and rage about a single BLM point from a 40min presentation. Those people don't get that Debian isn't just a distro build by a couple of people (e. g. Arch) or a company (e. g. Red Hat), but rather by a community with roughly 1000 developers and >2000 additional uploaders (probably more), which are all bound by the same _social_ contract (yes that is a real thing in Debian).

Mixing packages from testing/unstable and Debian stable? What could go wrong? And how easy can it be achieved? And what about installing Debian on your new shiny laptop/PC when it comes with an ancient kernel out of the box? How would you proceed to boot with a new kernel right away, huh? Maybe people should also learn to rebuild installation media? Yeah, I last used Debian in 2001. I quickly understood back then that it was a distro for servers. Nothing has really changed for the past two decades.

[Danielsan]

I have been Debian Testing + Pinning toward Sid forever the only point is during the freezing time, by the way Debian can do a lot thing to improve like:

1. Improving the release cycles like: rolling + snapshots + lts snapshots, just for instance.
2. Improving the package creation so people would be more involved in working upstream directly on Debian.
3. Better bug reporting and better attitude toward who report the bug.
4. Being init agnostic again or working together with other derivatives like Devuan.
5. Focus on general availability over all the platform, innovation and best practices.
6. Being again independent.

[skeevy420]

Release cycle is a toughie. There's always going to be some lag. If Debian released packages willy-nilly that'd be chaos. So some kind of a happy medium needs to be struck. Debian's reputation is that it is a bit behind, but solid. Which I am OK with. My priorities line up well with that. I'm on the platform because it is solid. If I didn't care about that there's plenty of other paths I could choose.

Can't say that I disagree with release cycles being tough. All you have to do is look at random releases from the past to see examples of crappy package choices because some project's stable release came 3 days after a version freeze or the distribution release was held up a bit to accommodate a new package release. That's why I'm against release cycles and prefer rolling release distributions. But I'm also thing from the point of view of a home user. I'd have a completely different point of view for a scenario where I'd be managing multiple systems.

The problem with Debian is that it isn't 2005 anymore. The world moves faster than their release cycles and guidelines can keep up with, at least in my experiences as a desktop/GUI user. Even CentOS, which IMHO is their biggest competitor since both aim for servers and stability, has a rolling release edition these days. Contrary to what a lot of people seem to think, Debian Testing just isn't adequate to use as a long term rolling release OS like Arch, Gentoo, Manjaro, or Tumbleweed are.

[danmcgrew]

"The Problems Debian Is Facing In 2020" are all contained in the simple statement and fact that none of the problems Debian is facing in 2020--dealing with the production of a world-class Linux distribution--are addressed by this talk and slide show.

ALL of the problems Debian is facing in 2020 are created by Debian, and ARE contained within this presentation.

[DRanged]

There's a lot wrong with the presentation....

• Most of the problems I encounter is because of an arbitrary definition of stability. The stable branch is basically just old, and although it is unlikely to fail, it doesn't offer everything people are looking for. So, you could jump to testing, but even that has relatively outdated packages, and is very unstable. Arch offers newer packages with (to my understanding) fewer paid developers, and I encounter far fewer problems due to updates.

I would disagree with very unstable. Depending on used case it has Debian testing has been rock solid for me in the last 9 years. See previous posts of me.

[mos87]

regarding the last slide... here's a list of some REALLY WONDERFUL twitter accounts to follow (for YOUR GREAT AMUSEMENT):

https://twitter.com/realShamima

https://twitter.com/Zara_Bukake_MP

https://twitter.com/X_R_Brum

https://twitter.com/Layla_MoronMP

[DRanged]

Debian should do something about there bug reports, because when I lookup a bug report I do stumble on reports from more than a decade ago while there are new versions out and defintely not showing anymore the behaviour when it was reported and not closed maybe due to laziness, incompetence or whatever.