Distro wars and that certain je-ne-sais-quoi toxicity of many uber-nerd environments, where elitism and holier-than-thou reign supreme.

I'm glad that a certain degree of pragmatism (e.g. W10's October deadline) is finally being embraced in our beloved ecosystem.

View this comment - View article - View full comments

Relative numbers are all fine and dandy.

But it's in the absolute values I find most satisfaction.

The growth from 0.75% (2018) to 2.57% (2025) should be considered also compounded with the growth from 90M MAU (2018) to 150M (2024 - can't find data for 2025).

Graphing the absolute numbers would show an even steeper growth curve.

View this comment - View article - View full comments

@sudoers

The fun part is that they learned their lesson and didn't try that again

So you think this isn't just delayed on Canonical's part, but they will support 32bit forever?

Ok.

View this comment - View article - View full comments

@sundoer

Fun (!) part: Ubuntu already tried to get rid of this more than five years ago (19.10 and 20.04 LTS) and received the same flak.

https://ubuntu.com/blog/statement-on-32-bit-i386-packages-for-ubuntu-19-10-and-20-04-lts [External Link]

It's just a matter of time before all distros will go this way, hopefully coordinating a cutoff release point together.

We're just delaying the inevitable, until the point it will be really hard to manage the fallout.

View this comment - View article - View full comments

I'm quite disappointed that he bailed out on this.

Honestly expected, given the negative press coverage and toxic, rabid comments from the "reee mu games" crowd, but nonetheless rather unimpressed about the outcome.

Instead of a strictly technical debate about the merits, the fallout consequences, and how to deal with them from a practical, actionable, "what can we do about it?" point of view, it turned out into a us-vs-them fistfight that helped nobody.

The issue remains, and of course will become worse every passing month, when more upstream drops support for legacy, and more people will expect support for obsolete, unmaintaned applications (for who commented about the kernel being 30 years old: UNMAINTAINED is the keyword here, and I don't think I need to explain why) which will become all the more burdensome the more time it passes.

Mark my words. This mess will rear its ugly head again, and more frequently every passing year.

A pyhrric victory for who cannot accept the hard fact that an early 2000 unmaintaned game belongs to a containerized or emulated environment, not to daily maintainer burden.

View this comment - View article - View full comments

@Caldathras & @ScottCarammell

Since the news of this being backtracked, I'd welcome you to continue the discussion in the new article's comment thread.

And never forget: we're all on the same side.

View this comment - View article - View full comments

@Caldathras

Besides, how do you know that this issue is not being influenced by IBM or RedHat in some way?

Because the engineer proposing the change says so in the very discussion thread this article is about?

https://discussion.fedoraproject.org/t/f44-change-proposal-drop-i686-support-system-wide/156324/145 [External Link]

No offense, but on a matter about volunteer Fedora maintainers, I'd tend to trust more a volunteer Fedora maintainer than a Joe H Random on GamingOnLinux. And so should you.

Our society has been devaluing the idea of maintenance against the "convenience" of a throwaway habit that only benefits the for-profit sector -- at the expense of our society, culture and environment.

I feel the "right to repair" sentiment here, and I broadly agree with it.

But a line needs to be drawn.

If you're against the Temu's and the Shein's of this crazy world, I hear you. If you're against planned obsolescence which would require you to buy a new iPhone every time the old one's warranty expires, I hear you. If you're against mad races to shareholder quarterly profits at the expense of everything else, I hear you.

But the technology we're discussing to get rid of (again and again) is so old that calling it "throwaway habit" after more than twenty years that the "next model" has been out is frankly embarassing.

Just for scoping what we're talking about: can you provide a list of games that you know for a fact they wouldn't work on WOW64 Proton, or without i386 compat .so's, and the date they were published?

Just to have an idea of the size of the problem.

View this comment - View article - View full comments

The more I read about this, the deeper the rabbit hole goes, and the messier it looks. And it gets worse every day.

Upstream 32bit versions of applications/libraries started to be unsupported, untested and/or unused by upstream since some time now. "Doing nothing" and maintaining the status quo is worse than difficult: it's dangerous.

Fedora volunteers cannot keep providing a better product downstream than they have access to upstream. As a side note: can we please stop blabbering about corporate greed and for-profit progressism? These are guys helping for free in their free time. It's not IBM, it's not RedHat, it's a different thing. The corporate world abandoned 32bit AGES ago (and support it only in front of quite substantial payments).

Resources are scarce. Whoever is on the roof shouting that we must keep things running like twenty years ago should come down and start helping contributors. Or become one.

Associated projects, like RPMFusion, have the same resource starvation (people/time/interest) as Fedora maintainers, and they wouldn't pick up the slack. After all, it's rather unrealistic to offload one volunteer to overburden another.

Fedora is too small for Valve to care? Maybe. Or maybe this could be the last straw to push Gaben (always be praised His Name) to make Steam 64bit only and including in itself some "soft virtualization" in the form of 32bit runtime environments for abandonware 32bit native games. This wouldn't help extra-Steam native games. But at some point, we should stop complaining we can't run Amiga games without an emulator.

Other important subsystems, like Gamescope and OBS, are equally critical for a modern, gaming-friendly environment. We cannot dictate what others do with their free time, but asking kindly is always an option. And the best option is start contributing actively to the projects to help them get out of 32bit technical debt. That's the beauty of Open Source, after all: not backward compatibility, but technical excellence.

This need to happen, sooner or later. It's a matter of when, not if. It's too early now? Most likely yes. So, when it's the right time? Ubuntu tried it in 2019-2020 and it backfired horribly. But even five years ago, the rationale about letting 32bit die was strong and pressing. We're half a decade later debating exactly the same stuff, with exactly the same problems. https://canonical.com/blog/statement-on-32-bit-i386-packages-for-ubuntu-19-10-and-20-04-lts [External Link]

Do I have a solution? Of course not. Will I miss some late 90's games unless there's an user-friendly way to run them? Probably. But in all honesty, I have no heart to ask Fedora maintainers (or any other distro maintaner, for that matter) to keep doing this maintenance work forever because Reeee mu games.

View this comment - View article - View full comments

Valve should go full 64bit, and drop/integrate the 32bit requirement/legacy.

And, tbh, so should WINE.

AMD64 is 22 years old, x86-64 is 21 years old.

After two decades, you can safely assume (the mother of fuckups be damned) that most of your userbase uses the "new" technology.

View this comment - View article - View full comments

For a short, exciting moment after reading the title, I thought that Valve was letting players use Proton outside of Steam.

That would be something I'd put next to the adverb finally.

View this comment - View article - View full comments