87
Linux doesn't use semantic versioning, right? So this is just an arbitrary number?
It's always backwards compatible so major versions don't mean anything.
Yes.
I absolutely love semver, but it can lead to absurdly high version numbers (a package that I maintain at work is now at something like 3.1.125). It contains mostly config for other things, so... This is somewhat expected.
I still think it's better than just naming every version after the year of its release (like 2026.1) or random arbitrary numbers, though
but it can lead to absurdly high version numbers
I have a really hard time seeing this as a problem. Why is 125 an "absurd" version number? You've presumably done 125 patch versions since the last minor version, so doesn't it just make sense?
It's just a number, it doesn't matter if it's "high".
I mean "absurdly high" in the context of the thread where Linus says at about 20 it's time for a new version.
But in the sense of semver, that's a completely reasonable version number, assuming you had many small fixes/additions.
What is he going to do after 19.19?
LinuxTwo 1.0 Should buy a few more decades before LinuxTwo gets to 19.19 and LinuxThree happens. Or perhaps we expand version numbers to use letters first, meaning you have 36*28 versions before 20?
There goes my polymarket gains, goddamnit.
this post was submitted on 08 Feb 2026
87 points (100.0% liked)
Opensource
5531 readers
76 users here now
A community for discussion about open source software! Ask questions, share knowledge, share news, or post interesting stuff related to it!
⠀
founded 2 years ago
MODERATORS