Linux 6.0 Released With Many Intel & AMD Driver Additions, IO_uring Keeps Advancing

[coder]

am I the only one that sees the trophy as shaped like a Wheel? Do I need glasses? I mean... it's a Wheel... as trophy for an Invention award... I can't even! LOL!

There's something really strange about the image Michael used, because it looks like you should be able to see his hand through the hole, but you can't!

Well, I just did an image search and indeed it's simply a hollow circle with a sphere resting inside. Not sure what that's supposed to symbolize, but oh well. I've seen some clever glass awards, over the years, that use refraction to good effect.

...which reminds me of an anecdote I once heard about the caustic refraction from a glass trophy leaving scorch marks on someone's desk, when left in bright sunlight. So, maybe it's better that Linus have a safer, more boring trophy.

[marlock]

coder if you check out Linus quotes from when he's about to decide or has recently decided on a major version change, or when someone talks about it to him and around him, the replies strongly hint at him knowing full well it's an absolutely arbitrary choice. He used to make fun and keep folks on their toes by taunting a change in pattern (iirc that was around 3.x >> 4.0, which is when I started reading more news about the kernel).

So yeah, it totally could change into a firefox style versioning and it totally wouldn't cause any loss in useful information... but I bet Linus has totally flewn over similar concerns exactly because it's arbitrary. Personally I just enjoy seeing him (a top notch software engineer) mock about needing his fingers to keep a count, so I hope it keeps working as it is

Onto more interesting bits of the news, am I the only one that sees the trophy as shaped like a Wheel? Do I need glasses? I mean... it's a Wheel... as trophy for an Invention award... I can't even! LOL!

[coder]

sdack last post reported, this was waaaay off-scale and very much uncalled for, especially after the last response from coder being totally unpretentious

Thanks. I didn't see it, probably for the best.

coder, on the Linux Kernel, major and minor versions have the same meaning, but Linux Kernel 6.0 looks better than Firefox 125...

Well, let's try mapping this to Linux. From what Linus said, the scheme is essentially base-20. So, if we just switch to base-10 and drop the decimal point, then you're right - we'd end up at somewhere around where Firefox is at.

I know it's a shift in what we're used to, but I don't see sequential release numbering as particularly more or less friendly than artificial X.Y numbering. Just my opinion.

Stallman provided the GPL tool but Linus wielded it as a true master

GCC and Linux are the only two examples that come to mind, where GPL exerted real power (i.e. coercing people to upstream patches, even when they're really rather not). That's not to say it hasn't worked for other projects, but I think those two examples benefited enough by the contributions they received that it helped sustain and entrench them. It's telling to look at just how much time & effort has had to go into LLVM, just to it could keep pace with GCC (and perhaps slightly overtake it).

[marlock]

sdack last post reported, this was waaaay off-scale and very much uncalled for, especially after the last response from coder being totally unpretentious

coder, on the Linux Kernel, major and minor versions have the same meaning, but Linux Kernel 6.0 looks better than Firefox 125... Linus usually says he can't count to a higher number than he has fingers and toes, and this refers to the minor version not going over x.20... as previosly mentioned, there is not much point in drawing the line between major and minor over API breakages... there would either be only major versions (counting kernel-land breakages, which happen all the time) or only minor versions (counting only userland breakages, which nearly never happen).

IMHO the "Opensource Godfather" moniker is well-earned, after all he did flip the infamous finger at Nvidia and now they started opensourcing their drivers... maybe that's even what motivated Intel as an office joke that became a serious proposal, LOL! Stallman provided the GPL tool but Linus wielded it as a true master

[coder]

Semantic is what you make it to mean. It may just not mean what you want it to be.

I use the most common definition of the term:

https://semver.org/

If you need it to be all uniform and abiding then you should visit China. You will love it there.

The principle of systems design hinges on common definitions and semantics. That's all we're talking about, here. If nobody agrees on what anything means, you cannot have APIs, you cannot have specifications, you therefore have no modularity and cannot scale. That's not a world I want to live in. Therefore, it's reasonable and sensible to discuss and debate semantics and try to form consensus around the most sensible interpretations.

As for Linux' freedom to use whatever version scheme it wants, I would defend that even if I think it's nonsense. We can debate what's most logical and sensible, but nobody is making anyone do anything.

There's a certain irony in using the specter of a totalitarian government to shut down a discussion about something completely unrelated.

[birdie]

I believe Intel is merely trying to give Linux the recognition it deserves, which has been long overdue. The interchange is staged and it might be what feels awkward. So is Linus being called the "godfather of open source", which he is not. It is Richard Stallman. I found it rather ignorant to just skip over the GNU project and to cancel Stallman out. And why only now? Intel has been supporting Linux for a while now and quite successfully. The only remarkable innovation in recent times between Linux and Intel is perhaps io_uring, which revolutionised SSD technology and in particular Intel's Optane series. Anyhow, I guess this is just what happens when you stage something like this.

I will however not celebrate before Microsoft and Windows is dead.

I'm not sure [the] Linux [kernel] needs recognition. Over 95% of Internet servers run Linux, pretty much all the supercomputers run Linux, it's only the desktop where Linux is nowhere to be seen but this conference is not catered to the average Joe, so they'd never know.

[coder]

Why is that? Because major/minor implies that a major version is backwards incompatible (or smth like this) and Linux ignores this convention?

Yeah, the kernel doesn't abide by semantic versioning in either its internal or its external APIs.

Linux famously doesn't have a stable internal API, which creates an incentive for drivers to get upstreamed (or disincentive to get maintained out-of-tree, depending on how you look at it). I think this is an underappreciated reason behind its success.

And the userspace API has breaking changes whenever it really needs to, and not otherwise.

So, there's no good argument why it should use Major.Minor. Just habit, I suppose.

[NathanG]

For the longest time, people used to talk about "Wintel" as a short-hand for Microsoft + Intel. I don't even know the last time I've heard that term.

Only just occurred to me I hadn't heard the term in ages, looked it up on Google books ngrams viewer. Looks like it steadily fell from 2000 to 2010, I never noticed!

[cl333r]

the concept of Major.Minor versioning is really out of place. I wish they'd switch to a sequential numbering scheme.

Why is that? Because major/minor implies that a major version is backwards incompatible (or smth like this) and Linux ignores this convention?

[coder]

Bad genetics; I saw his dad at a parent/teacher conference and he was 35-45 and looked 55-65.

Don't forget diet & lifestyle. Those are also heritable traits, to some degree.

15 years ago, I dropped like 10-15 pounds, just by improving my diet. It happened over a period of a couple years, where I was just trying to improve my focus and productivity by eating better. I didn't have an explicit goal of weight-loss, nor did my activity level change.

Then, I began regular exercise and started to learn just how much of my weight was excess.