Hacker News new | past | comments | ask | show | jobs | submit login
Linux 3.0-rc1 is here (kernel.org)
246 points by angusgr on May 30, 2011 | hide | past | web | favorite | 77 comments

For those wondering what's new:

“So what are the big changes?

NOTHING. Absolutely nothing. Sure, we have the usual two thirds driver changes, and a lot of random fixes, but the point is that 3.0 is just about renumbering, we are very much not doing a KDE-4 or a Gnome-3 here. No breakage, no special scary new features, nothing at all like that. We've been doing time-based releases for many years now, this is in no way about features. If you want an excuse for the renumbering, you really should look at the time-based one ("20 years") instead.”

by Linus Torvalds

source: ttp://permalink.gmane.org/gmane.linux.kernel/1147415

I didn't know you were quoting Linus, you should make it more obvious.

It's funny and upsetting how I get more upvotes for comments like these than for things I actually put some thought into.

I've noticed the same. Most of my posts that have been massively upvoted were suggestions or small corrections. The posts in which I take time to explain something non-trivial generally end up with 2-3 votes at most.

Amount of karma is not a measure of interestingness of the contribution, it's used mainly as an 'agree' button.

This is almost certainly caused by the decision to hide comment score. If people could see that this kind of comment (f.ex. a small correction) already had its fair share of upvotes they would leave it at that, but as it is now everybody who sees it upvotes it.

What's the difference? Are we running out of upvotes?

1. If, now or in the future, the HN comment-ordering algorithm pays attention to children's scores, then you get the following pathology: A says something slightly wrong, B makes a small correction, B's comment gets massively upvoted because everyone upvotes it, so that subthread gets treated as highly-scored even though there's nothing terribly insightful or interesting about it. (You could, but probably wouldn't, get the following even worse follow-on pathology: people start putting mistakes into their comments so that this will happen and they'll get extra visibility.)

2. People get more karma for posting minor corrections than for posting substantive comments. That's arguably unfair, and again it has a possible follow-on pathology: making nitpicking more popular and carefully thought-out insightful comments less popular.

I've noticed no such change in my own comment scores with or without scores displayed. I think it's a problem intrinsic to comment voting systems. I see the same problem at all such websites I've frequented.

You've provided a valuable service and saved people time. The first question anyone will have about Kernel 3.0 is "What's new/different in it from 2.6.39?". By summarizing/quoting you let people skip skimming the article, or a Google search, and saved them time.

Not really. It's a sign of an unwelcome demographic change.

added the name :)

That's a start, but people only read that name after they've read the quote. Helpful conventions for quotes are:

"quotes" - Linus Torvalds

quotes - Linus Torvalds

I think you'll agree with me when I say that people only read the name after they've read the quote in your examples too.

When I see something in "quotes" or quotes, I usually look to the end to see the attribution before reading it.

It would break many of the programs which rely on 2.6.x naming scheme. Linus always was on the side of not breaking things if they do work. It seems that he changed his views for this particular case.

He's also a huge proponent of "Don't be stupid". I think that applies to developers who did what you mentioned.

I don't know if that's fair. I mean, if you have a program that needs to parse out the Linux kernel version then you could very plausibly have assumed the format vX.X.X[-something|.X], given it's been that way consistently for 15 years (since 2.1.0 in '96). Even the kernel's own scripts do that, as Linus mentions in the email. Now it's going to have format vX.X for the first time since 2.0.

I can't actually think of the use case for needing to parse the complete version string outside the kernel, but it doesn't sound stupid to me that you'd assume a canonical format that's been there for a decade and a half. Apparently wrong, and decidedly unimportant & bikesheddy, but not necessarily stupid.

Kia specifically mentioned '2.6.x' naming. Obviously expecting three parts is AOK, expecting the kernel to remain at 2.6 forever is not.

Well he said the '2.6.x naming scheme', which I took to mean the 3 part versioning scheme (as used in 2.6 kernel versions.)

I agree with you that assuming the kernel would always be 2.6.x is not OK, I'd just assumed noone would be that shortsighted and stupid. :)

"Linux 3.0-rc1... except there are various scripts that really know that there are three numbers, so it calls itself "3.0.0-rc1".

Hopefully by the time the final 3.0 is out, we'll have that extra zero all figured out."

Source: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6...

Ultimately, you get the worst you're willing to put up with. When the Windows dev team bends over backwards to keep buggy applications working, they signal that those bugs are acceptable and turn them into the accepted way to develop Windows application software.

When Torvalds bumps the major version number with little warning, he signals that anything which depends on it always being '2.x.x' or, worse, '2.6.x', is not acceptable and will not be tolerated.

I've lived in both worlds. I prefer Linux. I'm not alone.

Linus' email is more interesting than the commit: http://permalink.gmane.org/gmane.linux.kernel/1147415

Although that's not very interesting either. Linus: "What's new?" Linus: "Nothing."

Yeah, nice. "Major version? Minor version? Revision? What're those? I like round numbers."

Why not just go to a basic increment (Linux 3, 4, 5, 6, ...), or a date-based (Linux 2011! Now with new ribbon interface!) numbering instead?

...eh. I dunno why this bothers me so much.

It shouldn't. Open source version numbers greater than 1.0 dont mean a whole lot.

Chrome, the browser, around version 8 (give or take). How much sense does that make?

Yeah, I know. At one time though, version numbering was used as a signal to other developers that relied on your software. Changing the major version number indicated a large change -- either API, or architectural, or both -- that was expected to break software that relied on it. So then, as an end-user, it was easy to keep track of which pieces were likely to be compatible with which other pieces; if your add-on or what-have-you worked with version 2.1, then it would work with 2.1.1, and probably work with 2.2, but probably not work with 3.0.

I guess people got bored with the sensibility of that, or something.

Linux has kept its main public interfaces stable for all of its several hundred post-1.0 releases, so using that scheme would mean Linux would have to use rather odd-looking version numbers like 1.632.5

Chrome is on version #11 in stable. #12 and #13 are in beta and dev respectively.


8? I'm running Chrome 11!

You are correct. I am on 11.0.696.71.

I was wrong when I said 8.

Isn't the point of this to start correctly using the minor version numbers again, to get back to a good versioning system? And since it's a different scheme it makes a lot of sense to increment the major version to separate it.

Because the number can still mean something. Does anyone have any major arguments against or issues with semantic versioning? http://semver.org/

Despite all criticism and cynicism all over the Linux communities about the numbering and all that stuff, for me, every release of the Kernel (as well as any other major / dominant open source platform / project) is a reason for celebration.

It simply means, openness and freedom won the software/internet game. The fact there are people out there, spend the best of their times, contributing code and manuals docs, debugging and filing bugs, etc. etc. Is a sign that RMS and alike were not _"a bunch of hippies who likes to code for free"_ or even worst, a bunch of communists as some used to say at the beginning of Linux breakthrough.

From my own personal experience, it also means, the more open you will be, the more open software you will rely on, the more money you will make by the end of the day.

Thank you Linus, and all kernel contributors, for the great tools and platform you provided us for FREE!

> It simply means, openness and freedom won the software/internet game.

Except... it didn't. The far majority of programmers still write closed source non-free code. The far majority of software is still closed source and non-free. The far majority of humans browse non-free websites on non-free web browsers running on non-free operating systems.

Friend, I am afraid you have missed my point. What I meant was that open source have change the software world. Nothing is the same anymore, and in my opinion, the "cloud-revolution" we are observing today has a lot to do with open/free software.

It might be true that most browsers are closed source running on top of closed source operating systems, but that is just one half of the picture. the other half tells a story of most web traffic, especially on most popular ones, is generated by open platform.

Vast majority of programmers are writing closed source. yet they do not affect the market as those who write open source.

The EFF disagrees with you about the cloud. They believe the cloud is a serious loss of freedom, partially because cloud-code is hidden (even modified GPL code) and partially because people are giving away their privacy.

The cloud has been great for the business of writing software, but not because it is more free or open.

Oh, I don't think he was saying that wasn't the case. I believe what he meant (perhaps I am projecting here) that because it continues to exist, and exist successfully, the idea of free software wins.

The mere continued existence of free software/open source validates the ideology behind it.

As a libertarian/dyed in the wool capitalist, I love the success myself.

I don't see how existence equates to winning. A winner implies a loser.

I doubt RMS considers the status quo a success. The free software movement that I know had a much loftier concept of "win" than mere existence.

I doubt any of us could accurately predict, understand, fully appreciate that which is RMS The Great.

I mean, http://www.youtube.com/watch?v=S76pHIYx3ik.

One the one hand, the "benevolent dictator" model has shown itself to work pretty well for Linux, compared to some major failings of the "committee" model on other projects. On the other hand, does anyone think the fact that "major" changes can be done by one person who is "just going all alpha-male" will potentially spook people? I mean, I doubt too many people here on hn have a problem with Linus' personality. I imagine most of us appreciate his sense of humor and respect him as a penultimate hacker... I'm just thinking about starched-shirt types who call the shots in big businesses -- can you imagine being an IT manager and trying to explain to your boss that after 15 years there's a new major version of the Linux kernel because "well the head guy got tired of calling it version 2."

Edit: Penultimate apparently doesn't mean what I thought it meant, so, what I meant was "...respect him as a very good hacker...".

I'm not entirely sure why you would... My boss doesn't know what kernel version we run, all he cares is that it works.

You think linus is the second best hacker?

It appears my understanding of that word was faulty. I always thought it meant something along the lines of first-rate, I didn't realize it had such a specific meaning. TIL, I suppose.

No worries, the majority of people make that mistake.


Adjective: Last but one in a series of things; second last: "the penultimate chapter of the book".

To quote Murray Walker, an F1 commentator, "they're on the penultimate second from last lap but one!"

It's free, it's open source. If you've got a problem with how Linus does things, you can always fork it.

Lots of commercial software has version numbers assigned for marketing reasons and largely independent of technical changes.

See Slackware's jump from version 4 to version 7, or more recently, Mozilla's reconfiguration of its versioning system to make the numbers more closely match Chrome and IE version numbers (where v11 and v9 are the current stable releases respectively -- Mozilla is on a measly v4). This doesn't just occur in commercial software.

For example, Windows 7 is internally Windows NT 6.2.

I don't think the 3.0 release is "just going all alpha-male" - I think it reflects the maturity of the linux kernel, and could have done some years ago with good conscience too.

That said, his Google Tech Talk on git, which turned into a rant against SVN/CVS users did lower my respect for his personal skills.

My first reaction to that talk was "wtf…". Then I tried git. And then I started to think that he might been right.

"Version numbers? We can increment them!"


  -VERSION = 2
  -SUBLEVEL = 39
  -NAME = Flesh-Eating Bats with Fangs
  +VERSION = 3
  +NAME = Sneaky Weasel


Some interesting version names in there. I find 2.6.22-rc3 particularly intruiging...

My favorite part is the commit message on the tag: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6...

"Version numbers? We can increment them!" -- Linus Torvalds

A summary of the changes can be found here: https://lkml.org/lkml/2011/5/29/204

So, when will Linus be released from mental institution? [0]


I don't know why, but I always get a good chuckle or two from reading or watching Linus' materials.

Its name was Sneaky Weasel.

Wow, he said he might go to 2.8. This is madness.

No, this is Linux!


Seems a bit strange to me. A major version increment is supposed to announce "this is significantly different and possibly incompatible".

But now it doesn't mean that. It means 20 years.

There are extremely good reasons not to increment the major version. For instance tools which take the conservative approach of checking for 2.6 will all be broken. Those tools correctly elected not to risk wrong behavior resulting from a major version change.

When you add together the time it takes to compensate for this change for all maintainers everywhere (tools developers, distro developers, everyone), surely it will run into the hundreds of hours, perhaps thousands. All for, functionally, nothing. And sorry for calling you Shirley.

Seems to be an objectively bad move. Oh well. This needless-on-purpose change will cause resentment, but after time it will dissipate.

checking for 2.6 no longer means much, 2.6.0 and a current 2.6.39 have so many changes that just checking for 2.6 is not enough anymore. that was one of the reasons for the new version.

You're saying there are tools which opt to break entirely, instead of work with the chance of bugs, if your kernel version number got too high?

What tools actually do this?

In theory, the Linux kernel has been binary stable since version 1.0 -- if it worked on an old version, it will work on a new one.

In practice, due to the vast number of configuration options, it's not entirely binary stable even across different people compiling the same code. With work, you can probably make things compatible, but it doesn't come for free.

I know they invented git, but that website sure is fugly. Time to switch to github, perhaps :-).

Why do you care what it looks like? It's whether it works well that is really important for a tool like this (that pretty much only the devs are going to use).

It's not actually great on that front either though, but the point still stands. I prefer cgit[1] myself (as it's a lot more usable). It's quite mature and used by Freedesktop [2] and Gnome [3]. If people are really worried by how it looks you can easily add your own header, footer, and alter the CSS.

I hadn't seen Gitalist (that marchdown posted) though. Looks pretty good but I wasn't that impressed by the usability of the demo though. But that will probably improve in time.

1. http://hjemli.net/git/cgit/ 2. http://cgit.freedesktop.org/ 3. http://git.gnome.org/browse/

There is a nice open-source (which github is not) web front-end for git repos, http://www.gitalist.com/, and would you believe it, even they use git-web on their website.

Unfortunately, http://example.gitalist.com/ doesn't load any content whatsoever in my browser (Google Chrome with JavaScript disabled/whitelisted). Just a white page. And I have a hunch browsing with JavaScript off is more common for kernel devs than, say, GitHub users.

Well, Chrome must be broken then, as it's a validated HTML 5 page. It works with FF with noscript (with the domain blocked). Frankly, I have a hunch that people who care about javascript security don't use Chrome.

No chrome until noscript.

I didn't disable JavaScript for security reasons, but for faster web pages on my slow EEE laptop.

It may be fugly, but it's also functional

It probably works in text based browsers too

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact