Beta

Slashdot: News for Nerds

×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

Aging Linux Kernel Community Is Looking For Younger Participants

Soulskill posted about 9 months ago | from the fresh-meat dept.

Linux 332

Lemeowski writes "Time has been good to Linux and the kernel community, with the level of participation and volume of activity reaching unprecedented levels. But as core Linux kernel developers grow older, there's a very real concern about ensuring younger generations are getting involved. In this post, Open Access supporter Luis Ibanez shares some exciting stats about recent releases of the Linux kernel, but also warns that 'Maintaining the vitality of this large community does not happen spontaneously. On the contrary, it requires dedication and attention by community members on how to bring new contributors on board, and how to train them and integrate them alongside the well-established developers.'"

cancel ×

332 comments

Well, I'll tell you why I'm not interested.. (5, Insightful)

Anonymous Coward | about 9 months ago | (#45404629)

I'm part of one of these younger generations, and I'm honestly not interested in getting involved because I've seen how much of a raging asshole Linuz can be. He's a great maintainer, but he could be honest and give constructive criticism in less condescending ways. I'm not as experienced as he is, but that doesn't give him the right to be a complete dick in public theater.

Re:Well, I'll tell you why I'm not interested.. (4, Insightful)

Joining Yet Again (2992179) | about 9 months ago | (#45404707)

This. I've tinkered with the kernel, written device drivers, blah, but there's no way in hell I'd ever try to contribute upstream, because I know I'm not an experienced kernel hacker, and frankly I'm not sewn for the sort of macho abuse that dorks like to give each other.

There are other things I do as a hobby where I'm surrounded by people who are highly experienced, well-respected, but also excellent teachers - e.g. ham radio. There, I'm happy to do as much as I can for the community.

N.B. I'm not saying that I'd necessarily be good enough to contribute to the official kernel, merely that I wouldn't even try in that sort of environment.

Re:Well, I'll tell you why I'm not interested.. (5, Informative)

Jah-Wren Ryel (80510) | about 9 months ago | (#45404843)

there's no way in hell I'd ever try to contribute upstream, because I know I'm not an experienced kernel hacker, and frankly I'm not sewn for the sort of macho abuse that dorks like to give each other.

Sounds like a matter of perception. Linus yells at the people high up in the hierarchy because they are experienced and shouldn't be making dumb mistakes - right or wrong you aren't likely to get on the wrong end of that. As a newbie contributor any work you would do would go through a couple of levels of people vetting it for you. If you make dumb mistakes chances are the person who notices them will be a lot more gentle in pointing them out because dealing with newbies is part of the role in the hierarchy. No system is perfect, I'm sure there are some newbies who have received overly harsh responses, but that's going to be rare.

Re:Well, I'll tell you why I'm not interested.. (1)

Anonymous Coward | about 9 months ago | (#45405093)

Even if that is true, others 'old-hands' are also quite 'prickly' and fast to shout down at people. I've seen it recently first hand myself and after a few times trying to get bits of kernel fixed, quite frankly, I'd can't be bothered with being made to prove myself at every turn in the face of stark criticism often of the NIH kind.

Re:Well, I'll tell you why I'm not interested.. (5, Informative)

LordNimon (85072) | about 9 months ago | (#45405219)

I've been working on the Linux kernel for 10 years with numerous commits upstream, and I've never communicated with Linus.

Re:Well, I'll tell you why I'm not interested.. (5, Insightful)

Anonymous Coward | about 9 months ago | (#45404971)

in that sort of environment.

Well clearly you have never once 'been to' the LKML but instead built your opinion on the basis of stories-posted-on-slashdot.
Otherwise you would know that the LKML receives around 400 mails per day, the vast majority of which are polite, friendly and helpful.
Compare that with the number of posts offensive enough to make a story on /.

Re:Well, I'll tell you why I'm not interested.. (5, Interesting)

Anonymous Coward | about 9 months ago | (#45405305)

Well clearly you have never once 'been to' the LKML but instead built your opinion on the basis of stories-posted-on-slashdot.
Otherwise you would know that the LKML receives around 400 mails per day, the vast majority of which are polite, friendly and helpful.
Compare that with the number of posts offensive enough to make a story on /.

I *have* posted bugs on LKML, and gotten responses. I have interacted with at least two high level developers, as well as Mr. Torvalds. The one time I got a reply (Len Brown, INTEL senior systems engineer) plus asked to download software to dump the rom from hardware, followed by an analysis and a change to the kernel (which I then applied, re-compiled and tested with reports. About 200,000 people were affected by that bug (and I got email from around the world). I've also gotten several very polite replies from Alan Cox and a few others. The trick is that you have to 1) know about computers, be able to describe the bug fully, what you have tried to fix the bug, and how it affects things. 2) be able to reply to questions / do more testing 3) re-compile a kernel with a fix and see if it fixes the bug. Most people can't do #3.

Re:Well, I'll tell you why I'm not interested.. (1)

Anonymous Coward | about 9 months ago | (#45404749)

It's a common trait really. Linus has been doing this for over 2decades now and I'm sure he's raised his fair share of people in that time. Not to mention he's also a father etc. At some point you just get tired of taking people under your wing all the time and you just expect them to be capable of working at your level.

People always change between up bringer and not; so be honest with yourself if you're put in his shoes. Given the same job, would you rather work with a diverse set of people less capable than yourself, or would you rather work with a diverse set of people as capable as yourself?

Re:Well, I'll tell you why I'm not interested.. (5, Funny)

DeBaas (470886) | about 9 months ago | (#45404769)

I'm part of one of these younger generations, and I'm honestly not interested in getting involved because I've seen how much of a raging asshole Linuz can be. He's a great maintainer, but he could be honest and give constructive criticism in less condescending ways. I'm not as experienced as he is, but that doesn't give him the right to be a complete dick in public theater.

You've managed to asses that he is 'a raging asshole', but now how to properly spell his name?

OP responding.. (3, Funny)

Anonymous Coward | about 9 months ago | (#45404805)

*assess. Bazinga.

Re:OP responding.. (0)

Anonymous Coward | about 9 months ago | (#45405297)

What's Bazinga a misspeling for?

Re:Well, I'll tell you why I'm not interested.. (1)

sanosuke001 (640243) | about 9 months ago | (#45404811)

I logged in to say this. Linus (with an 'S') needs to jump off his high horse and be a personable human being. I have nothing against riding someone for breaking stuff but he's intolerable.

Re:Well, I'll tell you why I'm not interested.. (1)

larry bagina (561269) | about 9 months ago | (#45405385)

Linu[sz] isn't an English word so we have to transcribe/Romanize it to our alphabet. Linus is more common but not more correct (they're both wrong). It's like Ubuntu/OOboontoo or Obama/Ubama.

Re:Well, I'll tell you why I'm not interested.. (0)

Anonymous Coward | about 9 months ago | (#45404877)

With such thin skin I doubt you have the experience to do much of value anyway.
 
And I'm not saying Linus is a nice guy, I'm just saying that he's about par for the course if you work with teams that actually get things done instead of teams that just putz along and on a rare change give up the goods.

Re:Well, I'll tell you why I'm not interested.. (0)

Anonymous Coward | about 9 months ago | (#45405141)

No, it's about personalities, not experience or getting things done. That's his personality, not a management style decision.

Re:Well, I'll tell you why I'm not interested.. (3, Interesting)

ebno-10db (1459097) | about 9 months ago | (#45405393)

I'm just saying that he's about par for the course if you work with teams that actually get things done instead of teams that just putz along

Bull. The best teams I've worked with are often composed of people that play nice with others. Sure tempers flare sometimes, but on the whole the people are reasonable. In fact that's part of the reason the teams are good - yelling and finger pointing are not very productive. It also turns people into stubborn defensive asses that play NIH. In a good team, even when somebody screws up, it's politely pointed out to them, even to the point of not directly blaming them. Good people know when they've screwed up, and work hard to fix it and make sure it doesn't happen again. If they don't do that, get rid of them. Want to vent? Go yell at your dog.

Re:Well, I'll tell you why I'm not interested.. (0)

Anonymous Coward | about 9 months ago | (#45404889)

And nothing of value was lost.

Re:Well, I'll tell you why I'm not interested.. (4, Interesting)

bobbied (2522392) | about 9 months ago | (#45404949)

After watching a few videos of "Linuz"... I can assure you that he's pretty harmless, at least in person. I think he puts on the aura of raging narcissist on purpose and if you think about it, the whole persona serves him and Linux well. So far the Kernel project hasn't been fragmented and the project has been extremely stable for many years. This is not the normal course of an open source project, especially one of this visibility. This is largely due to "Linuz" and his persona.

But this is not to say I think the kernel is in good hands with him at the wheel. I worry about succession should "Linuz" become unavailable (say he's hit by a bus to use his illustration about why you should use git). I worry that the succession battle would be bad for the Kernel and the transition from the dictator rule to something else would be bumpy. Linuz could fix that by starting to transition what he does to his trusted few, and publish a clear future succession plan. But the future is "Difficult to see. Always in motion is the future."

Re:Well, I'll tell you why I'm not interested.. (-1)

Anonymous Coward | about 9 months ago | (#45405205)

... I can assure you that he's pretty harmless, at least in person ...

True, his personality is such that he can only be a *bully* hiding behind a keyboard, hiding behind his fanbois.

Re:Well, I'll tell you why I'm not interested.. (0)

Anonymous Coward | about 9 months ago | (#45405207)

I read the Wikipedia page on Stargazy Pie... I can assure you that it is delicious!

Re:Well, I'll tell you why I'm not interested.. (0)

Anonymous Coward | about 9 months ago | (#45404999)

It's really not that, but an aging Linux kernel dev group will show how robust the kernel is. The Kernel maybe open source [for viewing] but it's somewhat of a wall-garden amongst the dev community (only 'valid' things get pushed, aside bugfixes not crashing the kernel).

If it's not robust, as new devs come in, we'll see more kernel rollbacks or folks staying off the bleeding edge (happening now). Of course, git can create a mess if large groups of folks stick to specific versions where you won't know what branch you need. But that'll be your fault anyway for not sticking to the latest branch, right?

Re:Well, I'll tell you why I'm not interested.. (5, Informative)

Jody Bruchon (3404363) | about 9 months ago | (#45405065)

I have contributed some bug reports and fixes to LKML and I have yet to encounter anything other than a terse but helpful and friendly nature amongst those that picked up my reports and directly communicated with me to fix the code. The only people who get reamed on LKML or get a middle finger are the ones that do egregiously foolish things and should know better. Linux is a massive project that spans thousands of cultures and subcultures in the meatspace department, and there is no time at all to address every error with compliment sandwiches and a facade of "bless your heart" pseudo-kindness.

"Show me the code" is the mantra. If your code is shit and you're new, you'll be politely pointed at a resource such as the coding style guide or KernelNewbies to correct it. If your code is shit and you manage a whole kernel subsystem, you can expect to be told "your code is shit and you know better!" by Linus directly, because....get this: you tried to feed shit code into the kernel (which hurts everyone else because they ALL have to maintain your code down the line) and you're high enough on the food chain that you know better.

Re:Well, I'll tell you why I'm not interested.. (-1)

Charliemopps (1157495) | about 9 months ago | (#45405155)

I was going to post this exact same message. He's a jerk and doesn't realize his fame means people interested in his work are effectively his "Yes men" so he's getting separated from reality. If he worked where I do he'd get walked out the door the first week... how good he was would be irrelevant. It doesn't matter how good your lead is if everyone else quits to get away from them.

Re:Well, I'll tell you why I'm not interested.. (0)

Anonymous Coward | about 9 months ago | (#45405165)

(Score:-1, Troll)

Re:Well, I'll tell you why I'm not interested.. (3, Informative)

poetmatt (793785) | about 9 months ago | (#45405257)

If this is what people think of upstream kernel maintaining, they should probably not troll anonymously.

This is about as far from truth as it is from reality. The man is abrasive, yes, but if you think he's just going to come after you then the problem is absolutely your own perception and not Linus.

Not really a hobbyist project anymore ... (0)

Anonymous Coward | about 9 months ago | (#45405381)

Linux is not much of a hobbyist project anymore. It seems more of a corporate and government subsidized project. Linux is not even in the top 100 kernel contributors anymore.

The kernel will be maintained by those being funded by IBM, Red Hat, etc.

Don't worry (2, Insightful)

Anonymous Coward | about 9 months ago | (#45404635)

Really don't worry. It is commercial enough and if the community just winds down, the companies will just staff the kernel developer ranks,

Re:Don't worry (0)

Anonymous Coward | about 9 months ago | (#45405291)

Really don't worry. It is commercial enough and if the community just winds down, the companies will just staff the kernel developer ranks,

Already there. Linus Torvalds is not even in the top 100 kernel contributors list anymore.

Get on my Lawn (5, Funny)

DeBaas (470886) | about 9 months ago | (#45404637)

Get on my Lawn!

Pöttering should take over the Kernel develop (1)

Anonymous Coward | about 9 months ago | (#45404639)

His controverse ideas will be loved by the majority of the linux tea party crowd :)

why waste time if its free? (-1)

Anonymous Coward | about 9 months ago | (#45404645)

why would i waste my time contributing when i can get the product for free?

Then switch language (0)

Anonymous Coward | about 9 months ago | (#45404657)

If you want to attract new people, then you will have to bite the bullet and switch to a language that does RAII or a similar predictable resource management technique. Nobody in his right mind will write those mind-numbing goto constructs, when the compiler can do this. It doesn't have to be C++, it could be some modern form of C with classes or D.

Re:Then switch language (0)

Anonymous Coward | about 9 months ago | (#45404773)

High level language cause garbage collection so, no-go!

Re:Then switch language (0)

Anonymous Coward | about 9 months ago | (#45405051)

Which part of "predictable resource management" did you not understand? Garbage collection is not predictable and therefore not relevant. Do you even know what RAII is? Neither C++, nor D, nor C with classes cause garbage collection, despite being much higher level than plain C. Your message was a total non-sequitur.

Re:Then switch language (1)

lgw (121541) | about 9 months ago | (#45405211)

If you want to attract new people, then you will have to bite the bullet and switch to a language that does RAII or a similar predictable resource management technique. Nobody in his right mind will write those mind-numbing goto constructs, when the compiler can do this. It doesn't have to be C++, it could be some modern form of C with classes or D.

There is a completely irrational hatred and fear of C++ among C kernel hackers. It's so steeped in the culture I'm not sure it can change (though a flux of young people would help). The funny thing is, when I've spent enough time with veteran kernel hackers to show exactly how RAII would work in their code, they were accepting that it actually works and would make life easier, but were still unwilling to change.

I suspect a big part of it is simply that so much of your day-to-day work habits as a good kernel guy get invalidated (because made unnecessary) that it feels like starting over.

It's not at all just RAII, BTW. There are times when "placement new" (and overloading the new operator) and where templates could massively clean up code by moving common stuff into the compiler.

Re:Then switch language (1)

bobbied (2522392) | about 9 months ago | (#45405287)

It doesn't have to be C++

How wrong can one be? I *might* give you C (but with classes is really just C++), but why would you do that? C++ has its issues, but as a language to write a kernel in I'll take the issues and get the predictable performance in return. You simply cannot do garbage collection in a kernel and get predictable performance for interrupt routines, context switching, signal delivery and the like. C and C++ are very common Kernel languages and for very good reason, pointers. Folks hate them, misuse them, cast them, and bad mouth them all the time, but they are *fast*, flexible and a great way to shoot yourself in the foot when you don't pay attention.

I suppose there is room for improving parts of the kernel by dropping into assembly, but I seriously doubt it will be worth the effort to do much of that. Certainly you'd not do the whole thing. Besides, C++ compilers are really very good at optimizing things anyway, so you'd not get much gain overall.

What's it pay? (2, Funny)

Anonymous Coward | about 9 months ago | (#45404659)

Benies? Dental? Vaction days? Sick days? Comp time or overtime? Weekends off? All national holidays?

Re:What's it pay? (1)

Immerman (2627577) | about 9 months ago | (#45405107)

The benies for being a volunteer are great - all the vacation you want, and you get paid triple-overtime for any work in excess of twenty minutes a week.

As someone who is taking OS course (5, Insightful)

Anonymous Coward | about 9 months ago | (#45404663)

This semester, I am taking OS course at UMBC.
Course is easy, material is easy. Hard part - figuring out how the fuck you should write Linux Kernel code.
Why there are no good tutorials that on how to write basic kernel code, good guides on its structure (many book sold on Amazon are outdated) ......there should be one, centralized place with all the useful materials for the beginners + it should be constantly updated.

Re:As someone who is taking OS course (1)

Joining Yet Again (2992179) | about 9 months ago | (#45404739)

At least it's not BSD.

Seriously, though, mainstream OS implementation is 10% OS theory and 90% careful engineering. So your course will be useful, but it won't be nearly sufficient.

Re:As someone who is taking OS course (1)

fisted (2295862) | about 9 months ago | (#45405027)

At least it's not BSD.

Seriously, though, mainstream OS implementation is 10% OS theory and 90% careful engineering. So your course will be useful, but it won't be nearly sufficient.

You are aware of that the BSDs are carefully engineered, well documented, etc?
Unlike GNU and Linux, which rather have grown.

Re:As someone who is taking OS course (1)

Darinbob (1142669) | about 9 months ago | (#45405157)

Too bad it's not BSD though. Linux code is very often very obtuse and difficult to understand whereas a lot of BSD kernel code is comparatively straight forward. If I were to teach an OS class using source code I'd point students to BSD first (netbsd or openbsd at least, or even 4.2BSD).

Re:As someone who is taking OS course (5, Informative)

DarkOx (621550) | about 9 months ago | (#45404997)

There is: http://kernelnewbies.org/ [kernelnewbies.org]

Re:As someone who is taking OS course (1)

Anonymous Coward | about 9 months ago | (#45405003)

Posting as AC to preserve mod points, I searched and found http://kernelnewbies.org/Documents [kernelnewbies.org] which you probably know. That is referenced by https://www.kernel.org/doc/Documentation/HOWTO [kernel.org] as a resource for kernel developers. Unfortunately it looks like a dying site. The last edited date in the footer is old for many pages. The list of Up To Date Books include stuff published in 2005 or about kernel 2.6.
Probably Linux need a Linuxpedia and developers should at least explain their code there.
However there are plenty of contributors so I'm quite sure there is a way to get up to speed.

Re:As someone who is taking OS course (1)

Immerman (2627577) | about 9 months ago | (#45405237)

Now I'm not a kernel developer, but I suspect there's a lot of old stuff that's still extremely relevant. How much has the kernel really changed in the last eight years? Lots of bug fixes, a few major features, a couple things cut, and lots of new modular components added in. All mostly irrelevant to a newbie-level overview, and if you want to get your hands dirty you need to dig into the sort of cutting-edge details of one particular aspect that has no place in a "getting started" overview. There might be room for some memorandums added to some of those resources, but outside of college textbook publishing houses nobody is going to publish a new book for that.

College Outreach (5, Funny)

Anonymous Coward | about 9 months ago | (#45404665)

Perhaps a campus tour where the senior kernel devs can personally tell prospective developers that they are retarded and kick them in the balls.

too busy (1)

Anonymous Coward | about 9 months ago | (#45404693)

They are too busy working 80 hour weeks to make 60k a year.

The older guys who cant get a job are bashing out the code for the kernel. For resume cred.

and be called perkeleen vittupää? (0)

Anonymous Coward | about 9 months ago | (#45404697)

Nope, think I'll pass.

Maybe when Linus grows up.

Linus' Three Step Plan (-1)

Anonymous Coward | about 9 months ago | (#45404713)

Step 1: Make a very public scene whenever someone does something slightly dumb. Be sure to use horrible names and as much profanity as possible.
Step 2: ???
Step 3: Newbies just beg to work with you. Profit!

I know it's been argued to death but I still think it's stupid.

Re:Linus' Three Step Plan (0)

fisted (2295862) | about 9 months ago | (#45405063)

but I still think it's stupid.

Well obviously you are stupid, as you are judging something which you don't know a wet shit about. Protip: subscribe to lkml and actually see how wrong you are.

Re:Linus' Three Step Plan (1)

Immerman (2627577) | about 9 months ago | (#45405379)

Really? My impression is that he very occasionally dishes out a ration of vitriol on one of the "upper management" who has insisted on doing something against the publicly established guidelines, usually despite repeated applications of more polite dissuasion. Of course some of the more polite stuff can still seem aggressive, but there's only so many ways you can say "you are wrong and your project will not be incorporated / your attempt to shift the blame will not be tolerated" while avoiding misunderstandings across cultural boundaries.

Which is ruder? To call someone out publicly, or try to dissuade them politely and let them dedicate hundreds or thousands of hours of work to a doomed project because they thought you meant "I need to be convinced by a working example". In his position I'm sure he's run up against plenty of those.

Not just young folk... (0)

hazeii (5702) | about 9 months ago | (#45404715)

Let's face it, linux isn't easy to hack now; the corporates are winning (complexity is their friend, if it was simple no-one would need a support contract). Why release a simple system, when you can bloat it with a zillion tweaks of dubious value and then charge money to keep the whole mess working?

Mind you, it's a strategy that's worked well for Microsoft (well, up till now anyway).

Re:Not just young folk... (4, Informative)

nine-times (778537) | about 9 months ago | (#45404823)

Why release a simple system, when you can bloat it with a zillion tweaks of dubious value and then charge money to keep the whole mess working?

I don't think it's really as malicious as that. The larger problem is that everyone has a slightly different definition of what makes a simple, stripped down system. You only want the features you want, I only want the features that I want. You want a rock-solid server; I want a responsive and feature-rich desktop system; my brother just wants to play video games. You can't do it all without a little bit of complexity.

And look at what happens when they try. Someone proposes a new window compositing system that will make development easier and performance more responsive, and people get all bent out of shape because it breaks the X11 spec.

Microsoft is a whole other ball of wax. Chronic mismanagement, perverse incentives to sabotage any product which might cannibalize the Windows/Office products, and an attempt to maintain backwards compatibility as much as possible, going back to DOS systems from a quarter century ago.

Re:Not just young folk... (0)

Anonymous Coward | about 9 months ago | (#45404897)

Feature-rich and stripped down are opposites. You want features to be available to you on a whim, learn to install new software. You want stripped down, learn how to use the core set of tools.

Re:Not just young folk... (3, Insightful)

vux984 (928602) | about 9 months ago | (#45405247)

Feature-rich and stripped down are opposites. You want features to be available to you on a whim, learn to install new software.

You are missing the point. A stripped down system can still have a feature rich kernel. If I want a feature rich desktop then I need a kernel that has features that enable the sort of high performance UX I need.

Right down to a scheduler that's friendly to interactive user processes. But maybe that scheduler's not as optimal for what you were doing with your server, so now we want a tunable scheduler that can be adjusted towards either.

And the complexity begins its lift off.

Re:Not just young folk... (1)

Anne Thwacks (531696) | about 9 months ago | (#45404909)

You only want the features you want, I only want the features that I want.

In which case, *BSD is the answer to your prayers. You only get the featues you actually ask for, and once you have got them, they stay got!

Re:Not just young folk... (0)

Anonymous Coward | about 9 months ago | (#45404863)

There is also the fact that there is an attitude change. The "stone soup" of days past is gone, with only the old guard in place. The younger people want to hear the ka-ching sound when writing code, not the fact that they wrote something to scratch an itch.

If it isn't easy like writing an app in Objective C, then there isn't an interest. Might as go back to playing WoW, since there are a couple alts they can gear up on Timeless Isle before the next expansion.

We see this in programming languages. If it isn't something point and drool that supports sloppy programming, you will find only the "old fossils" using it. Not just COBOL, but C and even perl are considered antiquated compared to Ruby on Rails, PHP, or whatever the latest thing is for coding SEO-optimized crap for more blog hits.

Re:Not just young folk... (1)

Mitchell314 (1576581) | about 9 months ago | (#45405007)

And maybe they go for the ka-ching because they have bills to pay, little savings to live off, and are trying to establish themselves so that they can continue that bill-paying thing. Don't be so bloody condescending.

Re:Not just young folk... (1)

lgw (121541) | about 9 months ago | (#45405269)

There is also the fact that there is an attitude change. The "stone soup" of days past is gone, with only the old guard in place. The younger people want to hear the ka-ching sound when writing code, not the fact that they wrote something to scratch an itch.

Being a skilled Linux kernel hacker is a great career move, though. I routinely get interviewers contacting me for top-tier Linux kernel architect jobs when there's barely a suggestion in my resume that I could actually do that work. Demand far outstrips supply there, even compared to other senior positions.

It's a bloated mess. Who wants to work on that? (0, Flamebait)

Animats (122034) | about 9 months ago | (#45404725)

Well, what do you expect? The thing is a bloated mess.

Re:It's a bloated mess. Who wants to work on that? (1)

arkane1234 (457605) | about 9 months ago | (#45405021)

Well, to be truthful it's a bulk of modules. The main kernel isn't any more bloated than any other Unix operating system with safeguards and security code within it.

Re:It's a bloated mess. Who wants to work on that? (0)

FudRucker (866063) | about 9 months ago | (#45405071)

it is not really that messy, it is well organized and categorized, but it is getting rather bloated, i just did a slackware-14.1 install and /usr/src/linux-3.10.17 weighs in at 589 megs, and after a quick look inside /usr/src/linux/* i wonder how much of that code is not even needed because slackware is only building for x86 & x86_64 architectures (i think there is an arm build too) i bet some of that code could have been left out, and Linus could have separated the kernel source in to source code bundles intended to be built on different architectures, because i know some of that code is not built and installed on x86 & x86_64.

It's so much fun! (1, Informative)

Anonymous Coward | about 9 months ago | (#45404727)

It seems like most young programmers are way more interested in developing mobile-apps than in getting yelled at, cursed at, and described as being in compromising situations involving Microsoft.

Re:It's so much fun! (1)

CanHasDIY (1672858) | about 9 months ago | (#45404941)

It seems like most young programmers are way more interested in getting paid.

FTFY.

Welcome to the Casino Economy - if it ain't making the House richer, it ain't worth paying anyone to do.

Re:It's so much fun! (1)

DarkOx (621550) | about 9 months ago | (#45405023)

Kids today... Geez

Consider the possibility it might be done (3, Interesting)

Anonymous Coward | about 9 months ago | (#45404731)

I know I've mentioned this before, but you need to consider the possibility that your software might be done.

Take TeX for example. The last stable release is 5 years old. It's done.

At some point even the OS kernel will switch from "active development" to "something people study". We studied the circuit diagrams for radio receivers, memory circuits, and even more complicated things like 8-bit ALUs. They're done. We weren't developing that stuff in school. We were just understanding it.

The Linux kernel will end up in a text book some day. People will want to understand it. Nobody will want to develop it. That's a good thing. It means that this phase of technology is approaching the done phase.

What's the next phase? If you're young that's where you should be looking.

Re:Consider the possibility it might be done (0)

Anonymous Coward | about 9 months ago | (#45405013)

Because computer technology continues to advance and new hardware is released constant requiring constant kernel updates to keep up with it?

Your comparison to TeX is makes no sense because these are two pieces of software who's only non-trivial (in this context) commonality is that they are software.

Re:Consider the possibility it might be done (0)

Anonymous Coward | about 9 months ago | (#45405061)

OS kernel development will never be "done" so long as things like device drivers and file systems exist in-kernel.

Re:Consider the possibility it might be done (4, Insightful)

fisted (2295862) | about 9 months ago | (#45405089)

I know I've mentioned this before, but you need to consider the possibility that your software might be done.

Considered and considered stupid, because suggested in the context of operating systems. Operating systems are only done when hardware is 'done', which is unlikely to happen any time soon IMO.

Re:Consider the possibility it might be done (0)

Anonymous Coward | about 9 months ago | (#45405101)

Except the kernel contains all its drivers within its source tree. Everybody who wants to use it on new hardware as to write a driver for that, and some of them are open sourced and committed into the tree.

As long as it stays structured that way it can only stop changing when it stops being used on new hardware. At which point it would seem to be dead.

Re:Consider the possibility it might be done (2)

dgatwood (11270) | about 9 months ago | (#45405229)

Except the kernel contains all its drivers within its source tree.

Not at all. I've built and installed drivers that aren't part of the kernel source tree on several occasions. The entire driver stack has to be part of the kernel, by necessity, but not the drivers themselves.

IMO, the Linux kernel should pull all of the drivers out of the kernel source tree and into separate projects so that they can be separately maintained. The "everything in the kernel project" model means that every fix to every driver is dependent upon the upstream maintainers taking the changes. That's a maintenance nightmare.

Re:Consider the possibility it might be done (1)

DarkOx (621550) | about 9 months ago | (#45405105)

The kernel is not like most software project though because of what it does. It provides an interface to hardware for other software. That other software might be done but as long as the hardware changes the kernel is never done.

You could argue that something like a scheduler might one day be done, but the rules change, memory is cheap in plentiful even on the smallest devices, it was a major constraint when Linus started Linux. Now its okay for your scheduler to use much more memory if that gets you to other properties you want.

Most computers had one CPU and one decoder, than it was one CPU and a couple decoders. Now its multiple cores with a couple decoders and shared cache. Even if what worked before still works now, whats best now is different.

Storage technology is evolving as well, as is network technology.

While some areas of the kernel might mature for a time until some sea change happens, I don't the project as a whole will ever be done.

Re:Consider the possibility it might be done (2)

dgatwood (11270) | about 9 months ago | (#45405139)

The next phase is almost always to replace the technology or some portion thereof with something that does the same thing, but in a better or easier to use way.

With TeX, the logical next step is HTML/CSS typesetting. I'm pretty sure you could replicate most of the interesting parts of LaTeX in only a few thousand lines of JavaScript, assuming you had a browser that supports most of CSS3, but you'd also get lots of stuff that LaTeX can't handle.

With Linux as a whole, the logical next step is to repeatedly shoot X11 in the head until it stays down, then build something modern to replace it. Possibly replace it with Wayland—hard to say.

Down at the kernel level, the next logical step is to reduce the over-reliance on unreliable device names. Compared with OS X, the hot plug configuration story seems downright miserable to me. Device node names in the filesystem change, Ethernet interface names change depending on what order stuff gets plugged in, and so on, and the only way to fix this involves creating cryptic configuration files that try to pin devices in position. Blech. Those names shouldn't matter; if they do, you're doing it wrong.

Basically, Linux needs to do something similar to OS X's kextd and I/O Kit frameworks. Start by working up a standard API for walking the device tree, searching for devices that match against specific criteria, and returning a usable reference to that device (e.g. an open file descriptor). Then, create network configuration tools that use that matching system to identify the interface and configure it appropriately using a configuration dictionary. Then the next logical step is to load drivers based on those matching dictionaries. And then you have something approaching a usable driver loading system.

Re:Consider the possibility it might be done (1)

Darinbob (1142669) | about 9 months ago | (#45405185)

Not really, because hardware isn't done. Operating systems keep evolving to keep up with the hardware. Also software isn't done, so operating systems keep evolving to keep up with newer paradigms in software. Linux is constantly changing, and not just to keep some old coders busy.

start over (0)

stenvar (2789879) | about 9 months ago | (#45404777)

The Linux kernel has served us reasonably well, but perhaps it's time for a new generation to create a new generation of kernel.

t works great! Throw it out and start again! (0)

Anonymous Coward | about 9 months ago | (#45404883)

"It works great! Throw it out and start again!"

Your idea is bad, and you should feel bad.

Re:start over (1)

Sesostris III (730910) | about 9 months ago | (#45405017)

The Linux kernel has served us reasonably well, but perhaps it's time for a new generation to create a new generation of kernel.

Hurd?

Re:start over (1)

sconeu (64226) | about 9 months ago | (#45405125)

Yeah, I understand that Hurd is going to be big and professional, unlike Linux [google.com] .

My reason for not getting involved. (5, Interesting)

hamster_nz (656572) | about 9 months ago | (#45404809)

It is just too damn big, hard and complex. Why would I want to learn the ins and outs of such a large codebase unless somebody is paying me to?

It is not like the old days when you could pick up a "... in a nutshell" book, start hacking up a driver, then get it accepted into the kernel. I don't want a three year unpaid intership while I get up to speed and gain respect in the comunity.

I'll spend my time working on my project on either a microcontroller (AVR, PIC...) or a bare-metal build on ARM.

Re:My reason for not getting involved. (1)

Anonymous Coward | about 9 months ago | (#45404901)

Nobody sane knows everything about all parts of the kernel. That's why its "compartimentalized", you just have to look at the small portion that interests you -- memory management, scheduling, vfs, a particular filesystem, networking (whatever part interests you), what have you.

Re:My reason for not getting involved. (0)

Anonymous Coward | about 9 months ago | (#45405059)

I'll spend my time working on my project on either a microcontroller (AVR, PIC...) or a bare-metal build on ARM.

Look, you're welcome to reinvent the wheel all you want, but FFS just give that as your reason instead of whining about "OMG OPERATING SYSTEMS IZ HARD". After you've made all the mistakes (on "bare-metal ARM") that the last gen of kernel developers already fixed, then maybe you'll have a point.

Re:My reason for not getting involved. (0)

Anonymous Coward | about 9 months ago | (#45405143)

Exactly. I'd rather write an iPhone app where my skills will be valued than contribute to some complex niche OS kernel that almost no one uses.

Completed System (0)

Anonymous Coward | about 9 months ago | (#45404895)

The problems with the Linux Kernel itself are fundamental and unfixable: too big, at a million lines, and not designed from the beginning for minimum trust of all the many components.

Not that it isn't better than the alternative with 30 million lines or more from Redmond -- but since that time, all manner of malware has been written, some of which also appears to accomplish useful work, or corrupts things that do useful work, and it is time for a system that intrinsically distrusts any programs or drivers it is running to do the right things, and ensures that the system owner can retain control.

And Linus, before you toss me some invective, tell me about the compiler optimizations that get disabled because of dependencies on dividing by zero raising exceptions in the kernel???

Learning curve is a 80 degree gradient (0)

Anonymous Coward | about 9 months ago | (#45404945)

I actually was looking at FOSS projects to contribute to for awhile and Linux was one of my first choices.

A lot of the experienced devs will gladly tell you that they're looking for fresh breath, but they aren't so willing to offer documentation or offer assistance in starting out. I'm not expecting my hand to be held, but even some kind of informal write-up on something aside from a style guide would help, yeah?

Re:Learning curve is a 80 degree gradient (1)

arkane1234 (457605) | about 9 months ago | (#45405253)

I'm no kernel code developer, but I'm a systems engineer with light C/C++ experience and that is the reason I stopped contributing to the Gentoo community as well.
While they don't parallel in many ways, the team thought process was kind of the same.
I created full documentation on logical volume management installation with raid setup and so forth during Gentoo installation from scratch and it amounted to them saying to learn their documentation coding language and put it into that. I had other things to do aside from learning a language (some odd hybrid SGML that wasn't well documented) that I'd only use for that purpose.
I used Gentoo for a bit after that but it got to the point of not compiling cleanly due to dependency issues which told me the package management system was suffering the same issues as the documentation more than likely. Haven't really used Gentoo since 2005 for those reasons.

Let me get this straight (1)

scottnix (951749) | about 9 months ago | (#45404951)

I'd have to learn a code base developed by near-religious zealots, written in a language 20 years beyond chic only to be treated like a small minded idiot when I make mistakes? I've quit *well paying* development gigs because they met only one of those criteria. There's no way I'd go through that for free. Not for all the resume fodder on Earth.

Re:Let me get this straight (0)

Anonymous Coward | about 9 months ago | (#45405099)

[...] written in a language 20 years beyond chic [...]

Right, right, style and fashion is what's really important in software engineering. My mistake. So, praytell, how's your ultra-hip and stylish Python port of the Linux kernel coming along? No performance issues, I'd imagine?

Re:Let me get this straight (0, Insightful)

Anonymous Coward | about 9 months ago | (#45405121)

Your input is not needed, obviously.

You have no sense of ethics ("code base developed by near-religious zealots"), you don't understand the value of C or C++ ("language 20 years beyond chic") and you don't have enough ambition ("resume fodder").

What I do agree with you is there is no need to treat people condescendingly.

Just sit back and watch the Linux kernel triumph.

Re:Let me get this straight (2)

arkane1234 (457605) | about 9 months ago | (#45405263)

written in a language 20 years beyond chic

No, you'd be coding in C.

Its a trap!!! (4, Funny)

FudRucker (866063) | about 9 months ago | (#45404953)

they just want to kick you off their lawn

Netcraft Has Confirmed: Linux is old. (1)

Scholasticus (567646) | about 9 months ago | (#45404989)

See subject line.

Linux really doesn't... (-1)

Anonymous Coward | about 9 months ago | (#45405075)

.... deserve younger generations IMHO. If I was linus I would be trying to get massive funds to make my OS useful to people (in terms of user interface/software compat) who use applications. Linux' big problem is that the world doesn't use linux as a main OS. You want contributors? Try making forked desktop OS that's actually a carbon copy of windows in terms of UI and can run all mainstream games.

People just don't get up and learn obtuse difficult and complex shit for no reason and especially while not being paid. The kind of community linux attracts is not attractive. Linus has had over a decade and he still hasn't gotten it into his skull that if you want an OS to be something more then an engineers toy you need to actually look at successful operating systems on the desktop and what they are doing right.

The fact that valve had to go and do Steam OS is more proof many engineers and geeks are lost without guidance of a visionary that wants to put Operating systems in people hands that are usable and mimic dominant OS's verbatim in terms of UI and usability.

That's one thing I never got with the linux crowd, why not just fucking clone windows desktop UI for linux where everything is under the same name but only slightly modified for linux? (control panel, my computer, hardware, etc, etc).

Re:Linux really doesn't... (1)

maxwell demon (590494) | about 9 months ago | (#45405367)

The last thing I would want is a clone of the Windows UI. If I wanted Windows, I'd know where to get it.

Kernel Development vs Linux Kernel Development (0)

Anonymous Coward | about 9 months ago | (#45405161)

I'm quite interested in very low level programming and believe it wouldn't be too hard to be able to do it based on my current skills however for me it's more the bigger picture than fixing up something else someone else has written.

While there are plenty of merits for working with the Linux kernel including all the prior effort that has been put into it, I would rather construct something entirely new from scratch. I'm sure many people will say that's crazy however it's both the challenge that I like and I probably will enjoy it more. Of course it would be hard to catch up to anything like Linux however it's not the point of every endeavour to beat every and all competition.

Another big thing is major architectural changes that I might want to write but would never ever make it into the mainstream Linux kernel. Perhaps I no longer want the current file permission model, perhaps I don't like '/' being the root to every folder on the server or even perhaps I don't like 'root' being called 'root'. I don't want to fork the kernel at some point and just make those changes myself, these are fairly large changes that would break so many things and it's likely I won't even enjoy it any more.

Big moves like writing a new kernel from scratch can allow new insights into problems, do drastic improvements throughout the system without backwards compatibility issues and allow the next generation of features and functions to make a cleaner implementation.

Younger kernel devs? (1)

Darinbob (1142669) | about 9 months ago | (#45405201)

If they get a lot of younger kernel devs, do we stand a chance of seeing the kernel equivalent of Unity?

they should be worried (0)

Anonymous Coward | about 9 months ago | (#45405243)

when I see young people at coffeeshops, college, gay bars, etc, I almost never see linux. They often use OS X, Windows 8, even FreeBSD or Illumos. But linux? Nope. I asked one, a twink named Christian, if he ever used Linux. "Linux? That's for like old people. It's not cool like Android." I tried to correct him, to no avail. "What? Android is Google." I dropped the subject, as he could give me orgasms but no answers.

Well well (0)

Anonymous Coward | about 9 months ago | (#45405279)

Who would have thought young people wouldn't want to be vitriolically spat at all day by people who don't shower for 3 weeks and have hair down to their waists and fingernails as long as a cats claws, who complain and whine nasally as they stuff another fucking handful of Cheetos into their gaping maw, pointing fingers and condescending to beat the band so they can fuel their egos to make up for the 30+ years they spent getting rejected by the opposite sex ( and the same sex for a generous portion of them ), only to settle for the neighbourhood whale with 300+ pounds and 3000+ cocks on her.

I always said the death of the Linux kernel would be the people themselves, not the lack of mainstream OS adoption or the mahoosive code base, but those fucking smelly, condescending, arrogant, asocial, unkind debased fucks who call themselves maintainers.

I, for one, welcome their demise with open arms.

Good fucking riddance.

Utter bollox! (0)

Anonymous Coward | about 9 months ago | (#45405331)

The Linux kernel is not a bunch of hackers mucking around, it's a project maintained by a small army of paid professionals. The days of getting involved are long gone unless you start working for a manufacturer that needs Linux drivers.

The seconds reality of today is that the young "developers" do little more than hack on scripting languages for basic web functionality, and using the likes of jQuery. They're as far removed from OS level knowledge are you can get.

If not Linux, then what? (0)

Anonymous Coward | about 9 months ago | (#45405365)

If Linux goes out of favor, then what replaces it? More Windows? A new kernel created as shareware/freeware that replaces Linux?

It just seems to me either Linux lives on, is replaced, or back to Windows. Just don't let it be a Windows only world.

youth subculture - we need it (1)

Pav (4298) | about 9 months ago | (#45405399)

Perhaps a major part, or a major branch (experimental revisited?) needs to be handed over to the control of young blood. It'd be difficult for the old guard... young guys take risks and make more mistakes, but they bring energy and ideas. In any case an area under the control of a new generation MUST happen for a young Linux subculture to further develop, and considering the importance of community as a motivation this is overdue. The size, momentum and commercial interest in Linux will could make this less than straightforward.
Load More Comments
Slashdot Account

Need an Account?

Forgot your password?

Don't worry, we never post anything without your permission.

Submission Text Formatting Tips

We support a small subset of HTML, namely these tags:

  • b
  • i
  • p
  • br
  • a
  • ol
  • ul
  • li
  • dl
  • dt
  • dd
  • em
  • strong
  • tt
  • blockquote
  • div
  • quote
  • ecode

"ecode" can be used for code snippets, for example:

<ecode>    while(1) { do_something(); } </ecode>
Create a Slashdot Account

Loading...