Kuro5hin.org: technology and culture, from the trenches
create account | help/FAQ | contact | links | search | IRC | site news
[ Everything | Diaries | Technology | Science | Culture | Politics | Media | News | Internet | Op-Ed | Fiction | Meta | MLP ]
We need your support: buy an ad | premium membership

[P]
Linux Kernel Politics

By Hobbes2100 in Op-Ed
Sun Feb 03, 2002 at 11:09:28 PM EST
Tags: Round Table (all tags)
Round Table

Recent events have questioned the practicality of the current Linux kernel development hierarchy. This article discusses the issue of Linus Torvalds' role and lays out some future possibilities.


Recently, there has been a fair amount of talk regarding Linus Torvalds' apparant inability to manage the volume of patch submissions to the Linux kernel. This has been discussed by kernel people and by some parts of the Internet community.

A very funny thing has happened in these discussions (or if it didn't happen, it has definitely come to light). Somewhere along the way, people started referring to Linux as Linus' property. From this comment: Of course Linus has every right in the world to remain the status quo, even if it damages Linux. After all Linux is his baby and he can do with it what he likes. I agree with part of this statement; I disagree with the justification. Linus is free to do whatever he wants but it is not because it is his baby. He is free to do whatever he wants with the kernel because 1) it is distributed under the GPL and 2) the Linux kernel community is willing to follow him.

Since the first of these points is self-protecting and will not be going away anytime soon, I'm interested in point number two: the Linux kernel community is willing to follow him. Here, I'll investigate three possibilities. I don't claim that these are exhaustive.

  1. This first possibility is based on the fact that more and more companies are making Linux part of their business strategy. This being the case, these companies may not be very happy with a one man show being both a bottleneck to improvements and a single point of failure.

    What action(s) might these businesses take? One possibility is that they would work together and form an open consortium to act as kernel maintainer. This would have the benefits of stability and openness [1] but it would likely suffer from a "design by committee" mentality.

    Certain individual businesses, like IBM, could certainly take over the whole task of kernel maintenance themselves but it is most unlikely they would even attempt to do so. Such corporate "participation" in the kernel would not be very well received by many individual end-users. However, on the other side, corporate clients might find many benefits in saying that they are using something "made" by Big Blue.

    I highly doubt either of these options will ever see the light of day. However, the consortium possibility is intriguing. It would be even more so (and more acceptable to me) if it included both big name companies and various individuals from the already existing kernel community (but I wouldn't hold your breath waiting for Torvalds to be among them).

  2. The second possibility is that a small, highly motivated group of people will start a kernel fork of their own, just for the fun and interest of it. They won't have any desire to become dominant figures in the Open Source community. They'll just want to have fun. And then, at some point, some outsiders will take note of their project and think, "hey, I think I'll try that out." Having tried it out and like it, these individuals will send it off to a few friends and all of a sudden, a new kernel player will emerge.

    Is this very likely? No. But given enough time, even unlikely things become possible.

  3. The last possibility, and the most likely, is that the status quo will remain ... for the time being. Perhaps the position will become hereditary, who knows? At some point, Torvalds will have to pass the torch on. This might be because of a coup d'etat or because of retirement. Either way, I think I would pay for a front row seat to that show.
All three of these scenarios have assumed that Torvalds won't work with the community to make improvements (or changes, if you prefer). There are other possibilities for more cooperative solutions but monarchs don't tend to give up authority very easily. Perhaps the Linux kernel does not need an overthrow; it just needs its Magna Carta.

[1] This might be a different sort of "open" then "Open Source". However, it should include the "Open Source" sense as well.

Sponsors

Voxel dot net
o Managed Hosting
o VoxCAST Content Delivery
o Raw Infrastructure

Login

Poll
In the future, I see the Linux kernel:
o being maintained by an immortal robot named Linus. 9%
o being overseen by the "Peers of the Source". 27%
o being ruled by a corporate oligarchy. 1%
o being run by me. 6%
o being run by RMS. 4%
o becoming self-replicating and intelligent. 26%
o disappearing into oblivion. 23%

Votes: 93
Results | Other Polls

Related Links
o Linus Torvalds'
o kernel people
o Internet community
o this comment
o Also by Hobbes2100


Display: Sort:
Linux Kernel Politics | 18 comments (18 topical, editorial, 0 hidden)
Comment (3.71 / 7) (#1)
by quartz on Sun Feb 03, 2002 at 03:18:52 PM EST

I'm confused. First you talk about how the Linux community "is willing to follow" Linus, then at the end you speak of monarchs and authority. Which one is it? IMO, since Linux is GPLed, the only authority Linus has over it is the one invested in him by the community. He's more like an elected leader than a monarch. Who cares if he's willing or not to give up that authority; if/when the community needs another leader, the authority will simply be taken away from him.

Now for the scenarios:

  1. If such an "open" consortium is ever formed to maintain the kernel, my bet is that the community will never follow; they love their freedom too much to trust a bunch of corporations, no matter how "open" they claim to be. The community will probably ignore the corporations and maintain its own version of the kernel. As for who will get the right to the name "Linux", your guess is as good as mine. <g>
  2. IMO such a group just can't form outside of the existing kernel community. When someone wants to become a kernel hacker, what do you think he'll do first: subscribe to the various kernel mailing lists or seek other motivated people outside the existing kernel community? The answer seems obvious to me. Now if we're talking about a part of the existing community splitting off, that's not only theoretically possible, but is a common occurence is most free software projects.
  3. Hereditary? Ha! That's a good one. You're not being serious, are you?

Talking about the future of the kernel, there's a possibility you didn't consider, namely that at some point in time HURD will become such a huge success that it will draw all the support (community and corporate) away from Linux. Yeah, I don't think that's gonna happen either. :) But seriously, Linus said it himself (too lazy to find a link, but he pretty much repeats it in every interview): he doesn't care about his personal authority over the thing; he will respect the will of the community and help whoever the community elects as the new leader(s) take over the role.



--
Fuck 'em if they can't take a joke, and fuck 'em even if they can.
Linux kernel (4.00 / 1) (#2)
by twodot72 on Sun Feb 03, 2002 at 03:36:04 PM EST

As for who will get the right to the name "Linux", your guess is as good as mine.
That would, again, depend on Linus. Linux is a registered trademark of his.

I think the most likely scenario is that the other core developers will eventually convince Linus of the need for a better system for handling patches and additions to the kernel, or, possibly, that Linus will choose to pass on his duties to someone else (possibly a group of people).

I think forking is less likely. A successful fork very rarely happens, especially in high-profile projects like this. There are simply too much vested interrests in keeping it all together.

[ Parent ]

Open vs. open vs. open ..... (none / 0) (#7)
by sgp on Sun Feb 03, 2002 at 09:10:13 PM EST

Re. Scenario 1: It depends on how "open" it is - if IBM suddenly declare, "We're going to take on Linux kernel development - submit your patches to us if you're sick of Linus dropping them", odds are they'll get quite a few submissions. IBM are happy with Apache, they're happy with Linux. Okay, you can sed s/IBM/My_Fave_Nominee/g since it's GPL, and only Linus himself can un-GPL the kernel.

But if IBM are happy with the GPL status of the kernel (as they certainly appear to be), then why not? They've got the experience, not just with Linux, but with OS development in general, they've got the staff, the money... and it'd say GPL'd. So the rest of us can stay happy.

I suppose if they made so many huge changes that they could prove there was no inherited GPL code left in their variant of the kernel, that'd be an interesting test for the GPL.

But there's nothing to say that IBM couldn't raise a coup if they wanted to, and if enough existing kernel gurus were behind it.
But that's a big "if".

There are 10 types of people in the world:
Those who understand binary, and those who don't.

[ Parent ]

GPL and the kernel (5.00 / 1) (#9)
by FattMattP on Mon Feb 04, 2002 at 12:22:57 AM EST

only Linus himself can un-GPL the kernel
Wrong. Linus can re-release the sections that he wrote and holds copyright for under a different license. However, many people have contributed to other parts of the kernel under the GPL. If someone wants to use the kernel under a different license, they need to receive the consent of each and every copyright holder. The Mozilla project already went through something like this.

[ Parent ]
Oh yeah. (3.00 / 1) (#13)
by sgp on Mon Feb 04, 2002 at 06:05:35 AM EST

I forgot. Me bad. :-(

There are 10 types of people in the world:
Those who understand binary, and those who don't.

[ Parent ]

Sorry if I was vague (none / 0) (#16)
by Hobbes2100 on Mon Feb 04, 2002 at 12:47:40 PM EST

Quartz,

I'm sorry if I was a little vague. I meant monarch in sense 2 over at Merriam-Webster. So, even if he (Linus) is the front position, people could start following someone else -- putting that person in the lead.

It also tied in nicely to what I see as the call for a Magna Carta of sorts. Clearly definied roles and a slightly expanded power base could do the kernel a large amount of good (IMHO, of course).

Regards,
Mark
Sed quis custodiet ipsos custodes? --Iuvenalis
But who will guard the guardians themselves? -- Juvenal
[ Parent ]

What they can do? (2.83 / 6) (#3)
by DeadBaby on Sun Feb 03, 2002 at 03:50:34 PM EST

A company who is upset with Linus running the show can use another OS. The BSD license is much more logical for business -- it's not a license virus and it runs all popular Linux apps as well or better than Linux in most cases with native versions of all the server applications that matter.





"Our planet is a lonely speck in the great enveloping cosmic dark. In our obscurity -- in all this vastness -- there is no hint that help will come from elsewhere to save us from ourselves. It is up to us." - Carl Sagan
Append (none / 0) (#4)
by DeadBaby on Sun Feb 03, 2002 at 03:54:01 PM EST

IT being some BSD variant, not the BSD license itself which... probably wouldn't run any software very well...
"Our planet is a lonely speck in the great enveloping cosmic dark. In our obscurity -- in all this vastness -- there is no hint that help will come from elsewhere to save us from ourselves. It is up to us." - Carl Sagan
[ Parent ]
Terminology (none / 0) (#8)
by fluffy grue on Sun Feb 03, 2002 at 11:33:51 PM EST

BSD:BSDL::Linux:GPL
--
"Is not a quine" is not a quine.
I have a master's degree in science!

[ Hug Your Trikuare ]
[ Parent ]

Just wondering (2.57 / 7) (#5)
by spacejack on Sun Feb 03, 2002 at 04:03:22 PM EST

What exactly is "Linux"? As I understand it, merely a simple one-word trademark owned by Linus Torvalds.

I wonder how much marketing money it would take to create another brand with greater recognition? I mean, it's not like you'd need to write or pay for the source to the OS casually known as "Linux"...

Linus doesn't scale - Repeat (3.87 / 8) (#6)
by Builder on Sun Feb 03, 2002 at 04:21:06 PM EST

You know, living on the Internet is like watching Sky tv. It's one bloody re-run after another.

We saw the Linus doesn't scale discussion a couple of years back. At the time it took a face to face at someone's house with some of the big names, including Linus to sort it out. Things came close to a fork then (according to rebel code). Funny thing is that one of the now trusted luitenants, Alan Cox was on the team working on the potential fork tree.

These things will happen. Loads of people are bitching about it, someone senior _might_ call for a fork, but I doubt we'll see one.

One conditional to all of this is that I think if a fork happened it could not be named Linux. Linux is a trademark of Linus Torvalds. I'm not sure how the code being GPL'd affects this.

One last thought for the inexperienced. Source control may work on some teams. It probably worked fine for your university project. But it is by no means a silver bullet. You cannot apply a technical solution to a social problem (paraphrase of Edwards' law).


--
Be nice to your daemons
Solutions Already Exist (3.00 / 4) (#10)
by miah on Mon Feb 04, 2002 at 12:39:48 AM EST

I think this problem is being voiced by a small minority. The core of the problem has already been solved.

It was solved by having different patch branches the existing source. Most major vendors apply at least a few patches to their "release" kernels. Alan Cox has one and I believe there are three or four other people that have big patches that affect major things like the VM system and the VFS layer.

Through the patches innovation and experimentation is solved while making the 'stock' kernel very stable (other than the devel branches). The different patches are like little spawning pools where the more successful offspring eat the less fortunate mutants and then grow up to be merged back with the rest of the school (for a really bad analogy).

Don't worry too much about it, it's fix itself. This 'Open Source' stuff has been around for a while and isn't going anywhere as long as people are still using it.


Religion is not the opiate of the masses. It is the biker grade crystal meth of the masses.
SLAVEWAGE
Linux kernel politics. (2.33 / 3) (#11)
by scanman on Mon Feb 04, 2002 at 04:53:35 AM EST

Recently, there has been discussed by kernel people and by some parts of the system's origin, history, and purpose? Yes--because people who forget history are often condemned to repeat it. The free world which has developed around gnu/linux operating system. People who love this system ought to know that it is not because it is his baby. He is free to do so. Such corporate "participation" in the kernel would not be going away anytime soon, i'm interested in point number two: the linux kernel community is willing to follow him.

Recently, there has been a fair amount of talk regarding linus torvalds' apparant inability to manage the volume of patch submissions to the linux kernel community is willing to follow him.

This would be wiser to let the matter drop. (But we are not in detail) an accurate idea. This first possibility is that they are using a system that came out of the internet community.

Since the first of these options will ever see the ght of day. However, the consortium possibility is based on the non-free software; the companies that package "linux" add non-free programs that use them.

Granted that the move toward open source software should be fueled by technical, rather than political, decisions." and caldera's ceo openly urged users to drop the goal of freedom. It is no time to take the future for granted. Our community's strength rests on commitment to freedom and their community, determination to keep on for years and not give up.

If "the job" really were done, if there were nothing at stake except credit, perhaps it would likely suffer from a "design by committee" mentality.

"[You are] a narrow-minded moron [and] a complete loser." - David Quartz
"scanman: The moron." - ucblockhead
"I prefer the term 'lifeskills impaired'" - Inoshiro

EGCS! (5.00 / 5) (#12)
by Ranieri on Mon Feb 04, 2002 at 04:57:49 AM EST

The second possibility is that a small, highly motivated group of people will start a kernel fork of their own, just for the fun and interest of it.
...
Is this very likely? No. But given enough time, even unlikely things become possible.

This is exactly what happened to GCC. The project was stagnating. Pentium code generation had been waiting for inclusion for months, as had other major achitectural changes. Forks were showing up all over the place, including the famous PentiumGCC. Then a few people decided to launch the EGCS project, an experimental GCC tree aimed at including all the features that the "official" GCC tree didn't have.
Eventually EGCS became the compiler of choice, and the EGCS team was appointed to lead the GCC development effort, replacing the old GCC. This is the announcement in question, from the FSF site.

In April 1999, the egcs steering committee was appointed by the FSF as the official GNU maintainer for GCC. At that time GCC was renamed from the "GNU C Compiler" to the "GNU Compiler Collection" and received a new mission statement.

Alan has already dropped a hint that Linux might be headed down the same path. Personally, i look forward to this change. I am confident that, just as EGCS made GCC a better compiler, an "E-Linux" might make Linux a better kernel.
--
Taste cold steel, feeble cannon restraint rope!

Linus will have to bow out... (3.00 / 1) (#14)
by maroberts on Mon Feb 04, 2002 at 09:53:20 AM EST

..at some point, but I tend to suspect that the second level of trusted lieutenants e.g. Alan Cox will take more of a leading role as time goes on.

Personally I would be deeply against any one person or corporate entity taking over after Linus does hand over the reigns of power, whether that person be Alan Cox or the entity be IBM or RedHat.

I believe that Linus, if he has any sense, will evolve from being a leader to a guide, and thus leave the development scene in the most graceful way possible. I think the next phase of Linux leadership will need some form of committee, but it will need to be a committee of people with an intense drive to see Linux forward, and thus ensure that Linux does not stagnate. I think that the best role corporations could play in such a committee would be to sponsor or to employ members of that committee, with such sponsorship to be limited to one member per corporate sponsor.

Whilst Linux is becoming too big for one person, it also does not need a whole raft of chiefs. Linux has succeeded in part because whilst many have contributed, only a few have been responsible for its direction, and this has allowed the vision of those few to be realised.


~~~
The greatest trick the Devil pulled was to convince the world he didn't exist -- Verbil Kint, The Usual Suspects
Linus owns Linux(tm) (4.00 / 1) (#15)
by Mr Fred Smoothie on Mon Feb 04, 2002 at 12:25:38 PM EST

"Linux" is a registered trademark of Linus Torvalds. Even if some Consortium of businesses were to fork & maintain a new "official" kernel, they couldn't call it Linux without Linus' permission.

Linux(tm) is a registered trademark of Linus... (5.00 / 1) (#17)
by Tau on Mon Feb 04, 2002 at 03:43:28 PM EST

...with the specific disclaimer that anyone may use it. The only reason he has one in the first place was that some instruction manual company tried to $$$ in by trademarking it and attempting to hijack the name. Linus won the right back and then went and trademarked it.

Besides, for him to close down competing forks by means of trademark strongarming would be very much against the grain of what Linux actually is. If that were to happen I'd get under cover cause the result wouldnt be so much the shit hitting the fan as the sewage duct hitting the hydroelectric turbine (just you wait till RMS hears about it.)

---
WHEN THE REVOLUTION COMES WE WILL MAKE SAUSAGES OUT OF YOUR FUCKING ENTRAILS - TRASG0
[ Parent ]
A 4th posssibility is ... (4.00 / 1) (#18)
by schrotie on Tue Feb 05, 2002 at 08:05:15 PM EST

... that Linu[x,s] makes it a couple more years until finally a new kernel establishes itself. Monolithic Linux is maintanance wise not the optimal architecture for an open source project, technologically not state of the art. And anyway, disregarding anything you might have heard, Unix and thus Linux is one big security nightmare. Granted the other OS is yet worse, but that's no good argument.
There will be some small, secure, persistent, real time microkernel (see e.g. eros) some day. And it will take over eventually, if only cause sysadmins are really sick of cleaning up break ins every week even though they read all security groups daily and act accordingly.
Everything will have to be recompiled though, and many things maybe even ported. But hell, we gotta move west sooner or later.

Thorsten

Linux Kernel Politics | 18 comments (18 topical, 0 editorial, 0 hidden)
Display: Sort:

kuro5hin.org

[XML]
All trademarks and copyrights on this page are owned by their respective companies. The Rest 2000 - Present Kuro5hin.org Inc.
See our legalese page for copyright policies. Please also read our Privacy Policy.
Kuro5hin.org is powered by Free Software, including Apache, Perl, and Linux, The Scoop Engine that runs this site is freely available, under the terms of the GPL.
Need some help? Email help@kuro5hin.org.
My heart's the long stairs.

Powered by Scoop create account | help/FAQ | mission | links | search | IRC | YOU choose the stories!