Who Wrote, and Paid For, 2.6.20

Corbet writes "LWN.net did some data mining through the kernel source repository and put together an analysis of where the patches came from. It turns out that most kernel code is contributed by people paid to do the work — but the list of companies sponsoring kernel development has a surprise or two." The article's conclusion: "The end result of all this is that a number of the widely-expressed opinions about kernel development turn out to be true. There really are thousands of developers — at least, almost 2,000 who put in at least one patch over the course of the last year. Linus Torvalds is directly responsible for a very small portion of the code which makes it into the kernel. Contemporary kernel development is spread out among a broad group of people, most of whom are paid for the work they do. Overall, the picture is of a broad-based and well-supported development community."

Comments

Define "volunteer."

[Rob T Firefly]

TFA: "It is not uncommon to see Linux referred to as a volunteer-created system, as opposed to the corporate-sponsored, proprietary alternatives. There has been little research, however, into how much work on Linux is truly 'volunteer' - done on a hacker's spare, unpaid time. In general, the assumption that Linux is created by volunteers is simply accepted."

Thing is, even though some of those changes were done by programmers in the course of their paid jobs, isn't the work still being "volunteered," albeit by the company rather than an individual? As companies, Red Hat, IBM, Novell, or Big Roy's Heating and Plumbing don't need to help improve the kernel, nor are they directly paid for their work on it. They simply do so because a better Linux kernel does benefit them directly or indirectly, as do many individual volunteers.

Re:Define "volunteer."

[krlynch]

So the work is not volunteered, it is a component of an agenda.

Doesn't the same argument apply to non-monetarily-compensated "volunteers"? Don't they have an agenda as well?

Re:Define "volunteer."

[prelelat]

I don't agree with your standpoint. I do think that IBM Oracle and other companies benifit from this kind of program. On the other hand most people benifit in some way from contributing code to linux in the first place. People use it for experiance to get a job, to make the OS that they run better, to be apart of something and make themselves feel better. Just because a company is volenteering programmers to the cause because its benifiting them doesn't mean its not volenteering. Its like saying donating to linux because you want it to work better for you so that you can produce more money is not really a donation.

Most people donate, volenteer for something because they know it will benifit them in the end(how many people at Harvard who have volenteering on their application to the school volenteered because it was something they wanted to do, I would guess half does that make their time in a soup kitchen less valuable or appreciated?). This doesn't mean that its any less noble in the end.

Re:Define "volunteer."

[morgan_greywolf]

Exactly. Most of the driver development in the early days were done by people who -- gasp, shock -- had that particular piece of hardware and needed it to work with the Linux kernel. Much driver development is *still* done that way, although some driver work is now sponsored by companies who develop the hardware (i.e., Broadcom)

Most everyone working on the kernel has an agenda and that's okay -- open source isn't about communism or pure philanthropy, it's more of a libertarian or anarchocapitalist philosophy.

Re:Define "volunteer."

[VWJedi]

The term "volunteer" in our culture generally carries implications of altruism rather than self interest.

If you put it that way, no one is a "volunteer developer" for linux. They write / change code for their own benefit (to add features, improve functionality). Once they've finished, they usually give their code to "the linux community", but the reason they do the work in the first place is because they want to fix / improve the way their system runs.

Re:Define "volunteer."

[xappax]

Most everyone working on the kernel has an agenda and that's okay -- open source isn't about communism or pure philanthropy, it's more of a libertarian or anarchocapitalist philosophy.

Lots of geeks are anarchocapitalists, so it makes sense that they'd want to claim the successful and popular open source movement as their own, but I don't think they're as similar as you assert.

Anarcho-capitalism is about profit and individual property as the central pillars of society. Open source is not about profit, and it's definitely not about private property.

Open source is a tradition that was established to fight back against those who sought to profit from proprietary computer code. It was introduced as a way to foster cooperation and support between those programmers who didn't seek to profit from their code, but did want to share it with other like-minded people. Open source has become so successful that entire profit-making industries have come to depend on it, but at its core Open Source is designed as a sort of "non-profit cooperative" for people who code for free. Open source is a gift economy - sure everyone gives gifts for different reasons, but they're still gifts.

The open source philosophy is also clearly against private property. Of course, the only form of property that open source involves is intellectual property, which many anarcho-capitalists claim is a special case, but I think the point should still be made that nobody owns open source code, and nobody can own it. Since private ownership of everything is a central tenet of anarcho-capitalism I can't see where the similarity is.

I know socialism is a bad word on Slashdot, because it means red commie soviets who are going to take away all our civil rights and make us live like in 1984, but personally, I see the open source movement as an example of voluntary socialism, or anarcho-socialism - programmers have decided that the existing market forces are abusing their property rights to producing crap software for ridiculous prices. So, they have voluntarily formed a network which allows them to share their resources in a non-market environment.

The reason open source software is so good is precisely because it's not driven by profit-oriented market forces, but by the diverse motivations and interests of many people and organizations. Obviously they're not doing it out of pure generosity, but in general when people develop open source code they're considering how to make good code primarily, not how to make lots of profit primarily.

Funding...

[Needs Food Badly]

It's really quite interesting the amount of funding that is sent in the direction of the devs working on the Linux kernel. I'm curious what would happen if the funding was spontaneously cut. Linux was built from scratch and supported for free back in the day, but would the main developers continue to work or even be interested at all if they weren't being paid?

Quite a paradox

[L. VeGas]

...the list of companies sponsoring kernel development has a surprise or two.... a number of the widely-expressed opinions about kernel development turn out to be true.

So... the surprise is that there is no surprise?

GPL vs. BSD

[ThePhilips]

Overall, the picture is of a broad-based and well-supported development community.

It is just confirmation of old statement that GPL(v2) provides better (at moment best) ground for cooperation between vendors.

Many companies are willing to control what OS does with their software and hardware - and Linux gives them that chance on cheap. But even more so, GPL allows Linux to "merge" back possible code base "forks". That's next to impossible with BSD licensed code most tend to keep closed.

Let's just hope Linux would be able to go on surviving the "snowball" effect of the merges.

Re:GPL vs. BSD

[Anonymous Coward]

Why can't you merge changes back to BSD-licensed code?

Because when Microsoft makes ftp.exe using BSD-licensed code, all they have to do is tell people that it had BSD-licensed code, not give them the code nor the changes they made. Or, take a look at the various commercial forks of postgres that add replication, live backups, or whatever. Sure, it'd be nice if they gave these features back to the postgresql database server, but the developers chose the BSD license knowing that the people who do stuff with the code don't have to give back.

This is why the GPL makes code Free, while the BSD license makes programmers Free.

Re:GPL vs. BSD

[jimicus]

Why can't you merge changes back to BSD-licensed code?

Nothing. But if I merge changes into a piece of BSD-licensed code, there's nothing to stop my competitor 20 miles down the road downloading this piece of code complete with my changes, making a few minor tweaks to it, keeping those tweaks private and selling it - even though I provided a lot of the code which his product is based on.

With the GPL, he can still make tweaks to it but he's got to make those tweaks freely available to anyone he distributes code to. Many choose to simply submit these tweaks as patches upstream rather than maintain their own fork of the software - not really a lot of point in being that anal when the license explicity allows your customer to do that anyway.

What we're seeing happening now - particularly in the embedded space - is that manufacturers are taking the free stuff in Linux, tweaking the kernel and submitting changes where necessary but keeping the majority of their proprietary code logically separate in userland so they don't have to GPL it. Hence why you can have a router based on Linux which is technically open source, but the clever stuff (eg. removing the complication from configuring iptables with a web app, a means of holding firewall rules and some glue to turn these into iptables commands) remains private.

Re:Fairly Interesting Overview

[LWATCDR]

Why is it shocking. Red Hat, IBM, and Novell hope to make a lot of money from Linux.
Then you have the expensive systems that use Linux
Intel and HP are still hopping that the Itantium will work out in the end and frankly Linux is the big OS for the Itantium. Not to many hobbiest have an Itantium sitting around so Intel and HP probably contribute a lot of code for the Itantium port.
IBM sells a lot of Power systems that run Linux so they probably contributed a lot of code to support the new Power6. Not to mention the the 360/370/Zmachine port.
Then you have Mips contributing for the embedded market.
Linux is now big business.

Re:BDFL

[JensenDied]

where does open source mean unpaid?

Interesting idea, but hardly accurate

[Critical Facilities]

From TFA

Finding an answer to that question is somewhat trickier than looking at who wrote the patches, mostly because very few developers say "I wrote this on behalf of my employer." The approach taken by your editor was relatively simplistic, but, perhaps, the best that is practical. Any patch whose author's given email address indicates a corporate affiliation is assumed to have been developed by an employee of that corporation. So any patch posted by somebody with an ibm.com email address is accounted as having been done by an IBM employee.

While I still find the result interesting, and while I also would like to know which organizations contribute the most to the kernel, I don't know that this method is really a good way to reflect whether the work was done in a "sponsored" fashion.

That is, just because someone's email address shows that they're from IBM, doesn't necessarily mean that they were being paid by IBM to explicitly work on the kernel. For all we know, they might have "15 minutes of real, actual work" like this guy and are just hacking away in their cube because they're bored. Maybe not, but still, for he purpose of determining which companies contribute most (or the individuals motivations to contribute), that seems like a shaky method of proving or disproving things.

This is the DARKEST secret!!!

[erroneus]

...The darkest secret that Microsoft and other software companies don't want the rest of the world to know.

Companies are shelling out billions of dollars each year just to run some software that needs to be renewed, updated and purchased again and again and again.

Some companies are investing their workers or their donations into the community software projects because in some way, it will truly benefit them in a way that will not expire the way proprietary software does. So when people start noticing that businesses do more than just "use" F/OSS, but they contribute to it in a way that makes it more usable for themselves. And depending on the way they contribute, they can also write off some on their taxes as part of a tax strategy.

So companies can spend their software budget in a way the keeps them locked in and paying ridiculous annual fees and subscriptions, or they can actually pay to get the software they actually want in the way they want it, benefit themselves, benefit the public and even build a lot of good will in various communities.

I am hopeful to see the rest of the F/OSS revolution in my life time...

The analysis is broken

[EmbeddedJanitor]

These are only the submitters, not necessarily the actual authors of the changes.

Many patches are fed in through email lists etc where the maintainer (more likely to be a "named person") picks it up and pushes it upstream. I expect many volunteers will be in that group.