How do you measure the value of free and open source software (FOSS)? That's a puzzler, because it's, well, free. Moreover, a popular distribution like Linux can incorporate the contributions of thousands of individuals working remotely from around the world. That means that there are almost no associated overhead costs over and above the time of the developers themselves. Even the question itself is a bit of a misnomer, because one measure of the value of FOSS is not the cost to build it, but rather the avoided cost of not having to do so. Because you don't have to pay anything to download FOSS, and since the same project that developed the software will continue to maintain it for you, using free software can allow you to launch products and services that, for economic reasons, you would never otherwise attempt. In an effect that's near and dear to my heart, that means that competition can reenter market niches that had become locked up and stagnant because entry costs to new participants were simply too high. Trying to get a handle on the value of FOSS is still a worthwhile effort, though, because it allows people to appreciate the beneficial effects of FOSS in general, as well as to have greater respect for something that arrives without a price tag. By appreciating the amount of effort that goes into FOSS, not only historically but on a weekly basis, it's easier to appreciate the robustness and responsiveness of the product as well. It also helps anticompetition regulators and legislators appreciate the significant pro-competitive effects that FOSS can have. Today, the Linux Foundation is releasing a report titled Estimating the Total Development Cost of a Linux Distribution that shows just how valuable FOSS really can be, using the Fedora distribution of Linux and the Linux kernel itself as examples. (Disclosure: I am legal counsel to LF.) The effort is particularly interesting as the authors (LF's Amanda McPherson, Brian Proffitt and Ron Hale-Evans) use the same methodology employed by respected industry expert David A. Wheeler in 2002 to value a related Linux Distribution (Red Hat 7.1). The run up in value wrought by six additional years of global collaboration is an eye popper.
Just how big a run up is that? Wheeler concluded in 2002 that it would have cost $1.2 billion to develop Red Hat 7.1 using typical proprietary means. Today, McPherson, Proffitt and Hale-Evans price it out as follows:
Using 2008 salary figures, the tests published in the paper revealed that if developed today, the full set of Fedora 9 distribution packages would cost $10.8 billion. The Fedora 9 distribution contains 204.5 million lines of code in 5547 application packages. The development effort estimate comes close to Person Years.
The paper makes interesting reading from a number of other perspectives, but it’s hard not to pause on some of the numbers themselves. Here is some more data on the Fedora 9 release:
Over 1,000 developers, from at least 100 different companies, contribute to every kernel release. In the past two years alone, over 3,200 developers from 200 companies have contributed to the kernel.2 It’s important to note that the kernel is just one small piece of a Linux distribution. A distribution is actually made up of multiple components including the kernel, the GNOME and KDE desktop environments, the GNU components, the X window system, and many more. The total of individual developers contributing to these projects surely numbers in the thousandsu
How about just the Linux kernel? Here’s the summary, measured in the same fashion:
Applying this test to the Linux kernel included in Fedora 9 found the value to be 6.8 million lines of code worth $1.4 billion. The development effort estimate for the kernel alone exceeds 7500 Person Years.
So how about those avoided costs? The value of FOSS is being particularly appreciated in the rapidly expanding mobile space. Here are some current examples of the impact of multiple Linux-based platforms being made available:
- Small but nimble platform companies like Asus have created the new, low cost, mini-laptop form factor. In its cheapest mode it runs on Linux and pre-loaded with other FOSS, such as OpenOffice.org
- Companies are creating products totally outside their historical competencies, such as Amazon’s Kindle reader, and the new Android mobile platform launched and supported by Google
- Mobile phone vendor Nokia has opened openedup its Symbian operating systems to developers
- One of the largest handset manufacturers (Taiwan’s HTC), which until now only sold its wares to others to sell under their own brands, is now launching its own branded products – based on the new Android platform.
In each case, it is the free OS that makes it all possible. And in many cases, the decision to go open was also driven by awareness of the further free benefits that FOSS can provide. That benefit, of course, derives from the hordes of independent software vendors that will take advantage of an open platform and a huge market to create own products and services to their own advantage. That makes the host platform, in turn, much more attractive to customers, and the virtuous circle continues to turn for the benefit of all.
The lessons, then, are clear: the benefits to be achieved through the FOSS development process can be huge. Not only does this method help vendors share costs through collaborative benefit, but it reopens old, consolidated market niches to new competition, and allows a wealth of innovative new companies, and even individual developers, to create new products and services in what can only be called an explosive fashion. The result is more choices, lower costs, greater innovation, more rapid technological progress, and a healthy and efficient marketplace.
Which, ahem, is something that we would all like to hang on to during these current turbulent times.
There is, of course, much more to be found in the report, which you can find here. The LF press release can be found on line here, and as usual, I have reproduced it in full below for archival purposes.
For further blog entries on Open Standards and Open Source
sign up for a free subscription to Standards Today today!
* * * * * * * * * * * * * * * * * * * *
Linux Foundation Publishes Study Estimating the Value of Linux.
New report finds the value of developing a Linux distribution to be worth $10.8 billion
SAN FRANCISCO – October 22, 2008 – The Linux Foundation (LF), the nonprofit organization dedicated to accelerating the growth of Linux, today announced it is publishing a new report written by Amanda McPherson, Brian Proffitt and Ron Hale-Evans on the value of Linux development. The paper finds that it would take approximately $10.8 billion to build the Linux community distribution Fedora 9 in today’s dollars with today’s software development costs. It would take $1.4 billion to develop the Linux kernel alone.
The report, titled “Estimating the Total Development Cost of a Linux Distribution,” is available today.
This report is an update of a 2002 study done by David A. Wheeler that examined the Software Lines of Code (SLOC) present in a typical Linux distribution (Red Hat Linux 7.1). At that time, Wheeler found that it would cost over $1.2 billion to develop a Linux distribution by conventional proprietary means in the U.S.
The authors examined the Fedora 9 distribution using Wheeler’s tools and methods, specifically the SLOCCount tool that estimates value and effort of software development based on the COnstructive COst MOdel (COCOMO). The report goes into detail on the methods used, how they specifically apply to the Fedora distribution and the Linux kernel, and what an estimate of Linux’ value really means.
Highlights of the paper include:
– How Much Does a Full Distribution Cost?
Using 2008 salary figures, the tests published in the paper revealed that if developed today, the full set of Fedora 9 distribution packages would cost $10.8 billion. The Fedora 9 distribution contains 204.5 million lines of code in 5547 application packages. The development effort estimate comes close to 60,000 Person-Years.
– How Much Does the Linux Kernel Cost?
Applying this test to the Linux kernel included in Fedora 9 found the value to be 6.8 million lines of code worth $1.4 billion. The development effort estimate for the kernel alone exceeds 7500 Person-Years.
– How Does This Really Measure the Value of Linux?
This study reveals that collaborative development creates enormous economic value. In the past two years alone, over 3,200 developers from 200 companies have contributed to the kernel. An even larger number has contributed to full Linux distributions. Measuring the economic effort involved is imperfect, but this report clarifies why the methodology is the best approach and some of the limitations.
”This year has seen an incredible proliferation of Linux-powered devices outside of traditional Linux strongholds: netbooks like the eeePC, mobile phones like Android and the Gphone, and consumer devices like the Amazon Kindle. Would these products be possible without Linux?” said McPherson. “I think this points to the power of the collaborative development model. Monopolistic software companies used to be able to fund heavy R&D budgets, keeping out competition. Given the cost associated with building an OS like Linux, one wonders if proprietary companies will ever go it alone again.”
Amanda McPherson is vice president, marketing and developer programs, at the LF and leads its promotion, developer, and community-relations activities. Brian Proffitt is community manager with the LF, managing the Linux Developer Network. Ron Hale-Evans is senior specifications writer with the LF and works closely with the Linux Standard Base (LSB) developer team to create LSB specifications.
About the Linux Foundation
The Linux Foundation is a nonprofit consortium dedicated to fostering the growth of Linux. Founded in 2007, the Linux Foundation sponsors the work of Linux creator Linus Torvalds and is supported by leading Linux and open source companies and developers from around the world. The Linux Foundation promotes, protects and standardizes Linux by providing unified resources and services needed for open source to successfully compete with closed platforms. For more information, please visit www.linux-foundation.org.
###
Trademarks: The Linux Foundation and Linux Standard Base are trademarks of The Linux Foundation. Linux is a trademark of Linus Torvalds. Third party marks and brands are the property of their respective holders.
If you want an apple tree, you can either build a strand of DNA molecule-by-molecule until you have something that will replicate and grow; or you can find a ‘windfall’ apple and plant that.
Finding ‘windfalls’ avoids a lot of costs and brings competitive benefits.
But there are lots of ways to get an apple tree. I don’t think any one way will exclude any other.
Andrew,
A good comment. But after being tweaked by you on so many minor word choices, I can’t resist: do you want to reconsider the gender selection in the phrase, "a gaggle mostly consisting of idealistic young _men_?"
While that statement may be arithmetically true, (any proportion comprising more than 50% can, I guess, be accurately described as "mostly" that), last time I checked there were an awful lot of very talented young _and older_ women sitting at keyboards as well.
Just trying to keep you on your toes, and thanks for the comment.
– Andy
Appreciated 🙂
I wondered about that whole statement actually – RedHat is hardly a gaggle, Oracle is hardly idealistic. Since you offered, maybe it would be better to go for “a community dominated by idealistic young men with a flawless grasp of English”, as those are the people whose innovative ideas tend to get traction. While the groups I’ve just offended are no less smart or hard-working, the community could do with some optimisation around letting those people achieve their potential in positions of leadership.
– Andrew Sayers
P.S. if you think I’m hard on word choices by non-me people, you’re … hmm… misinformed? No, wrong… underinformed? Less wrong… you ain’t seen nothing yet? No, wildly inappropriate. Let’s just say that there’s quite a gap between when I finish writing the content of a post, and when my disdain for its minutiae subsides to a level where other people are allowed to see it.
Thanks for the pointer to that rule.
I believe that the Sixth Edition was adopted a few years ago (2007?), so it sounds like this is a rule that would have been in force at the time. That said, though, one of the problems was that many National Bodies either joined, or upgraded, just before the OOXML vote. As a result, any rule that would eventually take an NB out of voting consideration would require a string of votes before the rule kicked in. Until that happened, SC 34 would still be out of luck.
– Andy