Allison Randal

The Linux Foundation has released a report estimating the Linux kernel to be worth $1.4 billion, and the Fedora 9 distribution to be worth just over $10 billion. The report is an update of a 2002 report estimating the worth of Red Hat Linux 7.1 (Fedora is the community edition of Red Hat Linux, renamed in 2003). The report doesn't attempt to estimate the worth of the total Linux ecosystem, as some early speculators suggested it would, but instead sticks to the hard facts. The estimation is an industry standard formula-based analysis of current lines of code in the Linux kernel (7 million) and Fedora packages (200 million), calculating what it would cost to develop the kernel and Fedora distribution from scratch under present-day developer salaries and operating costs. The original and updated reports both include the kernel in the calculation of the distribution as a whole, but the new report breaks out a separate figure for the kernel, highlighting its significance within the distribution.

One important perspective to carry away from this report is the astounding pace of Linux development. The 2002 report, which used the same tools for counting lines of code, and the same formulas for calculating total cost of development, listed Red Hat Linux 7.1 as worth $1.2 billion. There was some variation in the Fedora packages used for the the two reports, and for the sake of comparison the current report offers an alternate calculation using the lower operating costs of 2002 ($8 billion instead of $10 billion). Even so, we're looking at a distribution that's worth seven times what it was worth six years ago. Averaged out, that would mean every year sees an investment of time, money, and effort roughly equivalent to the entire history of Red Hat up to 2002.

In acknowledging the community and corporate contributions that made development of the Linux kernel and distributions possible, the report briefly touches on broader implications for open source development in general.

The companies and individuals who work on Linux-related projects and build this value profit by sharing the development burden with their peers (and sometimes competitors.) Increasingly it’s becoming clear that shouldering this research and development burden individually, as Microsoft has done, is an expensive approach to building software. While monopoly position in the past has allowed them to fund this massive development, we believe that in the future competition from collaborative forces will make such an isolated position untenable.

More and more companies are turning to Linux and open source software to radically cut R&D or licensing costs. In the current economy, that trend is sure to increase.

The official word is out, OSCON 2009 will be moving from Portland, Oregon to San Jose, California. We've received significant positive feedback on the move, and messages of welcome from Bay Area open source contacts, but also some messages of disappointment from the local Portland open source community, and from non-local attendees who enjoyed visiting Portland every year. We're also sad to leave, Portland has been an incredible incubator for the conference, as it is for many open source ventures. OSCON was first launched in Monterey, California, and then moved to San Diego, California. In 2003 we moved to Portland, to the Marriott Hotel on the Waterfront, and in 2005 we moved to the Oregon Convention Center on the other side of the river. During our time in Portland, we received incredible support from the local open source and technical communities, and from the Portland Development Commission. In Portland, OSCON grew from 1300 registered attendees in 2003, to 3000 registered attendees in 2008.

Clearly Portland was good for OSCON, but at the same time, we move most of our conferences every few years, to allow new local communities and organizations to participate and to provide new activities for non-local attendees. And while every conference planner likes to see yearly growth as a sign of a healthy conference, it's a challenge to find space for 3000 people plus a projected growth of ~20% (based on previous years). The largest available keynote space at the Oregon Convention Center (the largest conference facility in Oregon) holds approximately 2500 people in our current layout (with a stage and airwall space to divide into smaller rooms for day sessions), and has an absolute maximum limit of 3600 people. 2008 marked a record high of OSCON attendees being turned away from sessions they wanted to see, due to firecode restrictions on the maximum occupancy of the rooms. Even Tim O'Reilly was turned away from two sessions. Given a choice between restricting conference registrations to a pre-set limit (like we do with the Web 2.0 Summit) and finding a bigger space, we knew the right choice for a community-oriented conference like OSCON was to find a bigger space.

We're thrilled to see LinuxCon starting up in Portland, and hope that the energy of yearly Portland-local open source events that grew to complement OSCON will naturally migrate to LinuxCon. Jim Zemlin, who I've known for years, emailed me as soon as an article was released speculating that LinuxCon was started as a reaction to OSCON leaving Portland, to assure me that they had been planning LinuxCon in Portland long before they knew OSCON was moving, and reiterating the Linux Foundation's full and continued support for OSCON. We plan to do cross-promotion between OSCON and LinuxCon, and maybe even cross-conference discounts (if we can work out the practicalities of verifying registration at an unrelated event).

We don't know yet how long we'll stay in San Jose. Some aspects of the space are ideal: it's got a strong local open source community, it's certainly large enough to host us now and for a few years in the future, and it'll be convenient for the attendees and staff to have the conference hotels connected to the session space again. And, with the rising cost of fuel and travel, a huge local open source population in the greater Bay Area is a definite advantage. But, it's hard to tell exactly how good a fit any space is for a given conference until you've actually held the conference there. I'm sure some things about the space will be less-than-ideal, and we look forward to the feedback from OSCON 2009 attendees (both good and bad) to help us make an educated decision on the location of OSCON 2010 and beyond. We also welcome suggestions from open source communities in other cities (with conference facilities to host 3500+ people), as candidates for future years.

Thanks for your part in ten fabulous years of OSCON and open source. Here's to another fabulous ten ahead.

The early registration discount for RailsConf Europe ends on July 15th. You can save €150 by registering in the next week.

One great feature of the new conference website software O'Reilly is using this year (developed by my co-chair Edd Dumbill) is the "Personal Schedule". When you're surfing the schedule or any list of talks you can click the star beside it to add it to a private list. During the conference you can quickly view your list to make sure you don't miss any talks you want to see.

The aggregate data from all the personal schedules also helps the conference organizers. One classic problem in scheduling a conference program is predicting which talks will draw a large crowd and need a large room, and which talks will fit in a smaller room. No matter how hard you try, it's common to end up with at least one session in a small room packed to the gills and overflowing into the hallway, and at least one session in a big room attended by 25 people. With the personal schedule data, we can track the popularity of a talk before the conference and swap the "hot" sessions into larger rooms. I've already moved up "Do You Believe in the Users?" by Ben Collins-Sussman and Brian Fitzpatrick, and am keeping an eye on several other popularity spikes: "Skimmable Code: Fast to Read, Safe to Change" by Michael Schwern, "Top 10 Scalability Mistakes" by John Coggeshall, and "Even Faster Web Sites" by Steve Souders. If you take a moment to mark off the sessions you plan to attend before the conference, you can join in the power of collective intelligence, applied to improve your conference experience.

And don't forget, the Early Registration discount ends next Monday, June 2nd.

In light of Amazon's attempts to lock print-on-demand publishers into their own printing services, I've made a personal decision not to buy from Amazon any more. Since the site first launched over a decade ago, I've spent thousands and thousands of dollars on Amazon feeding my addiction to tech books and fiction, on music, DVDs, electronics, and gifts for friends and family. I realize my spending is a tiny drop in the bucket of Amazon's total revenue, but it's a decision I feel good about, the same way I feel good about using low-energy lightbulbs, reusing plastic bags, and buying a car with environmentally friendly fuel economy and emissions ratings. One of the fundamental principles of capitalism is that when one source of goods and services isn't meeting your needs, you switch to another. The power to decide which businesses succeed and which fail lies in the collective hands of millions of individual consumers.

I've mainly switched to Books-A-Million for the prices (fair disclosure: I developed a good portion of the site back in the heady dot-com days), but I shop around at Barnes & Noble, Bookpool, Powell's, Alibris, BookFinder, and here in South Africa Exclusive Books. There's no shortage of alternatives, all over the world.

In my very first order, I bought some Xhosa language learning CDs, and on a whim added a print-on-demand book of Xhosa folk tales. It just goes to show that by restricting print-on-demand publishers, Amazon isn't only damaging the publishing ecosystem, it's also hurting its own business.

For the 4th year running, Google and O'Reilly will present a set of Open Source Awards at OSCON 2008. The awards recognize individual contributors who have demonstrated exceptional leadership, creativity, and collaboration in the development of Open Source Software. Past recipients for 2005-2007 include Doc Searls, Jeff Waugh, Gerv Markham, Julian Seward, David Heinemeier Hansson, Karl Fogel, David Recordon, and Paul Vixie.

The nomination process is open to the entire open source community, closing May 15th, 2008. Send your nominations to osawards AT oreilly DOT com. Nominations should include the name of the recipient, any associated project/org, suggested title for the award (“Best Hacker”, “Best Community Builder”, etc.), and a description of why you are nominating the individual. Google and O'Reilly employees cannot be nominated.