ApacheCon US 2006 kicked off its general session this morning in Austin, Texas, following two days of tutorials. Apache Software Foundation (ASF) president Sander Striker opened the proceedings with his "State of the Feather" address. Cliff Stoll, the hacker-catching, planetary astronomer, author, and volunteer 7th grade science teacher, followed Striker with a keynote address which included a demonstration of how he taught a 7th grade science class to measure the speed of light.ApacheCon is small compared to commercialized conferences like LinuxWorld or Black Hat. The number of attendees at this one is less than 500, according to Sally Khudairi of HALO Worldwide. The crowd is a mix of Apache hackers, users, and others involved in the ecosystem surrounding the dominant HTTP server on the planet. Size may be one reason the sense of community is so strong here compared to more commercialized shows I've attended the past few years.
Of course, commercial firms are here as sponsors and exhibitors, and as users and participants in the various open source projects. Before the first speakers this morning, Google, Sun, Covalent, Simula Labs, and others had staffed their booths in the exhibit area. Novell and IBM were present, too, though I didn't see their booths.
I asked Dave Johnson at the Sun booth when they were going to open source Java. "Soon," he replied quietly. I couldn't weasel any more information out of him, though I tried my best. I did hear an announcement of an open meeting between Sun and Apache about licensing issues around Java, however, so it really might be happening sooner rather than later.
Attendees were still straggling into the hall as Striker went to the podium to give his "State of the Feather" address. His presentation detailed the major events of the past year within ASF, including projects moving up to become one of the ASF's 37 TLPs (top level projects), as well as those just entering the ASF fold by joining the Apache Incubator.
As Stoll warmed up the crowd during his keynote, he mused briefly about why Apache had asked him to speak. His books, Silicon Snake Oil, and High Tech Heretic, demonstrate that Stoll is not the biggest fan of the Internet, or computers, out there. It must be something in the air, because Stoll's talk reminded me very much of Nicholas Negroponte's remarks earlier this year during his talk on the OLPC project at the Red Hat Summit.
Stoll is very much the teacher, and an activist at that. He roamed the room constantly, seeming to be stalking the outlier geeks on the edges of the classroom, letting them know that they should be up in front where they can really see what's going on, urging them to join the action.
Stoll's keynote had more action than any I've seen lately. Stoll is a non-stop human dynamo, leaping here and there, getting on and off the podium, moving a chair over in front of one of the two large displays projected against screens at the front of the room, so he could point at something with his finger instead of simply aiming a laser pointer at it, and the aforementioned circling action.
He started off with a blowgun and a stuffed monkey, demonstrating the physics involved in simultaneously firing a "dart" from the blowgun and dropping the stuffed monkey from where it was suspended twenty feet away, and still managing to hit it with a dart. Don't ask me any questions about it, I'm an English major, but it had something to do with horizontal velocity and vertical acceleration. But it was fun.
But that was just the warm-up. He explained how he wound up teaching a 7th grade science class for a school where he lives. He agreed to do it under the following conditions: he could teach whatever he wanted, however he wanted. What he wanted was to teach the kids how to measure the speed of light.
The demonstration for this project included a laser sending pulses of light over two different mirror-controlled routes: one the long way, one the short way. It also included an oscilloscope with two inputs, one for each of the light paths. Once everything was properly set up and running, it was as simple as measuring the distance between the tops of the sine waves displayed, and explaining that horizontal axis on the oscilloscope measured time, so now we knew both the distance each path took, and how much longer it took over the longer path, so we could easily compute the speed of light.
I'm glad I was there. Now I understand both what Stoll's complaint is about computers in schools, and what he suggests would be better. Stoll argues that the computers are bad because the students use them primarily to play games, rather than learn, so it's better to expose them to teachers who teach. If all teachers were as gifted as he is, I would agree wholeheartedly. But that's no more true than the assertion students use computers only to play.
The business of open source
Khudairi gave each panelist a chance to describe their own history in open source and how they have used it since entering corporate life, then made the talk interactive by asking questions of them and inviting audience members to do the same.
I found Thompson's story of migrating their datacenter from Sun to Red Hat Enterprise Linux the most compelling tale of the three. Thompson says that not only was it one of the largest, most visible migrations to open source in history, it was also a very successful one. Successful not only in terms of cost savings, but in performance and quality as well.
I also appreciated Damarillo's observation that it was a little harder to make a profit selling open source than it is selling closed. In my opinion, that fact underscores that the value of open source is not nearly so much that you can make a mint selling it, but in that you can make more selling whatever it is that you sell by using it.
Brian Behlendorf: Bringing open source to the enterprise
Brian Behlendorf has been part of the Apache and open source community from the beginning. The crowd that gathered to hear him speak on "Bringing Open Source Software Development Processes and Principles to the Enterprise" this morning wasn't large, but it included Striker, the current ASF president, and others from Apache's core team.
The topic is near and dear to his heart, especially since Collabnet, which he founded in 1999 and where he serves as CTO, does exactly that.
Behlendorf spent a little time bad-mouthing what has become traditional software development methodology. If you're a professional developer, you can probably relate to the negative effects of project management styles where the project itself is secondary to the project management application producing all those Gantt charts and requiring daily input from everyone involved. Behlendorf offered a single, chilling tidbit of data from the Burton Group to show the state of traditional software development today: it has a 75% failure rate.
He also addressed a real problem still being encountered today in bringing open source into the business place, showing a poster equating open source software development with communism.
The key difference between closed and open source development is, Behlendorf noted, the transparency. Not just in the code, but in the design decisions as well. The transparency in turn leads to a progressive growth of involvement, from new user to patch submitter, within a project. He also noted that while the right to fork open source code is often seen as something to be feared, it actually forces a change in management style which seeks to find and build consensus rather than forcing the will of a leader on everyone else.
Final observations and still to come
I saw more Apple laptops today than I've ever seen in my life. They are everywhere at ApacheCon. Secondly, and this is a very good thing, I didn't hear or see mention of Web 2.0 a single time today. ApacheCon continues through Friday, and based on what I saw today, I'm looking forward to it. Keynotes, presentations, BOFs, and parties: it's all good.