I've spent the last couple days recuperating from the sleep deprivation during OSCON. While I had an absolute blast in Portland, it has been really nice sleeping in my own bed (as late as I want) the last couple nights. In any case, before I go back to work tomorrow I figured I'd better do the write-up for the final day (Friday) at OSCON.
I actually woke up late and missed the first keynote of the day which was unfortunate since I heard it was actually really interesting. However, when you've been up until 4 - 5 AM every night for three nights in a row and set your alarm for 8 PM instead of 8 AM, you're going to wake up late. Fortunately, I made it to the convention only halfway through the second keynote. Unfortunately, the Ximian (now Novell) guys were right last night. The presentation was really boring. Apparently Novell had a different presenter lined up but something fell through and they had to go to a back-up. Or something like that. Like I said, I showed up in the middle and almost fell back asleep.
After the keynotes were over I went to a very poorly attended but interesting presentation from Edward Moy from Apple. He started by describing his history, how he worked into Berkley then Parc then Apple, which was interesting if not a bit narcissistic. The second half of the presentation was far more interesting when he actually started talking about software management, a topic near to my heart.
Edward said two things that I found extremely amusing. First, "exhaustive analysis of software management is difficult and often fails." With more explanation it became clear that Edward was noting that there are lots of rules around software management and cataloging them has never proven successful. Not sure I completely agree with that, but it certainly sounds a lot like my First Axiom of Setup. The second point Edward made was that "new issues always pop up." I seriously started chuckling when he said that. Here was an engineer from Apple talking to a room with about six or so geeks in it about the Second Axiom of Setup. I was tickled to get such validation of my personal theories.
From there Edward walked into how they were building on XAR, Arusha, YAP, and ZODB and how he would go in to add the appropriate features as necessary. This process absolutely fascinated me. He was improving Open Source projects because that was helping him get his job done. Oh, how I hope one day Microsoft reaches a point where we can partake in projects at this intimate a level.
After Edward's talk I went to a presentation titled "Open Source: Economic Nonsense". Kent Beck made a bunch of interesting and sometimes contrary statements trying to stir up the idea that somehow there is a way to make money while working directly on Open Source projects. I can't say I remember most of his points (they weren't really hanging together for me), but I was amused when he tried to argue that working on Open Source works the opposite from Maslow's Hierarchy of Needs. In other words, Open Source developers were already "self-actualized" and were working down the pyramid to get improve their esteem (respect of others), find love (others really respect and maybe a gf), and eventually work themselves into a shelter, food, clothing. It is an interesting theory.
There was a short break while I listened to r0ml and Stephen discuss the "Economic Nonsense" presentation. R0ml seemed to think Kent had borrowed a couple pages from Stephen's playbook and they went along discussing that. Soon enough the main ballroom was set up again and we went in for the last keynote of the day.
Stephen had told me that last year Milton's presentation behind the scenes of Lord Of The Rings (LOTR) was amazing. Well, some of the numbers Milton was throwing around--like the number of rendering machines going from 500 for the first movie to 1400 for the second to 3300 for the third--were interesting it was a very impassioned presentation. However, right in the middle of his talk Milton started a rather scathing discussion about RedHat's latest service agreements. Apparently, the price point for RedHat is not as attractive with their latest changes, especially when you have as many machines as Milton does to render LOTR and future projects. Milton also noted that switching off of RedHat wasn't necessarily an option because independent software vendors (ISVs) want to minimize the number of Linux platforms they certify on. RedHat is obviously one of the few vendors that ISVs consider. So, Milton is finding that it is very easy to end up locked-in to a certain Linux vendor if you want your systems supported.
It was fascinating to see someone from a company who only uses computers because it makes their day to day business functions easier or, in this case, even possible so clearly describe the problems that platform companies (like RedHat) encounter when the platform becomes successful. It'll be interesting to see how RedHat manages their future now that they are starting to have similar problems faced by companies like Microsoft and Apple for years.
So, that's it. There is my "trip report" of OSCON 2004. Maybe I'll go back and add some of my smaller insights over time but I feel like I've captured the general feel of OSCON for me in the last few blog entries. I hope you enjoyed the trip as much as I did. Now it is time to get back to talking about the Windows Installer XML toolset since 2.0.1927.1 was released late week with tons of cool new stuff in it.
If you have questions or comments about OSCON, please feel free to leave them here. If there are enough interesting questions I'll definitely reach back into the last week and try to pull out some memories. Until then, keep coding, you know I am (now that I'm home again).