Today's (note, today is still Wednesday for me) events massively made up for the lack of excitement in yesterday's blog entry. Unfortunately, today started at 7:30 AM and ended about 2:30 AM. Tomorrow starts at 7:30 AM again because I'm definitely not going to miss the three Dysons' keynote. So, I'll write as much as I can right now and then pass out for another exciting day.
Tim O'Reilly kicked off the morning with his keynote. I took a bunch of notes during his talk but I was never really engaged. Tim's explanation that the O'Reilly company isn't so much about publishing books but "capturing the knowledge of innovators" by following the "hackers" around was interesting. And his discourse about the dangers of data getting locked into proprietary systems seemed rather relevant. But, honestly, the thing I found most interesting in his talk was the quote from William Gibson, "The future is here. It's just not evenly distributed yet." I don't remember exactly how that quote pertained to anything in Tim's speech but I really enjoy Gibson's work so the quote piqued my interest.
On the other Robert Lefkowitz, better known as r0ml, gave an incredibly fascinating and amazingly well crafted keynote that I really feel at a loss to even attempt to capture it. But, try I will. R0ml started by explaining the title of his talk "The Semasiology of Open Source." He defined "semasiology" (with help from Webster's) as "the science of development of words". He went on to liken the word "denotation" to the "past meaning of a word" and "connotation" as the "future meaning of a word". This is a fascinating analogy and I was completely hooked.
R0ml then used a series of clips from The Princess Bride where Vizzini keeps using the word "inconceivable" to describe acts that were occurring in front of him until Inigo Montoya says, "You keep using that word. I do not think it means what you think it means" to point out the discrepancy in the way people sometimes use the words "open source". R0ml also pointed out that the word "patent" in the legal sense actually means "open for inspection." "Open" is hardly the connotation the Open Source community usually seems to use with the word "patent". R0ml continued to use ever increasingly detailed examples from many aspects of life (his comparison to some Open Source vendor's licensing agreements to how Muslim's handle mortgages [or rather the lack thereof] was incredibly fascinating) to truly drill into what "Open Source" means. I'm tired and can't make all of his arguments, but r0ml systematically worked his way down to the conclusion that "Source" == "Requirements". I'm not doing the presentation justice (I've tried for the last hour) but trust me it was fantastic.
After that, I attended a couple sessions that ended up being very disappointing. The first was supposed to be about the Gentoo Linux distribution and its Portage package deployment system. Unfortunately, the presenters got caught up with some technical difficulties and postpone enough random questions so they ran out of time before ever demonstrating Portage. That was terribly disappointing.
The second session turned out to be something of a lecture where the speaker introduced some interesting assertions that related software development to the theory of evolution. My problem with the presentation was that he assumed everyone would take his assertions as fact and proceeded to build a huge number of "statements of truth" on the assertions. While I found his assertions interesting, I think they were flawed in a few ways and I spent the entire presentation dissecting his truth statements and coming up with completely different results. Again, another terribly disappointing presentation.
Over lunch, I ran into Mike Hunter who was a fellow UMR graduate. We talked for a while and caught up. Mike did a great job keeping me distracted from the anxiousness of presenting my own session. After lunch I really wanted to go to Miguel de Icaza's session on Mono but decided it would be more prudent to spend the time getting into "presentation mode".
After much anticipation, it was finally my turn to present. One of my major concerns was that no one would show up to the session. As it was, there were about fifteen to twenty developers in a room that sat maybe sixty. I was pretty happy with that turn out.
The presentation went along fairly smoothly. I gave a short description of the Windows Installer (namely, it installs stuff on Windows), a short description of the Windows Installer XML toolset (namely, it builds MSI files from source code), and then did a demo. I actually deviated from one of the suggestions on my blog and did a demo where I used some .wxs files from another demo to install a Windows application, a Windows service, and a Web service with a nice little UI. The demos actually went really well and I finished off why the Windows Installer XML toolset was built and why it was released as Open Source.
I finished up my presentation by noting that the entire demo had been done with the WiX toolset running on Mono. The look of amusement on everyone's face in the room was priceless. I'll provide more instructions how to build the WiX toolset to run on Mono in a future blog entry. It really is quite easy although there are a few things Mono can't do. I spent a bit of the afternoon chatting with Miguel about the issues and he acknowledged that some COM Interop scenarios just aren't supported on Mono. That was a fun conversation. I've got a lot of respect for the guys from Ximian (now Novell), more on that below.
The last session I attended today was about SQLite. I like the idea behind the small embeddable SQL engine so I thought I'd drop in and listen. Not a terribly interesting talk except for when the presenter mentioned that AOL contacted him sometime late last year about getting Unicode support done. He said he expected it would take until mid-2005. AOL asked what it would take to get Unicode support by July 1, 2004. He told them and apparently AOL (in one way or another) funded the project to get Unicode support into SQLite last month. I found that kinda' interesting.
After all the presentations were done, Jason Matusow, Stephen Walli, Jon Rosenberg, and I went out for dinner. We discussed all kinds of things related to life, OSCON and SSI. Once again, it was great food and great company. Little did I know the night was just beginning.
After dinner, Stephen, Jon and I went over to crash the "Stonehenge" party at Bar 71. There, Stephen introduced me to the ActiveState developer (I feel terrible, I've forgotten his name) who wrote the MSI generation code for ActiveState's Perl implementation. He and I had a great little conversation about the Windows Installer, deployment, and life in general. That was fun.
Then, I found Stephen talking to Miguel de Icaza, Eric Dasque, and Pete Goodall (all from Ximian, now Novell). The rest of the night, the group of us just sat around drinking and discussing Open Source, Microsoft's relationship to Open Source, Ximian, Ximian's acquisition by Novell, and where all the good pubs are in Boston. I really like these Ximian guys.
It was a truly great night.