You can also see the photos from our trip to
Thailand, or our sailing adventure around the
Virgin Islands. You can also take a look at some of the
articles Ive thrown together about politics
and technology. Beware, I'm considered somewhat opinionated...
Tue, 03 Aug 2004
Final Thoughts on OSCON 2004
As usual, some of the best talks were on Friday, to get everybody
to stay at the conference as long as possible. The keynote about
the David Rumsey map archive was unbelieveably cool. It recieved
a well deserved standing ovation. The security talk was excellent
as well, and the
best practices help solidify my resolve to evangelize their product
in our company...
Overall, the whole worthless 'Emerging Technology' jive that was so
at ETCON 2003
wasn't as prominent. This conference was about real technology, and
the presenters were not just some slick sounding snake-oil salesmen
looking for venture capital. They were actual programmers with
actual code. How novel.
Another thing that is becoming more and more obvious is that
the 'Alpha Geeks' dont really seem to care about Java anymore.
The panel with Eric Raymond, and Tim OReilly was not well attended.
There wasn't a single Java talk on Thursday, and the rest of them
were only partially about Java. The more Sun tries to clutch onto
control, the more people will move away from it. That's why Im
becoming a Python hacker.
What stuck Pete as funny was how political this place can be...
any time somebody is presenting new technology, they need to explain
why its needed. The answer is either because nothing like it ever existed,
or because what does exist is crappy and wrong. Well... the problem with
the latter is that the guy who wrote the
'crappy and wrong'
software is most likely in the audience. There are two kinds of
presenters - those who walk on eggshells, and those who suffer from
angry glares. Even Microsoft bashing is a little tough, because they
are a huge sponser of the event! On Open Source. Double plus wierd.
In closing, here are a handful of links to the tech goodies that
I learned about at the conference:
OSCON 2004 -
the home page for the conference. The presentations should be online
Todays talks were quite good. I heard some cool stuff about the
state of Python 2.4, and how great a language it is for
The talk about
Tsearch2 and POSGRESQL was interesting... although I dont
think its quite ready for us to suggest to customers. The talk on
PHP performance was a good high-level talk, and I picked up a few
new Apache performance tricks.
I picked up a few things about whats currently available in
the Python lightning talks, and about what's next at the state of
the Python union talk. Its a big Python love-fest. Making
something like CPAN for Python
would be a good idea to help the community.
I also had the opportunity to meet Jeremy White, my boss's
brother. He's one of the lead developers on the
Wine project, which allows
people to run Windows apps on Linux. Quite an important and
incredibly difficult project indeed.
The Dyson trio keynote was interesting. Its cool to see a great
Freeman Dyson talk about topics like the future of technology.
Although he's a bit of a science worshipper for my taste. If the
purpose of science is to serve mankind, then humility should be
considered a virtue on par with innovation. Just because we know
our way around a research lab, doesnt mean we always know what's
best for the world. Nevertheless, I was able to get a
photo of the two of us.
The second talk was pretty good as well. Bdale Garbee reinforced
the idea that its hackers and inventors that push the envelope
of science the best. People like Tesla, Goddard, or the Wright
Brothers did things that everybody else said was crazy... Bdale
tooted his own horn a bit about the amature satelites that he
helped launch. Dang. Just like the
the hackers are now after the final frontier!
The best sessions today were Mono-centric. The
release talk was interesting, hostorically speaking...
and the talk about
was well attended. Everybody hissed a bit when the author told
us that he was just hired by Microsoft to fix the .NET
engine... Im sure they'd rather he work on Mono, but we can't
all work for Novell!
The future of Java in the open source community is looking
somewhat grim. There was a panel with Eric Raymond and OReilly
and a few guys from Apache and Sun... developers are miffed
that Sun hasn't made Java more open... but at Java One the
developers were miffed that Sun hasn't put the kaibash on
forked versions of J2EE. Rock and a hard place, it seems.
The XML talks I attended were fairly good - PHP has some
really nice SOAP hooks in
Python hooks are not as good, and
looks like its trying to replace the verbose SOAP
altogether. It is a lot more powerful, and easier to secure
and use... but people love their SOAP because it goes over
HTTP, dont they?
The first keynote by OReilly was pretty good. He's still
pushing the internet-as-an-operating-system jive. His latest
theory is that the quality of the large databases (such as user
reviews of movies or books) is what seperates the 'ok'
websites from the 'great' websites. Obvious, sure, but frequently
overlooked. His theory is that they might be able to brand themselves,
and become the next 'Intel Inside' for web applications. If you
dont have one of the name brands, your web services will suffer greatly.
I agree to a point... but there's a big difference between Intel
and a data warehouse. Intel made a ton of money because they had
the manufacturing infastructure and research and talent to keep
the PC manufacturers supplies with chips. Its really hard to enter
that market... just look at how long AMD took to get off the ground.
Compare that to a database. Their power hinges greatly on who legally
owns the data. If I make a book review on Amazon, do they own
the data? Or do I own it? Should these reviews be totally open, and usable
by Barnes and Noble? Its a legal fine line, which makes it subject to
copyright laws and the whims of Washington. Their power could collapse
like a house of cards if they are not careful...
The second keynote was very funny, but he rambled quite a bit.
Robert Lefkowitz had a good point that 'Open Source' means different
things to different people, which causes a great deal of confusion in the
business world. However, he offered no real resolution. He did have
150 slides, however... so it may have been buried in there somewhere,
and I just missed it...
Like Monday, the Tuesday tutorials were a mixed bag...
The first tutorial was about
wxPython widgets, and it was very good. The widgets looked very
professional, and the wx libraries have been around for almost a
decade... so its stable and performs quite well. The only thing
that concerns me is deployment. Getting all the right libraries
on a vanilla system to get the UI to look good would be hard.
The second talk was disappointing. I had wanted to go to the talk on
advanced wxPython, but that was sold out. So I had to go to the one one
getting DBAs and programmers to play nicely. I have no idea who
the target audience was for that talk, but it sure wasn't me. I saw
very little practical advice for either DBAs or programmers. This
was very surprising, because the presenter worked on the human genome
project. He obviously knew his stuff... I just dont think he's a good
My first tutorial was about
object oriented programming in PHP5
was less about PHP5, and more about teaching PHP4 people
what object oriented programming is. I didn't learn much,
and had a hard time convincing the PHP5 guys that the
polymorphism in PHP5 is awkward and wierd. Oh well...
its not like its a real language anyway.
The second tutorial, about advanced topics in the
Python core framework was very good. It was a great
overview of tips and tricks for writing Python, as well
as a long brain dump of killer third party Python apps.
Paul Prescod really knows his Python. It really got me
fired up to use advanced Python a lot more.