Ted Leung on the air: Open Source, Java, Python, and ...
Here's my roundup of CodeCon for today.
The most interesting talks that I saw were on the UltraGleeper, and H20.
The UltraGleeper is recommendation engine for web pages. The key idea was to interpret links as recommendations (and yes, "nofollow" did come up.). Instead of asking a user to fill out a questionnaire about their preferences, or do a bunch of up front ratings, the UltraGleeper uses a person's weblog and del.icio.us bookmarks as the "seed" for web page content that the user likes. After this, the usual feedback style of rating kicks in -- something familiar to many recommendation engines. It's good to see work on this sort of thing going on. At some point I think that it would be cool to have recommendation engine style functionality available in Chandler, but that's something for the future.
H2O is a system that was built at Harvard Law School to create a good environment for university classes. The point that stuck out to me the most was the throttling system that they built in order to regulated both the rate/tempo and structure of electronic discussions. The goal of this was to level out the playing field for discussions so that early and/or loud voices didn't dominate, as they are prone to do. Having recently commented on a discussion of small group dynamics on "The Wisdom of Crowds", I was predisposed to see the value of this design choice in H2O.
I viewed the talk on Mappr as a good example of the usefulness of folksonomies (even though they weren't perfect). I think it's also an example of the success of web services (in the most general sense of the term).
Kevin Crowston is here observing as part of his ongoing research on FLOSS. Kevin and his team have become a fixture at ApacheCon, and it was good to see him. We had a very interesting discussion on our perceptions of CodeCon versus ApacheCon and the respective communities.
The small size of CodeCon makes it somewhat less about the presentations and more about the people. I'm grateful to some of the great introducers who have been helping me meet interesting people in the CodeCon community. It's such a helpful thing when people are thoughtful about doing this. I want to build a stronger habit of doing this when I'm able to.
![Other blogs commenting on this post](/py-bin/images/bubble.gif)
So today was the first day of CodeCon... This (cell phone camera) picture gives you an idea of the atomosphere...
CodeCon is definitely a coder's conference. It starts at noon, which gives the nocturnal hackers (such as myself) a chance to get some decent rest. No presentations are allowed without running code, and the content so far is highly technical.
There's a decent number of ASF folk here. Ben Laurie and David Reid gave a presentation on the new certificate authority that they are working on for the ASF. I paid a decent amount of attention because I'm going to be a user of the system once it goes into production (I also submitted my request for a cert)
I was interested in the talk on the Aura reputation system. I'm going to take a closer look at this -- reputation systems are something that I'm very interested in, so it was good to see that people are out there working on them.
My powerbook decided to crash very hard right at the beginning of the ArX talk, so I only heard parts of the talk, since I spent most of it with CliffClassic doing deep surgery on the machine. Fortunately, Cliff knows his Powerbook kit, and he got me back up and running. It's a mystery to me as to why the machine decided to just lock up and die, but that's what happened. It has been behaving oddly -- the Cylon like sleep LED has been working incorrectly or not at all since December, and I discovered that the keyboard backlight was not working right (although Cliff's Powerbook-fu seems to have gotten that going again). Cliff also thinks that I need a new inverter for my LCD because a portion of the screen is dimmer than the rest -- I'm reluctant to give up the machine to get it repaired -- and of course, I live a goodly distance from an Apple store.
I missed much of the OTR presentation, but the parts that I did hear seemed like they have something really useful (at least if you value your privacy). I'm especially interested in their proxy for iChat. I'll be interested to see the demos tomorrow.
Reusable Proofs of Work (RPOW) seems like an interesting method for doing resource allocation (and/or payment). Unfortunately, it seems like it will be some time before RPOW tokens could be used in the real world.
![Other blogs commenting on this post](/py-bin/images/bubble.gif)