degrees of separation and open source
First up, I've been exploring Orkut a bit more and umm, I know this is going to sound trite, but after that initial "oh, he's here too, cool!" (ala Darshan and a few others) reaction, I'm a bit clueless as to what you actually do there. Help a brutha out here. One thing that would be quite cool to do (and even vaguely related to my research, the horror) is map the "semantic distance" between two given people. I do this for words now and it occasionally gives you some scary results.. for instance, how are houses related to dogs ? well, dogs usually live in houses. Sounds obvious, right ? Except that the crawler discovered that link automagically without knowing what a "dog" or a "house" means in the real world. How does that translate to Orkut ? Well, I am not sure yet. I do know that they are very down on automated crawls though, so I best proceed with caution *grin*.
Second, I have to hack and bugfix a piece of software written in PHP. It's a fairly complicated beastie and it does lots of stuff well. Unfortunately, when it barfs, it leaves spectacular craters (and organic matter) all over the page. Debugging it is like using ground glass for a tasty pre-lunch snack because there are templates, models and controllers littered all over. Now, I *know* people are probably thinking "quit yer bitching and just get the job done, ya whiner", so I saved the best for last. Found in the README is the following paragraph, which I shall reproduce in it's entirety.
"Part of the site was originally written in French so the code contains several parts with french comments. A good occasion to learn this great language".