August 17th 2007 And That’s a Wrap
As the days of summer dwindle, so, too, does my summer of research working on SWAN. Overall, it’s been a fun time. There were a few hiccups along the way. My initial tactic didn’t pan out; turns out SWAN doesn’t spend as much time doing gain calculations as either I or my advisor thought. I then moved on to adding support for using precomputed gain maps in SWAN. These maps are computed using techniques from ray-tracing (generally found in computer graphics) and can be quite huge. This required the implementation of a structure known as a k-dimensional binary search tree, or k-d tree for short. This is a really cool data structure that I implemented very quickly in Java (a few hours, in fact), but took weeks to do in C++ because I am a novice C++ programmer and C++ is a clunky, ugly language.
The good news is that today, I finally completed the implementation of this feature in SWAN. SWAN can now use precomputed gain maps for modeling wireless networks. This is much more flexible than the old system. I haven’t completely tested it, nor have I begun the “science-y” aspect of my work. And, unfortunately, since its the end of the summer, I might never get to that. I’m going to tie up some loose ends regarding the analysis of the usefulness of my new feature, but the “hardcore” analysis might have to wait.
I had planned to continue this work as part of my honors thesis, but I don’t expect to do this now. Much of it has to do with my desire to get back to work with which I am more comfortable (and I am not comfortable with C++!), as well as to do something fun. My current plan is to do an in-depth security analysis of online web apps platforms. Basically, this means a case study of Facebook. More specifically, it is a case study of Facebook apps, those little things everyone is adding to their profiles. I really don’t like Facebook, and I am anxious to see just how secure its public applications platform is.
Unfortunately, this means parting with my advisor, Professor Perrone, and finding a new honors thesis advisor. I had looked forward to working with him (and, I should note, I have not completely ruled out the possibility of continuing work on SWAN, which was a lot of fun), but this Facebook idea seems intriguing and I’d like very much to pursue it.
The research this summer was, however, quite fruitful and enlightening. That, coupled with a newfound enthusiasm which began with a chance meeting at a party on July 4th, has made me much more excited about pursuing a Ph.D. for another 4–6 years after Bucknell—and eventually becoming a professor/researcher as well. While I hope to have a great senior year at Bucknell, my post-Bucknell prospects are looking better and better.