The thing that I value most in a really great class is that it pushes you out of your comfort zone. Many of the core classes in computer science are not fields that I am extremely interested in, however I am forced to endure them a second time to complete a Master's degree. The benefit to this pain is that two of my classes have good instructors. One instructor is not using a book - he doesn't need it - I think he is Romanian. He is fabulous. He presents the topic in a complete and logical way, but is teaching us hardcore Operating Systems the old school way, and I respect this. In the other class, I have a Asian teacher whose field is on the hardware level - far below where I normally venture. In his class we have to write a 20 page term paper! Journal papers aren't usually that long. He suggested that we work in Quantum Dot Cellular Automata or Clock-less systems if we didn't have anything else that we want to study. I've read 4 papers on QCAs now, and they are interesting, but I don't see anything there that I really want to do. In the process one paper talked about implementing FPGAs with Quantum-dots... then I started thinking about FPGA processors again... I'm going to spend most of Thursday and Friday learning about those. I think I'm leaning toward implementing some kind of automata on an FPGA, but I have to narrow the paper down to a very specific topic. DFAs and NFAs have already been implemented for some string searching algorithms because after constructing the automata you can do string matching in O(1), which makes them nice... and when you can implement them in reprogrammable hardware, well, you can just imagine how much faster that would be.
Check out
this news headline; very cool.
