Last year in Japan I bought a Cowon iAudio D2 player, which have proven to be quite good. But a few days ago, I thought I'd try to upgrade the firmware of it. I then discovered that there are four different types of firmware depending on where you bought it. As I bought it in Japan, my firmware was not compatible with other firmwares. The reason for this are mostly due to small differences in hardware. In my case, I have the possiblity of watching Japanese television (not really useful in Norway).

Therefore, I thought I would try and upgrade to the european firmware (a lot more fixes get through to this firmware it seems), but I was a bit afraid I would brick it if it was the case. I looked around at the iaudiophile forums, and finally I found someone with the same attempt, and they succeeded. The procedure was easy, but to be able to use the european firmware, I had to rename them to have the same file name as the Japanese, in order for the player to pick them up. Luckily, it worked for me too. Phew!

Today, I'm sitting in a café in Oslo, waiting until we're leaving for Gardermoen and our flight back to the Netherlands after one week vacation in Norway. The weather was nice, and I got to do some skiing at least. However, I was actually supposed to be in the Netherlands already. The reason that I'm not, is that we (my girlfriend and I) missed the flight on Sunday. We actually missed it by one day, as we were 100% sure that we were leaving yesterday, so when we showed up at the airport, we were shocked to learn that we were 24 hours late! This was a silly mistake, as neither of us really looked at the date, we always assumed that we were leaving on Monday. Unfortunately, to be able to board the flight that we assumed was our flight, we had to pay 3000 NOK extra per ticket! In other words, we had to find other ways of getting back. Luckily, we got to stay at my sister place last night, and got new tickets for today's flight at approximately the same price as our original tickets (700 NOK per ticket). Hopefully, we'll be back in our apartment tonight

Damn, reading for exams is really not my favorite thing. It's not that it's very hard material, but the motivation is the problem. I always tend to get a bit sloppy with classes where the only form of assessment is the exam, and if the class is not very interesting either, it gets hard. However, these kind of classes are typically very theoretical courses, and one way I cope with it is to make them practical. For instance, in this course there are lot of distributed algorithms that the student is expected to know. Some of them are almost several pages long, and I'm really not the type for keeping all that in my head, and if I did, it would only be because I memorized it. So instead, I tried to implement the algorithm, as it helps with understanding because you can see how it works in action! What I did in this case was to create a node abstraction/class which I could re-use in several algorithms. The nodes definition is something like this:

void send(Message, nodeid); // Send message to a single node
void multicast(Message);    // Multicast message to all neighbouring nodes
void deliver(Message);      // RMI method called by other nodes via their send method
Message receive();      // Blocking receive method to fetch contents from buffer

The node creation itself adds necessary neighbours, and connections are specified at startup time. The Message class contains most info necessary, but is extended in some algorithms that need extra stuff. I implemented these algorithms using the interface:

• Ricart-Agrawala's mutex algorithm
• Maekawas mutex algorithm
• Peterson election in unidirectional ring

Some algorithms are really tricky, and I end up spending more time wondering how to implement it than actually doing it, so I guess this technique is not good always

Phew, the first quarter of my exchange study is almost over. So far, the stay here in the Netherlands have been very exciting. First of all, we did an awesome project creating a quad rotor controller using a joystick to fly. A demo of the previous years group can be found here. We were actually able to make it work like in this video. The hardware consists of a Xilinx Spartan 3E starter kit running the X32 CPU core developed here at TU-Delft, a PC with serial link to the FPGA board, a joystick connected to the PC, and the Quad Rotor itself connected to the FPGA board via a modified serial link. We implemented the control software, signal filtering etc on the X32 in C, and after optimizations, we had a cycle time almost half of the required, and it flew!

The other course I've been taking is a seminar on wireless sensor network, which handled nearly all aspects of this topic, having students present a paper on a certain topic each week. I presented a paper on reliable energy aware routing, which was very interesting.

Lastly, I have a course in distributed algorithms, which will finish on April the 4th with an exam. The course teaches various distributed algorithms for synchronization, global state detection, deadlock, locking etc, and goes through several P2P protocols as well.

After this quarter I'll also go home to Norway for a short vacation, finally