So, today I turned 23. To celebrate, I bought myself a present, Sony DSC-H3. It takes rather good photos (says an 101-class amateur-level beginner photographer newbie like me) for the price, except the lack of an uncompressed format is a definite minus. Nevertheless, I hope to be a diligent tourist and take lots of photos during my half-year stay in Australia, so expect to see some of them here.

I don't get presents too often any more, since anything useful I'd really want would cost more than a few dozen euros. But yesterday I got a very nice gift from the company that shows that a good present doesn't necessarily need to be measured by its material value.



I wanted to show it just for you to see what I was working on this summer at Farmind. I did some programming on the graphics engine, game logic and AI. It doesn't matter that the title isn't the #1 in sales, just that we had great fun doing it is enough. And of course, WSOP2008 being the first Nintendo DS game to come out from Finland makes me feel a bit proud.

So last year was jolly good, but I'm anticipating 2008 to be even better. I'm leaving Finland and Farmind for 6 months in Newcastle, Australia. I'll be damned if I don't catch a single kangaroo on my camera or don't get to drink the Australian equivalent of Pina Colada at a beach. The visit is not purely holiday though. I go there as an exchange student and plan on taking a few courses in computer science and mathematics. And finally when I get back, I'll be already working on my master's thesis. A busy year.

Development stuff: I was doing the median-of-medians algorithm for the Kth smallest element in an array, take a peek. While analyzing it, I amused myself by coming up with the following fallacious proof that you can use any %$\Theta(n^2)$% algorithm to sort an array of integers in a %$\Theta(n)$% time. You proceed as follows: Take a %$\Theta(n^2)$% sorting algorithm, say, Selection Sort. Divide the array of %$n$% items into %$\lceil$$\frac{n}{k}$$\rceil$% subarrays, where each subarray contains at most %$k$% elements, where %$k\geq2$% is an arbitrary constant. Now we can sort each of these in %$c \in \Theta(1)$% time with Selection Sort, since the subarrays each contain a fixed number of elements, at most %$k$%. We need to do at most %$\lceil$$\frac{n}{k}$$\rceil$% of these, so the total time to sort is %$c \cdot \lceil$$\frac{n}{k}$$\rceil \in \Theta(n)$%. What is wrong, if anything, with this line of reasoning?



Last Updated (Thursday, 17 January 2008 02:31)