Back after a long break, mostly because of exams.

Got eliminated after round-2 in this year's GCJ. It was fun, but I was hoping for a better performance this year.

Read an old blog post by Steve Yegge, that got me trying something I've been putting off doing for a long time. Started by coding a simple arithmetic expression evaluator. I'll probably start on something slightly bigger next.

ncm: Really good points. I read that blog post, and it is hard to believe that guy was actually involved in HotSpot. Most of what he says sound like "I refuse to believe it is that bad. It isn't Java's fault, it's your code that's doing it in a bad way". You comment wasn't offensive at all. I don't hate Java, and actually was a big fan not long ago, but simply denying Java has any weak points sounds highly idiotic.

Very happy today. Did very well in today's topcoder match. This is also the first time I got myself a slight advantage there by using a language with built in big number support.

Reading a lot of open source code. I must admit reading code from real world, living breathing projects (as opposed to quick and dirty hacks to solve a particular algorithmic problem, often deliberately aiming to minimize characters typed as well) has greatly improved my coding style.

While reading about memoization, I stumbled into decorators and decorator libraries for python. Really useful. Wrote my own memoization decorator for python, using ideas lifted from some perl and python code and a few additions of my own. I'll probably make it available online soon.

I've started reading SICP. Fascinating book. But can't build mit-scheme on my machine because I simply don't have enough memory. The ubuntu repositories are no help either because I use 7.04, which they just stopped supporting. I can find compiled binaries elsewhere, but I guess I might as well use this excuse to install 8.04LTS.

Moved my blog from LiveJournal to WordPress. WordPress looks so much cooler, and its open source to boot. The only catch was that LiveJournal only allows one month to be exported at a time. But this wasn't a problem for me since I only started there this month.

Just solved this problem at SPOJ. Pretty easy, but didn't risk getting a time limit exceeded verdict with Python and coded in C++ instead. The problem could be made slightly more interesting by increasing the number of jobs and the range of priorities (e.g. 10^5 perhaps).

Downloaded and built from source xmlrpc-c. I still prefer Python (or even Java) for xmlrpc, but knowing how to do it in C will surely prove useful later. Its fun to set up clients and servers in different languages and have them talk to each other.

I'm certified as an Apprentice now! Thanks to fzort, badvogato, zanee, and ncm.

Got pwned at the last TopCoder single round match. The easy problem could be solved by using a O(n) string search method. I tried Rabin-Karp, but made a very stupid mistake. A learning experience nonetheless. Posted a longer rant/post-match analysis in my LiveJournal blog.

Downloaded source of quite a few open source projects in the last few days. Going through them very carefully. mod_virgule code still baffles me. But I'm slowly making progress. The fact that many of the projects in the projects page here are either dead or have broken links (some even sound bogus) is also causing me some frustration.

Solved another problem at SPOJ. Currently mine is the only accepted Python solution there, but probably not for long.

Just noticed that fzort certified me as apprentice. Thanks fzort! However, advogato still considers me an observer.