Slashdot Mirror
Google Code Jam 2005 Winners Announced
Ember writes "The results of Google Code Jam are in. The winner is Marek Cygan from Warsaw University. Second prize goes to Erik-Jan Krijgsman from University of twente (Holland) and third to Pyotr Mitritchew from Moscow State University." Registration for the event took place back in July and Google reported a total of 14,500 registrants which is almost twice as many as last year, making for some stiff competition.
On the Code Jam page there is a link at the bottom to a recruiting page.
In Soviet Russia, asses suck this joke.
The US never had a monopoly on education. I mean, don't forget that the earliest modern-style universities were formed in Europe around 1200. Many were operating for around 500 years before the US was even formed. Today many of those institutions have been around twice as long as the United States, let alone the American educational institutions.
And before that there were centres of learning in Arabia, Egypt, Asia and Greece.
Cyric Zndovzny at your service.
First of all, it cost them over $155,000, as that is how much prize money they gave out. They also spent quite a bit to fly 100 people to the bay area, put us up in a hotel, etc. Second, everyone competes on the same problems, and they are problems that most people could solve given enough time. Google does this to encourage programming, as well as to recruit programmers, not to benefit from the work these programmers do.
I beta tested TopCoder once upon a time (they paid well for a few hours of college student's time). It has all the above problems plus a few not mentioned (limited ability to compile and test, limited access to tools, limited languages, etc). The two biggest things not mentioned here are that:
1)The problems aren't real world. They're heavily algorithmic, and generally a google search can find you pseudocode. The competitions are generally won by whomever knows the algorithm already.
2)Their code frequently requires heavy knowledge of the standard library for that language. If you don't know the StringTokenizer class in Java or wierd STL calls in C++, don't bother. Perhaps not an issue for everyone, but I learned C++ before templates existed, and never really liked the STL.
3)Diving right into code is generally a bad way to program, but in this competition spending time on design is a losing proposition.
THat said, it can be a fun thing to try out. I enjoyed it back in my beta testing days, even though back then they only allowed Java (one of my least favorite languages).
I still have more fans than freaks. WTF is wrong with you people?
Minor potential nitpick: I think you might have "steel" and "industrial" reversed. According to everyone's favorite source, The Industrial Revolution started around 1700 AD. While steel was invented by the Chinese around 300 BC, the mass production of Steel didn't occur in Europe until about 1855. I would argue that steel is more a product OF the Industrial Age than a CAUSE of the Industrial Age.
It is mostly highschool aged students pushed by their respective schools so the school can get some local press.
This guy Cygan is from the Warsaw University, not from a high school. His colleagues from the same departament already won other prizes: ACM IPC and Top Coder 2003.
"Long run is a misleading guide to current affairs. In the long run we are all dead." (John Maynard Keynes)
There were 3 competitors from the US in the top 10 (4th, 5th, and 8th). Also, there were more coders from the US in the finals than from any other country.
I believe that Poland had the second largest contingent. Poland has been doing quite well in programming competitions, as the competitors there get press more like sports players do in the US, which attracts other talented people to the field.
Sun already does. NSA, Yahoo, and Verisign are also sponsors. Looks like TopCoder runs it's own coding challenge independently of Google as well.
Men occasionally stumble over the truth, but most of them pick themselves up and hurry off as if nothing ever happened.
Brass??!?!?
It's bronze! Bronze age! Can you imagine people going around trying to use brass tipped spears?
-----
I just wanted to mention couple of things that might have influenced the recent results of polish programmers. Firstly, quite a few years ago a few people started to train promising young programmers in the art of algorithm design. Note, that they start very early, often in the primary school. These boys (usually) are trained by the best polish lecturers and older colleagues. Later, they often represent Poland in the international contests, with quite some achievements. But it's not only this group. Few years ago Warsaw University started a national programming contest in cooperation with the biggest polish newspaper. The contest takes about a week, and each day new tasks are given. It always starts with fairly easy ones, only later to go into really hard problems. This attracts many young people, that know how to implement e.g. simple sorting but don't really know much about e.g. graph algorithms. Competition in these contests gives these young people opportunity to extend their knowledge, and since it is a recurring event, they learn in the meantime and get better and better. Also, people from countries like Russia and Poland for a long time didn't have access to the modern technology, so instead of playing with machines, they played with theory. And in algorithms, the real problems lie in the design, not in the implementation. Finally, we're smart :)
This isn't Summer of Code, it's CodeJam. CodeJam is an algorithmic competition, it's much more akin to a track meet than a barn-raising. So there were no "winning projects", only winning solutions. You can read the final round problems, if you'd like.
If other reasons we do lack, we swear no one will die when we attack
I'd just like to comment on some stuff here that isn't quite right.
Sun, IBM, and many others have sponsored TopCoder competitions in the past. They have since backed off of them.
Sun still do. IBM haven't sponsored one just recently, but all companies can't sponsor every competition, there's a lot of companies trying to get in on it.
The problem with their approach is that it only proves that the coder can think and type fast. It does nothing to address teamwork, cleanliness of code, design capabilities, engineering ability, or many other areas that are critical to a real world programming job.
Actually that's not true at all. TopCoder now has multiple tracks of competition, each focusing on different areas, and they include *every one* of the things you mention.
These contests can be a fun way to compete with your peers, but my guess is that a lot of companies have found that placing too much empahsis on the results is a good way to get burned.
All of the company reps I've talked to are delighted with all of their hires from TopCoder events, with only one exception, out of dozens of coders.
If there's anything at all that these companies might find irritating about TC tournaments, it's that their tournaments are so accurate that 50% of the people who make the finals, and hence are around to be recruited and interviewed, will have already been to one before. But it's worth it to get the top talent, and sponsors rotate around.
I beta tested TopCoder once upon a time (they paid well for a few hours of college student's time). It has all the above problems plus a few not mentioned (limited ability to compile and test, limited access to tools, limited languages, etc).
Yes, that was years ago in beta testing. TC now lets you use whatever tools, compilers, and editors you like. Yes, there's limited languages, but if you're that good, you will know or can easily pick up C++ or Java or one of the other languages they allow. And that old myth of "a Perl coder would kick everyone's ass" doesn't hold up, except for extremely easy text-processing problems.
1)The problems aren't real world. They're heavily algorithmic, and generally a google search can find you pseudocode. The competitions are generally won by whomever knows the algorithm already.
The easier problems might use a standard algorithm as PART of the code, but googling for psuedocode doesn't help if you can't implement it damned fast. For the hard problems, you usually have to come up with something totally new and tailored for the problem. That's where the real test is.
2)Their code frequently requires heavy knowledge of the standard library for that language. If you don't know the StringTokenizer class in Java or wierd STL calls in C++, don't bother. Perhaps not an issue for everyone, but I learned C++ before templates existed, and never really liked the STL.
This isn't true either. StringTokenizer sure is useful, but if you don't know it, you'll just have to code it yourself, and it isn't hard. "weird STL calls" certainly aren't needed, but you might need to use and return vectors or strings.
This implies that writing hard-to-read code is actually a survival trait, since if your bugs are hard to find and the submitted test cases don't find the bug, you have a better chance of winning your room.
If your code *does* contain a bug, you're going to lose so many points that you won't win anyway, so this doesn't help you.
The ability to solve problems that, while "complex" , are still solveable in mere hours, is not really indicative of a truly great talent in either field, mathematics or computer programming
Some of these problems wouldn't be solvable in "mere hours" at all by ordinary programmers. It could take them days, and even then you may have a bug-ridden or inefficient solution.
The top programmers in these competitions aren't just slightly faster, they're *orders of magni