Slashdot Mirror
TopCoder, Math, and Game Programming
reiners writes "DevX.com has an interesting interview with David Arthur (dgarthur), the 2003 TopCoder Collegiate Challenge winner. Arthur discusses many interesting topics: the similarities between TopCoder problems and math problems, why TopCoder performance is positively correlated with 'real-life' programming performance, and why game programming is where the action is."
programming 3D rendering engine. that is where all the action is. i learned more about linear algebra while writting 3D rendering libraries, then i did during the course of my degree. :)
Consensus is good, but informed dictatorship is better
I find it interesting that a math double-major, who's considering becoming a math professor, uses C++ as his language of choice, with Java coming second. Not Lisp, not Scheme, not Haskell - C++.
I'm not sure what conclusion to draw from that fact, I just find it interesting.
--
CPAN rules. - Guido van Rossum
I would look at a Top Coder victor the same way I would look at someone who can answer trivia questions correctly. The experience is incredibly valuable, but I wouldn't say that they are parallel at all. Most of the questions and tests are biased against people who have experience doing competitions. A veteran programmer would probably perform 10x better in a real world environment, and is much more valuable than a TopCoder winner who is still in school... but I could be wrong.
Actually it's not as simple as that. In order to be successful in TopCoder competitions, one has to be a rather smart person (in fact the 2001 winner is "the smartest guy in the world"), with good analysis, generalization and abstraction skills, and above all, a quick thinker. The problems are not trivial to dissect, and time pressure is strong.
So for a long term employment (3+ years), I would rather hire a young successful TopCoder participant (one can always gain experience but not smarts) than a regular but experienced guy.
When men used to be men
It really depends on how you define "action". Encryption seems to me to be even more exciting a field. It isn't as glamorous as game programming, but the math involved is amazingly interesting (advanced number theory, primality), and good encryption tends to last for longer than good game engines.
3D rendering is not entirely about math (probably a lot more to do with studying the brain and how people generally interpret images that they see). Encryption however is ALL math. Anyhow, that's my 2 cents.
Yup. There wasn't a challenge. I did them for awhile. When they used to pay for all competitions, I learned it was better to stay in the bottom ranks. There you could take a room on one question, the easiest because all you needed were language tricks. When it was just Java, this basically meant that you learned all of java.util and java.lang packages. Basically my strategy boiled down to this:
1. Read the easy question.
2. Recall the Java class/method that shortcutted the problem.
3. Write 3-5 lines of code (not including the class and method header).
4. Score near max points and be done before everyone else and sit for the rest of the round.
5. Challenge round, slam the people that tried to finish as fast as you. Chances are these people made a mistake and if you knew the problem, you knew what to look for right away.
Most of the time I'd also just open the last two questions so I knew what to expect if someone did finish them, but the bulk of the people in the lower rooms would never even get to finish them in the allotted time, if they did, you could almost always count on it being wrong. So long as I didn't answer the big questions and let my scores inflate, I never moved out of those rooms and I never saw a point in it. The only point I saw in getting to the higher rooms was to make the invitational. Yeah I'd love to win $100,000 but at 1 in 64 kids (back a year and a half), my odds weren't that good anyway even if I thought I was talented enough to win (which I know I'm not).
This is not real world coding and I would NOT encourage colleges to become involved in Topcoder because most of their philosophies go directly against what professors are trying to instill in up and coming programmers. You are not encouraged to design, comment, test or even read the problem thoroughly in these competitions because it all costs you valuable points. Whether or not these are valuable skills in the real world, I'm still growing up and learning.
well after 69 comments (hehe), there has not been a SINGLE one discussing the competition problems, all three of which are quite interesting.
/. lowers productivity. making people spending way too much brain power on stuff that's completely unrelated and time consuming. heck; i might lose sleep over this.
especially the hard one, probably, because my mind is drawing a blank on how to have it implemented... (no i didn't cheat and look at the solution).
heh, actually they go like this:
*easy* - okay, i can think of a algorithm. probably not the fastest thing in the world, but it should work out.
*medium* - have a haze of an idea on what an algorithm might look like. with enough caffine it MIGHT solidify.
*hard* - at least I understand the problem, but curses on the restrictions of a binary tree =)... no idea on algorithm that would finish executing before the end of the universe. (granted, only 50 elements, so maybe it's possible brute-force)
Damn; this is exactly how
My life in the land of the rising sun.
The top level guys competing in topcoder are some of the smartest guys you will ever meet. Whoever thinks they are a coder, go ahead and try a competition. Its free and they do it a couple of times a week. See if you can even get the easy problem right. I dare you.
p.s. Topcoder also has the best Java client side applications going. Their competition arena application/applet is a masterpiece.
no i don't work for them. Yes I have competed.
Say what you want, but for the math gifted, most of them will code in Occaml, or one of the Meta Languages (ML), if they ever come across them.
Muchas Gracias, Señor Edward Snowden !
If students were taught how to visualise the math they are suppose to be learning, they would pick it up much faster. You will find that ALL people who are good with math have a natural capability for visualizing the concepts. It is my firm belief that visualization technique can be taught. I have done just this while tutoring. I found Linear Algebra extremely easy because it was so visual.
I'm sure you know that the skills you demonstrated in the contest are only a small part of the skills needed by a professional coder. I'm also pretty sure you'd make a great coder. On the other hand, based on what I've heard about you, I suspect you're too smart to be just a game programmer or something like that, and that better things are in store for you. Of course, coding is still a lot of fun, good preparation for management or research, and a decent profession. Whatever you decide to do, thanks for stopping by on Slashdot, and best wishes.