mich

Some of his (non-programming) achievements: http://www.the-elite.net/thehall/snap.htm

He has been long one of the very top TopCoder members (for which you need absolutely superior algorithm skills to be ranked that high), I think he has also won ACM finals one year. He is very well known in the problem-solving community.

Another guy I rank very highly is Steve Newman, he also had a rating over 3000 at TopCoder, but he is not competing anymore and has started a startup Writely (acquired by Google) - now Google Docs.

Yes, programming contests aren't everything -- but I can say I've learned almost everything I know about algorithms by competing there. And algorithms are important, if you are not doing just some simple web app.

Yes, Tomek and SnapDragon are great, however the rest of the programmers mentioned (checked their websites and stuff) don't seem anything special, let alone rock star. I consider a rock star someone who can solve tough problems, not create a basic database-backed website in Rails over the weekend.

I am interested in a MS/PhD program. So are these test the most important whether you get accepted to top programs or is it something else? Are perfect grades more important? Or research work / competitions / etc? How difficult is it to get accepted?

You are right, of course. My mistake.

I find the elitism here too funny. Most people seem to think that using Ruby or LISP makes them superior to thosepoor Java programmers while in the end all they are creating is a web-app like twitter (and of course think they are changing the world with this!). There are many Java programmers that are better than you.

And yes, as mentioned before they are using JVM which means you aren't limited to a single language(Java).