http://qs1969.pair.com?node_id=491393

I like the premiss behind Topcoder.

In SRM's, you get 3 problems and about an hour to solve them. You get points proportional to (solved : bool)*(time to solve : int).

After the round is over, you get to view the other member's code and try to find test cases on which they will fail (and get more points when you do). Right now, it is only possible to use one of C++, VB or Java.

The SRM's are actually fun in a video game sorta way, but it really seems that other language families (perl, lisp, prolog) would complete the picture better. For many problems, it would be an advantage to be able to choose.

I'd like to ask them to add Perl, but I was wondering if anyone has a better idea on how to do so than just by sending them an email. Thanks

Franck Binard

Replies are listed 'Best First'.
Re: Perl on Topcoder
by pg (Canon) on Sep 12, 2005 at 22:30 UTC

    You probably noticed that, most of those problems are more about algorithm not programming language. Essentially it is not a competition between different programming languages, but rather a competitition between the brain power of the coders. Language really is secondary in this kind of competition.

    But still it is not a bad idea to add Perl, so that programmers whose primary is Perl can compete with peers.

[reply]
Re: Perl on Topcoder
by Anonymous Monk on Sep 12, 2005 at 22:34 UTC
    You get points proportional to (solved : bool)*(time to solve : int).

    The longer it takes to solve a problem, the more points you get? Hmmm... that sounds... odd! Are these demerit points?

    After the round is over, you get to view the other member's code and try to find test cases on which they will fail (and get more points when you do).

    Apparently not... hmmm... How do you win? By doing a really bad job, and pointing out how others do a really bad job, too?

[reply]
      How do you win? By doing a really bad job, and pointing out how others do a really bad job, too?

      That sounds suspiciously similar to politics. :-)

[reply]
      sorry:

      1st part: inverse proportional to (solved)?time to solve:max time

      2nd part is correct. You get additional points for any test case that faults the code. You get demerits for trying a test case that fails (by being handled correctly by the adverse program)

      Many reasons to want to use Perl sometimes. Just 2:

      • There are problems that are just made for regex handling.
      • creating and using a map using the stl (in C++) is way longer than using a hash (and time is a factor)

      Franck Binard
[reply]
        using regexes is IMHO cheating when it comes to showing how much brain power you have :) They are too convenient and a huge shortcut in many cases.
[reply]

Back to Meditations