Comments on: And the winner is… http://blogs.mathworks.com/contest/2006/12/06/and-the-winner-is/ The MATLAB Programming Contest is a semi-annual competition where contestants submit MATLAB code to try to solve a challenge. For more information, see http://www.mathworks.com/contest/overview.html Sat, 21 Jan 2012 14:45:27 +0000 hourly 1 http://wordpress.org/?v=3.2.1 By: Rajiv Narayan http://blogs.mathworks.com/contest/2006/12/06/and-the-winner-is/#comment-2079 Rajiv Narayan Fri, 08 Dec 2006 23:57:25 +0000 http://blogs.mathworks.com/contest/?p=56#comment-2079 I've really enjoyed participating in these contests, and though my entries rarely figure at the top, I have found writing them to be addictive, and a fun learning experience. One of the aspects that interests me is to examine different algorithmic approaches to a problem. I think it would be useful to be able to cluster or group submissions by using a source code similarity measure. I like the idea of a generality prize, and would also like to suggest a prize for the most readable/commented entry, in order to encourage clarity rather than obfuscation. Looking forward to the next puzzle! I’ve really enjoyed participating in these contests, and though my entries rarely figure at the top, I have found writing them to be addictive, and a fun learning experience. One of the aspects that interests me is to examine different algorithmic approaches to a problem. I think it would be useful to be able to cluster or group submissions by using a source code similarity measure. I like the idea of a generality prize, and would also like to suggest a prize for the most readable/commented entry, in order to encourage clarity rather than obfuscation. Looking forward to the next puzzle!

]]> By: Niilo Sirola http://blogs.mathworks.com/contest/2006/12/06/and-the-winner-is/#comment-2078 Niilo Sirola Thu, 07 Dec 2006 18:39:16 +0000 http://blogs.mathworks.com/contest/?p=56#comment-2078 Wasn't able to contribute much this time around, but the problem sure was interesting, as was the "creative cheating" discussions in the newsgroup. I'm still looking forward to see some future challenges in continuous variables rather than discrete. This would bring lots of new things to be tweaked like discretization and maybe finally favor Matlab BLAS routines over spaghetti code. Also, continuous test cases would have more "information content" making probing more difficult Wasn’t able to contribute much this time around, but the problem sure was interesting, as was the “creative cheating” discussions in the newsgroup.

I’m still looking forward to see some future challenges in continuous variables rather than discrete. This would bring lots of new things to be tweaked like discretization and maybe finally favor Matlab BLAS routines over spaghetti code. Also, continuous test cases would have more “information content” making probing more difficult

]]> By: Daniel Nordlund http://blogs.mathworks.com/contest/2006/12/06/and-the-winner-is/#comment-2076 Daniel Nordlund Thu, 07 Dec 2006 14:40:38 +0000 http://blogs.mathworks.com/contest/?p=56#comment-2076 I think it would be more rewarding with a contest that focuses on Matlab and it's language rather than probing and constant tweaking. During this contest I found that the most interesting part was the cheating entries. For example, that way I learned how powerful dynamic regexps are. I did not learn anything from looking at the winning entry. Like someone wrote in the newsgroup, another round of Matlab Golf would be great! It's a lot of fun and you learn plenty about the Matlab language, syntax and functions you perhaps didn't know about such as TRACE and KRON. Another idea, perhaps even more useful, would be a series of pure optimization challenges. It could be similar to the golf contest only the point would be to write the fastest code that produces the correct solution - e.g. finds the longest palindrom in a string or "the heaviest box" of a matrix. Just use a similar but slightly different data set for the final evaluation of entries to make sure people can't probe and then use lookup tables. I think it would be more rewarding with a contest that focuses on Matlab and it’s language rather than probing and constant tweaking. During this contest I found that the most interesting part was the cheating entries. For example, that way I learned how powerful dynamic regexps are. I did not learn anything from looking at the winning entry.

Like someone wrote in the newsgroup, another round of Matlab Golf would be great! It’s a lot of fun and you learn plenty about the Matlab language, syntax and functions you perhaps didn’t know about such as TRACE and KRON.

Another idea, perhaps even more useful, would be a series of pure optimization challenges. It could be similar to the golf contest only the point would be to write the fastest code that produces the correct solution – e.g. finds the longest palindrom in a string or “the heaviest box” of a matrix. Just use a similar but slightly different data set for the final evaluation of entries to make sure people can’t probe and then use lookup tables.

]]>