ICFP contest entry the first

[gwillen]

Welp, the 2010 ICFP contest happened. I was on team CBV (Cult of the Bound Variable) with a bunch of CMU grad students and otherwise CMU-affiliated people. We placed in the top 5, which is exciting; we won't know how we did beyond that until the ICFP happens and they announce the results

Comments

Congrats from team vorpal

[175560]

stereotype441 and I gave it a shot with two other Portland software guys. Despite me coming down with acute appendicitis on the day of the contest, the other three team members managed to get to 16th place without using any external software.

Unraveling the problem to its mathematical core by Saturday night was crucial, as was writing automated scripts that could scrape for cars, solve them, and submit solutions without intervention. Unfortunately, the solvers we were able to come up with on our own were not very sophisticated (especially compared to what you guys were doing), but fortunately, many teams submitted many cars that were not mathematically interesting. We solved over 2,000 cars using only 1x1 and 2x2 matrices.

Congratulations, CBV!

[stereotype441]

As regards the ternary encoding: I was one of the rare people who really enjoyed puzzling through it

I too enjoyed puzzling through the ternary encoding. I love doing that kind of detective work. When we figured out the natural number encoding

'a list = 0 | 1 * 'a | 22 * N:nat * 'a ^ N

jes5199 and I looked at each other and said "that is exactly the kind of encoding a functional programmer would choose".

Anyway, congratulations on doing so well. Being in the top five is a commendable feat. We didn't think of integrating with a 3rd party math package like Mathematica or AMPL. I can see why that would have made a huge difference in your score.

Paul (team vorpal)

[_adept_]

Correction:
Length encoded after "22" is (number of elements in a list - 2). That is, 22{0} is a list of length (2+0).

[gwillen]

Thanks, I have edited the post. I forgot the +2 when I was unrolling "wnat" into the definition of "list".

[gwillen]

Thanks! All credit goes to the people who did the actual work, I just did the perl hacking. ;-) I am glad we had someone on the team who knew how to apply Mathematica+AMPL to the problem, and had a copy handy. (And the AMPL company is letting us borrow a copy of the full version for free if we want to enter the termination competition with it.)

BTW, I can't remember how we LJ-met before -- was it a previous ICFP, or was it the MIT Mystery Hunt?

[gwillen]

Haha, very true.