02:33 pm, 21 Jul 07
icfp!
This weekend I've been working on the ICFP contest again.
I'm taking much more casually than last year -- I started after dinner last night and gave it just a few hours -- but the fun is not in being a strong competitor but in solving the problems.
I lost a good four hours today tracking down a C++ precedence bug. I needed to test whether a number was even and wrote
which unfortunately parses as
which istrue for all nonzero i always false.
The worst part of it is, they provided a check program that told me nearly exactly where the problem was, but I looked over that code and convinced myself it was so obviously correct that a bug elsewhere must be cascading into the failure here.
That's what I get for being too clever, I guess! Shoulda just used mod in the first place.
(I'm also a little bummed that speed concerns have kept me from using any languages other than C++, though really we should've gotten to this point midday yesterday...)
I'm taking much more casually than last year -- I started after dinner last night and gave it just a few hours -- but the fun is not in being a strong competitor but in solving the problems.
I lost a good four hours today tracking down a C++ precedence bug. I needed to test whether a number was even and wrote
if (i & 1 == 0)which unfortunately parses as
if (i & (1 == 0))which is
The worst part of it is, they provided a check program that told me nearly exactly where the problem was, but I looked over that code and convinced myself it was so obviously correct that a bug elsewhere must be cascading into the failure here.
That's what I get for being too clever, I guess! Shoulda just used mod in the first place.
(I'm also a little bummed that speed concerns have kept me from using any languages other than C++, though really we should've gotten to this point midday yesterday...)
Good luck!
It's seriously a lot of fun, and cross-timezone teams are great because sometimes problems are serial.
Now it's Haskell (which was being worked on in parallel). The DNA => RNA conversion time is great. RNA => image, less so.
Any luck there?
You need to run potentially millions of steps -- ours takes tens of seconds for hundreds of seconds.
I hope you'll post a reminder a few weeks before ICFP next year. (Or maybe that's part of your meta-strategy to keep competition down... :) )