The winners have been announced at ICFP in Copenhagen. They are:
- First place: Frictionless bananas
C++ is the programming language of choice for discriminating hackers.
- Second place (Honorable mention): jabber.ru
OCaml is a fine tool for many applications
- Lightning division: HITORI
Java is very suitable for rapid prototyping
- Judges’ Prize: Eger a Marson are an extremely cool bunch of hackers
The final score tables:
Further details, including final maps, slides and a video of the presentation will be available here soon.
We’ve run the top 50% of entries on a new, larger, set of maps, and you can find the results as follows:
This time, we’ve hidden the top 5 scores. If you’ve made it into the last round, congratulations! We’ll be contacting the final winners very shortly, and announcing the results at ICFP in Copenhagen in mid September.
Here are the maps we used for Round 2:
Since the contest, some people have been enjoying themselves making their own implementations of the Lambda Lifter game. Here’s a couple:
- By Stefan Bühler
- By Jochen Hoenicke
Enjoy!
(EDIT: We have made a couple of updates to the score table based on queries from one or two teams. However, nobody will be eliminated who has already been announced to be through to round 2.)
We’ve now successfully installed and run the submissions and tested them on some small maps. You can see how you did in the first round here:
For the next round, we’ll only include the top 50%. We’ve hidden the top 10 scores since we don’t want to give too much away. However, it is pretty close at the top! We used the following maps in this round, some taken from the ones published with the contest task, some slightly modified to add new features and traps, and some completely new:
If you’ve made it through, congratulations, and good luck with the next set of maps (they will be bigger…).
Note that some teams scored 0 (and some even lower). We have verified every one of these by hand to ensure that they were installed and running properly. Many of these were simply missing installation scripts, or had given the lifter executable the wrong name, which we fixed. In the remaining cases, a score of 0 was caused either by not dealing with a spec extension correctly, not producing output on receiving SIGINT, by producing ill-formed output, or simply failing to find an answer at all.
The contest deadline has now passed. We hope you’ve all had fun, and we’ll keep posting news and statistics as they come in…
We’ve noticed that a lot of you are having fun creating devious maps. We’d love to see more of these, and we might even consider using them to help judging. So if you want to include them as part of your submission, we’d suggest putting them in a directory called ./maps.
We’ve fixed a minor bug in the web validator regarding update precedence. It was applying updates in the wrong order (how embarrassing :)).
Thanks to wilee on #icfp-contest for bringing this to our attention. You may want to recheck your routes on bearded maps.
We have made our surveying equipment more sensitive, and we are now beginning to find some growth underground which is hampering Robot movement. Details of these growths, and how to combat them, are available here, and you can also see this on the task page.
Here are some new test maps:
(EDIT 1115GMT Sunday: Added beard5 as a representation of this, just for fun…)
ICFP Programming Contest 2012 
