While I'm a fan of the ICFP, I do feel as if the questions are a bit too full of themselves (for lack of a better term).
While I'm sure they are trying to remove ambiguity, but the way they are couched and their length make them a lot less approachable than they should be. I feel like in about 1/10 the wording and narrative complexity they could create an equally interesting programming challenge.
The whole task was just formulated as "save Endo". A spec of Fuun DNA (a virtual machine), Endo's DNA and a two pictures were provided. That's about all contestants were told. Then, we had to discover Endo's DNA contained loads of fun things inside.
Oh, ICFPC'06 was incredibly cool, too, but I didn't participated at that time and only took the task years after that. Don't want to sound whining but other years are more conventional (a complex, but mostly well-defined task) so less fun in my opinion.
In the interest of promoting friends/coworkers: The contest this year is a really cool challenge, and was set up by one of my partners in crime, Duncan. (Of course, ICFP is always a cool challenge and is always set up by good people :)
> (Of course, ICFP is always a cool challenge and is always set up by good people :)
I agree in general. Only that car engine / fuel challenge a few years ago was really obtuse. (And when I met the guys who set it up, they were so happy about it. ;o)
Lambda the Ultimate and the satelites were really cool!
[The ICFP Programming Contest 2014 is the 17th instance of the annual programming contest series sponsored by The ACM SIGPLAN] International Conference on Functional Programming.
So the strategy for the GCC (the Lambda-man controller) is that you actually build a compiler/pre-processor for it in some other tool, versus just logically solving the problem with the given instruction set?
The details on the ghosts are dramatically clearer than the details on the Lambda man.
(Posted about the contest an hour earlier, only got two votes. Oh well.)
If I am reading the problem correctly, there is no interactive component this year? You get one submission and don't get to test it against anything else.
> It is not essential that the judges be able to run your code
It makes a little more sense if you don't think about the game as an assembly of parts, which isn't really explained anywhere. For example there is some strategy with trying to make the ghosts be as bad as possible, so they are in a nice neat line immediately after a power pill.
As usual, kudos to the team for making a puzzle with so many layers.
While I'm sure they are trying to remove ambiguity, but the way they are couched and their length make them a lot less approachable than they should be. I feel like in about 1/10 the wording and narrative complexity they could create an equally interesting programming challenge.