Naive game simulator (vlad's version, to distinguish it from possible alternative implementations)

[Vlad-Shcherbina]

Perfomance does not matter, I'm trying to keep implementation close to their description of the rules when it makes sense.

[Vlad-Shcherbina]

Status:

• it's in https://github.com/Vlad-Shcherbina/icfpc2015-tbd/blob/master/production/game.py
• all logic seems to be implemented
• interface is, let's say, not designed
• only rudimentary tests

Please start thinking about proper API, especially @fj128 .

I plan to improve testing in the immediate future.

[Vlad-Shcherbina]

There are reasons to think that it's buggy, and it blocks solvers.
If you can improve correctness, please do!

[Lomir]

I am thinking of making an alternative implementation, this way it should be easier to find bugs by comparing behavior.

[manpages]

@Lomir, how do you feel about writing universal tests for invariants (akin to what was done in tbd-playground) that will cover current core library?

It's way faster and these tests will be reused by other implementations.

Thing is that now we have golden testing for score mismatches and organizers hinted that they might improve reporting of errors by the end of the day. So I'd say that the priorities are —

1. Write C++ implementation
2. Test current implementation
3. Write a strategy
4. Call golden from core (not sure, @graphite might have already done that)
5. Write another python implementation
6. Do phrase search scavenge hunt
   ...
   N. Write tricky maps for potential battles in the finals