# Advent of Code 2020 Solutions Here's to my first 50 star year since 2017! The goal was to make it to top 100 at least once this year, and it finally came true on day 22 (and again in day 25 part 2, but that hardly counts). Honestly, I am not at all happy with myself, though. ## Kinds of Solutions I "raced" in [Crystal](https://crystal-lang.org), but also tried my hand at formal verification in [Coq](https://coq.inria.fr/), and tried out an APL dialect called [J](https://jsoftware.com/) for fun. I didn't always clean my race day solutions, particularly the ones to the "hard" days. I will write about my Coq solutions on [my site](https://danilafe.com). ## Rankings Here's the (rather embarassing) table with my times. ``` -------Part 1-------- -------Part 2-------- Day Time Rank Score Time Rank Score 25 00:07:18 105 0 00:07:24 94 7 24 00:25:27 1496 0 00:43:46 1165 0 23 00:27:30 887 0 01:35:00 1057 0 22 00:03:44 42 59 01:16:52 2242 0 21 00:21:57 612 0 00:26:05 439 0 20 00:27:52 465 0 03:00:50 845 0 19 00:14:56 123 0 00:48:50 439 0 18 00:26:17 1209 0 00:28:01 574 0 17 00:17:23 400 0 00:19:14 281 0 16 00:11:50 627 0 01:03:40 2040 0 15 00:18:11 1772 0 00:30:38 1924 0 14 00:18:05 1220 0 00:39:57 1173 0 13 00:04:59 198 0 01:05:11 1606 0 12 00:06:19 206 0 00:14:17 245 0 11 00:13:23 414 0 00:19:11 261 0 10 00:05:04 293 0 00:31:04 1493 0 9 00:08:55 1721 0 00:26:50 3540 0 8 00:04:23 275 0 00:09:56 256 0 7 00:17:46 832 0 00:24:22 615 0 6 00:03:31 385 0 00:09:38 999 0 5 00:08:21 839 0 00:27:23 3404 0 4 00:07:40 786 0 00:20:26 440 0 3 00:02:46 119 0 00:07:40 448 0 2 00:03:48 341 0 00:06:01 272 0 1 00:07:52 1072 0 00:09:00 748 0 ``` ## Crystal for Competitive Programming I really enjoyed writing Crystal for the Advent of Code, but there were a few reasons why it wasn't perfect for the task. * When you're not banking on brute force speed, you lose some time to the compiler. * Numbers are _always_ 32-bit by default, and require constant `_i64` suffixes everywhere in the code when they're involved. * Type annotations (however necessary they are) for hashes and arrays make refactoring a little bit slower.