Okay, I don’t want to finish this month without writing something on this blog, this website of mine is pretty much that, which is lame. I created this blog by myself for myself, so if I’m not putting the effort to register my researches and discoveries here, it means that I’m not putting much effort on myself.

Well, but it doesn’t necessarily mean that I’m just addicted to games and series (even though I am) – Which is worth mentioning, to be honest, because I noticed that the last two series that I watched (Elementary and Welcome to NHK)¹ reflects a lot about what I’m feeling right now. In the second day of this month, Youtube remembered me that the Advent of Code exists, so I tried to solve each puzzle more seriously.

And I learned a lot in this process, I would love to share a little bit of this wisdom that I acquired.

What And Why

In a nutshell, Advent of Code is a collection of 25 algorithm challenges that you need to solve to save Christmas (yeah, it has a lore). The answer for each puzzle is a number, and you should calculate this number based on the puzzle it self and an input, which is a text file with a lot of data. This input file is different for every user and every challenge.

Each challenge has 2 parts; the second part uses the same input file as the first one, but the result should be different. If you succeed, both of the two parts, you’ll get 2 starts, which means you need 50 starts to win. It starts easy and the difficulty increases for each level.

You can solve the puzzles using any programming language you want. Which is neat, I see a lot of people who aren’t particularly attracted by the competitive programming part of those challenges that uses AoC to learn new programming languages. And that’s a valid point of view, but I think I like another thing.

I really don’t care about the competitive part either, so I use it to train my design muscles to write readable and documented code. And it’s pretty convenient that each puzzle has two parts, because I can see how well I designed the first part of the puzzle, based on how much I had to change it to solve the second part.

My Adventure

I already knew about AoC last year (2023), but I was able to solve only 4 puzzles. This year I wanted to compensate that and put a realistic goal of getting 25 starts, half of the total, and I made it this time. I didn’t want to get all the 50 stars because I knew I would just burn out, some of the puzzles just isn’t suited for my level of programming yet.

When I was solving each puzzle, I got curious about other people’s solutions, because I’d like to know if my solution was not a crime in some country, I don’t know, and that’s when I discovered new Discord servers, met new Twitch streamers and found out the Advent of Code thread on lainchan.org – they were solving the puzzles using Haskel and Lisp, my beloved Lisp.

It was a lot of fun, and I’m happy to say that my code has the “readable” badge. I think I should be proud.

What I Learned

I documented all my solutions in a repository on Github. The main branch stores my 2023 attempt some of my 2024 solutions – I started with Python (the easiest for this kind of challenge), then I jumped to C++, then JavaScript and finally stuck with Elixir, my favorite language of all time.

I used C++23 for some of the challenges, and I loved it. I mean, it still smells like an old procedural language, but the std::ranges library helped a lot, and I also liked that C++ now has pipe operators like Bash or Elixir. But nothing beats Elixir, this language was designed to solve this kind of problem.

Elixir has a really strong standard library and a strong community as well, the pattern matching syntax is perfect, this month I officially love the functional pattern – now every problem to me is an array-ish problem. It even was easy to run Elixir as a standalone script; look, the code below plays all .mp3 songs from the music board of lainchan.org using MPV:

 1#!/usr/bin/env elixir
 2
 3Mix.install([
 4  {:floki, "~> 0.37.0"},
 5  {:req, "~> 0.5.5"}
 6])
 7
 8songs =
 9  Req.get!("https://lainchan.org/music/")
10  |> Map.get(:body)
11  |> Floki.parse()
12  |> Floki.find("a")
13  |> Floki.attribute("href")
14  |> Enum.filter(&String.contains?(&1, "/music/src"))
15  |> Enum.filter(&String.contains?(&1, ".mp3"))
16  |> Enum.uniq()
17  |> Enum.map(&("https://lainchan.org" <> &1))
18
19songs
20|> length()
21|> dbg()
22
23System.cmd(
24    "alacritty",
25    ["-e", "mpv", "--no-video", "--shuffle", "--log-file=uomagaa.log" | songs]
26)

I wish Python had a simple dependency manager for those kinds of standalone scripts, that would help me a lot.

But still about the repository, the better branch stores my last solutions, those were all written in Elixir, and, this time, I didn’t use a simple script file to solve, I tried to build an entire Mix project with tests included for each day. Which paid off, but I still need to improve my unit testing workflow.

And speaking of improving, I discovered that my weaknesses was code optimization and, the most critical one, linear algebra. Besides that, I also figured that my strengths are recursion – the other languages annoyed me because of that, they don’t have tail recursion like Elixir – and arrays – like, spamming .split(), .map(), .filter(), .reduce(), etc. in JavaScript; and Both of those of my strengths was the influence of Elixir, because That’s basically how you code everything in Elixir.

The Real AoC Was The Friends We Made Along The Way

Yeah, I know that’s sounds cliché, but I can’t help it, it was a lot of fun interacting with those passionate people. Like, we’re both selfish nerds who really love the art of creating and hacking stuff, it’s quite poetical how this selfish curiosity in highly specific interests can bring us together. Just saying.

Happy New Year!

That’s it for today, I didn’t do any research for that. As I said in the note, I want to put more effort into myself to post more today. This adventure of mine, mainly in lainchan.org – I love this place, as you can see –, inspired me to abandon an old project idea and start it again with a clearer goal in mind. I’ll write about it later; I just need to sketch some stuff and see what I really want, and I plan to create a simple social media focused on group discussion.

Besides that, I also want to learn more about Lisp and document my backup strategy, about my Arch Linux setup, etc. I had a lot in mind, I need to stop wasting these thoughts.

Happy new year! And thanks for your attention.