During the Advent of year 2020 I realised that I finally have enough spare time to enjoy the legendary Advent of Code. While coding through the newly released puzzles, I decided to open one of the earlier years' set. Just to check a couple of those puzzles and see if it is any different. Fast-forward few days I found myself admiring the animated ASCII art that gets displayed on the calendar main page after completing the whole set. Solving these little, well-thought-through tasks became a habit, rendering me overjoyed and satisfied.
Hence comes this repo, documenting my path to gathering now 450★ stars and beyond, as I was fortunate enough to enjoy the fun even in the years that followed.
- While in trouble, go back to the puzzle description. Chances are that you've missed or misread a key information. Many hours were wasted because of not following this rule closely.
- Brute-force is often a working solution, if not a desired approach. In those times where the crucial observation that would reduce the computation complexity by orders of magnitude just did not occur to you yet, maybe it is possible to optimize your code, maybe run it in parallel, and have a solution in couple of (tens of) minutes. Not pretty nor elegant, thats for sure. Working? May be.
- There are times where brute-force just doesn't work, as the projected run time is in days even with all the easily available optimalizations. In those times go to point 1; chances are you've missed something in the puzzle description. If not, a good nights rest should do it. Solutions sometimes lie outside the box.
- If part 1 of the puzzle smells of a linked list, it will require a linked list approach for part 2. Granted.
- A number of puzzles are alike. Abstracting meaningful parts of your solution might be worth it. Talking Distance maps, Graph traversing algorithms or ASCII-art rendering here.
- This one is even mentioned in one of the puzzle descriptions: Do not always aim to solve all the imaginable puzzle configurations. Look for simplifications, boundaries, try to cut the corners.
- Visualization and observed simulation often helps moving forward even with the toughest problems!
- Real puzzle input is the important one; it might contain properties the sample input lacked, and may lead to observations necessary to solve the puzzle.
- And last but not least, remember that you don't have to solve everything by your code. Cumbersome inputs can be converted into more manageable state by a tool of choice rather than by lengthy regexes, difficult math problems can be passed to Wolfram Alpha, complicated logic can be solved manually, often even as a game, Graphviz is your good friend as well. Your task is to find the answers to the puzzles, and the means and measures you take to do so are also yours.
- For general puzzle categorization and difficulty roster, check this awesome summary.
https://adventofcode.com/events
JavaScript solutions, executed in browser or node.
Code is pretty raw most of the time, with few exceptions.
Contains far more than healthy amount of visualisations, also here and here (captured on YouTube too) or playable games where possible or available.
Whole repo is deployed on Github Pages, links to individual puzzle solutions are in their respective readmes. Run only at your own risk of course.
- Summary for 2015
- Summary for 2016
- Summary for 2017
- Summary for 2018
- Summary for 2019
- Summary for 2020
- Summary for 2021
- Summary for 2022
- Summary for 2023
