advent of code 2023
 kdb+   software 

The lower end of the difficulty scale was definitely increased this year - usually the first couple of days are simple one-liners in kdb+. A rough heursitic on days 1-5 of this year vs last year gives a 100% increase in lines of code needed to solve both parts. At the same time, I felt that that the upper end of the difficulty scale was decreased, most of the problems (at least the first parts) seemed to be quite doable to me - other years, I have drawn a blank on higher numbered dates.

By the end of December, I had collected 36/50 stars, which is a personal best for me for a given year. I do have intentions to try my hand at the remaining problems throughout the year. As I mentioned, I think I can manage the first part of each day.

What did I learn?

Day 17 was my favourite problem this year, as I had never implemeted Dijkstra's algorithm in kdb+ before. It took me until the 28th of December to produce a solution that eventually returned an answer, previous implementations were far too slow. My current solution still leaves a lot to be desired though, running in ~30 mins for each part. I also quite enjoyed Day 7, Day 8 and Day 13.

My solutions to this year's problems are available on my Github profile.