Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Advent of Code 2023 is nigh

adventofcode.com

Post reply on HN
Print view View original

161–170 of 319 posts

Re: Advent of Code 2023 is nigh

#161

Post by Avshalom »

greymalik wrote: post #52

I've never done AoC before. It seems like the success criteria is primarily about getting the correct answer, and secondarily about submitting a solution as quickly as possible if you want to be on the leaderboard. Is that right? Is there any centralized place for seeing other people's solutions? I'd like to be able to learn from how others approach the problem, and what more elegant or performant solutions exist tha…

submitting as quickly as possible is very much not a success criteria. Each problem becomes available midnight EST so unless you're a competitive weirdo you won't even see the problem until the leaderboard is full.

Re: Advent of Code 2023 is nigh

#162

Post by epiccoleman »

sixothree wrote: post #95
Waterluvian wrote: post #26

Every year I want to love this. Every year I get four days in before it feels like work. I think I’m just the wrong audience, but I really do want something this well-produced but with perhaps a very shallow diff little curve, bordering on just effortless fun.

Agreed. Games like TIS-100 just feel unrewarding to me because I'd rather be solving my own problems.

I had a similar realization a while back with Factorio. I caught myself watching Youtube videos on how various systems worked and how to optimize things and solve problems, and I just thought... "man, think what I could build if I put this same energy into my programming projects."

Still, I like Factorio and similar games, and I'm excited about the upcoming expansion, but it is just kind of funny to notice that it basically feeds on the same kind of drive that I use to build little apps and whatnot, and with those projects, I end up with something a bit more tangible than an elaborate virtual factory.

Re: Advent of Code 2023 is nigh

#163

Post by imperialdrive »

Gosh I'm feeling extra special this morning as an absolute infant beginner tinkerer that figured solution in powershell to part II after a few minutes of closing my eyes, but I highly doubt I'll be able to say that about the next one if they get harder. The example was so excellently written though, which helped a ton.

Re: Advent of Code 2023 is nigh

#164

Post by Brentward »

AdamH12113 wrote: post #125
jjice wrote: post #101

Earlier quoted context omitted.

Same here. I would've really like if the spec specifically mentioned the possibility of that one edge case ahead of time instead of having to sift through the 1000 lines of input. No hate on AOC though, I really respect all the hard work that goes into it.

Unless the question has been edited recently, it did. There are multiple lines in the second example input that show the overlap: > eightwothree > 4nineeightseven2 > zoneight234 I test my AoC solutions incrementally by printing output, so I found that I was failing to produce the correct list of numbers in a line right away. I suppose if you're taking a faster approach and just trying to extract the first and last nu…

Those examples contain overlap, but not in the first or last digit that is given as the solution. So from the instructions you can't tell that the intermediate list of numbers should be 8, 2, 3, you only know that the final answer is 83.

Re: Advent of Code 2023 is nigh

#165

Post by mlsu »

bryancoxwell wrote: post #25

I think I’d like to try this year’s in a language I haven’t touched before. What languages should I consider if I want something paradigmatically different from Go, Python, etc?

If you haven't tried Rust:

Someone put together a very nice template in Rust that automatically downloads tests, solutions, creates a scaffold for the binaries, etc and submits solutions through CLI. I used this template last year to learn Rust and it got me "up and running" quickly and easily.

https://github.com/fspoettel/advent-of-code-rust

Re: Advent of Code 2023 is nigh

#166

Post by Tainnor »

Day 1 part 2 was harder than is common for the first day. I ended up using parser combinators (attoparsec for Haskell), which morally feels like the "right" way to do this (so you don't have to manually keep track of how much you've read or somehow abuse regex lookaheads), but I don't know the library well and it took me some time to implement it correctly.

Re: Advent of Code 2023 is nigh

#167

Post by jetrink »

shever73 wrote: post #92

As others have said, part 2 of today's was really difficult. I finally solved it using Python regex `overlapped=true`, but it was very tricky. The irritation of having all of the test cases passing, but it failing for my challenge input! I hope it doesn't scare off newcomers, but I already know a few who have given up on part 2.

> As others have said, part 2 of today's was really difficult.

I suspect it was purpose-built to foil ChatGPT. I solved it on my own first and then went back and tried to help GPT4 write a solution. Even after carefully walking it through a strategy, discussing edge cases, and having it outline a solution in pseudocode, it still failed completely to translate our conversation into working code. (It didn't even work for non-tricky input.) It did anticipate the edge cases though, so that's something.

Did anyone have any better luck with ChatGPT? I wonder if LLM-resistant puzzles and generally greater difficulty (at least for the second star) will be a theme this year.

Re: Advent of Code 2023 is nigh

#168

Post by burntsushi »

j1elo wrote: post #127
burntsushi wrote: post #80

Earlier quoted context omitted.

Rust regex crate author here. fancy-regex is built on top of the regex crate and supports look-around. The regex crate doesn't support arbitrary look-around because it isn't known how to implement efficiently. See: https://swtch.com/~rsc/regexp/regexp1.html

> The regex crate doesn't support arbitrary look-around because it isn't known how to implement efficiently. A bit of a philosophical question: If how to write an efficient implementation is yet not known to man, ie. not a matter of the library's author time or skills, but literally a limit on human knowledge: why not at least provide the functionality with a good enough implementation? (with caveats just possibly me…

ReDoS is one answer. Another answer is that it would effectively require supporting another regex engine with very different semantics internally. That's a tall order, and with the existence of fancy-regex (and other engines), there's no real compelling reason to do it.

You don't need to be all things to all people. Embrace the fact that there are other choices!

Re: Advent of Code 2023 is nigh

#169

Post by philomath_mn »

jetrink wrote: post #167
shever73 wrote: post #92

As others have said, part 2 of today's was really difficult. I finally solved it using Python regex `overlapped=true`, but it was very tricky. The irritation of having all of the test cases passing, but it failing for my challenge input! I hope it doesn't scare off newcomers, but I already know a few who have given up on part 2.

> As others have said, part 2 of today's was really difficult. I suspect it was purpose-built to foil ChatGPT. I solved it on my own first and then went back and tried to help GPT4 write a solution. Even after carefully walking it through a strategy, discussing edge cases, and having it outline a solution in pseudocode, it still failed completely to translate our conversation into working code. (It didn't even work f…

Yeah I had GPT4 take like 5 stabs at it with explicit test cases and explanations and none of them worked

Re: Advent of Code 2023 is nigh

#170

Post by nick0garvey »

Chobilet wrote: post #119

Without going into spoilers, its interesting that people jumped to regex to solve this. For me that was a fairly non intuitive when I first saw the problem(both parts). What jumped to me is the problem statement indicated a finite number of states and I crafted a solution based on that information. But its really cool to see how we all jump to different implementations.

I also started to write a finite state machine for part 2 but found it too tedious to craft by hand. How did you do it?
Post reply on HN