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

Advent of Code 2022 is nigh

adventofcode.com

Post reply on HN
Print view View original

61–70 of 150 posts

Re: Advent of Code 2022 is nigh

#61

Post by xigoi »

jiggawatts wrote: post #51

Every time I do the AoC puzzles I wish for two things: 1) The site should ask for the language used to solve the problem and then use this as semi-scientific data for comparing "time-to-solution" for various languages. The sheer volume of data and the relatively high difficulty of cheating on novel problems would make this data set interesting and maybe even useful. 2) I wish all the puzzles had a "hardcore" mode whe…

> site should ask for the language used to solve the problem and then use this as semi-scientific data for comparing "time-to-solution" for various languages.

I think this would just show you the average time zone of the language's users.

Re: Advent of Code 2022 is nigh

#62

Post by masklinn »

birdfood wrote: post #27
javajosh wrote: post #3

I did several of these last year and found them deeply unpleasant. I like puzzles but some puzzles are just...bad. Like the ones that require you to just brute force a solution, or write lots of detailed code to capture the requirements...that shouldn't be what this is. Just my opinion, of course, but like I said I found last year an unpleasant chore that felt more like work. Hope it's not like that this year.

I felt similar last year. I was trying to do them in Zig and Rust, and it felt like most of the challenge was just writing custom data parsers for poorly formatted data which felt too much like my day job.

Fwiw there’s a few recurring formats, the old hands have a bunch of helpers for that and other things. Sometimes there’s something completely bespoke but the most commons are space-separated lines of things, usually ints, commonly fixed-size.

Re: Advent of Code 2022 is nigh

#63

Post by pedrosorio »

xigoi wrote: post #61
jiggawatts wrote: post #51

Every time I do the AoC puzzles I wish for two things: 1) The site should ask for the language used to solve the problem and then use this as semi-scientific data for comparing "time-to-solution" for various languages. The sheer volume of data and the relatively high difficulty of cheating on novel problems would make this data set interesting and maybe even useful. 2) I wish all the puzzles had a "hardcore" mode whe…

> site should ask for the language used to solve the problem and then use this as semi-scientific data for comparing "time-to-solution" for various languages. I think this would just show you the average time zone of the language's users.

Today 4 minutes gets you top 1000. There’s more than enough users starting when the problem gets released that an analysis could be done by checking for solutions submitted within the first hour.

Re: Advent of Code 2022 is nigh

#64

Post by matsemann »

pedrosorio wrote: post #58
Jtsummers wrote: post #49

Earlier quoted context omitted.

1. Have everything in place at the start (a script to grab the inputs helps). 2. Use a language that makes the edit/run cycle short. 3. Be very familiar with that language and how to process input in it. 4. Have solved many similar problems in the past so you can parse these problem statements quickly. 5. Have solved many similar problems in the past so you know exactly what code to write or use (if you have a utilit…

“Have solved many similar problems in the past so you can parse these problem statements quickly” This sounds like an advent of code specific thing. The problem itself is trivial, no need to have solved anything similar in the past, but the description is so verbose (because it’s a cute Santa-related story) that it becomes hard to parse the problem we’re trying to solve, quickly. I don’t think this is common in any o…

Feel I have to disagree on this one. Firstly, most programming competitions I've been to have this kind of problem description as well, just check Kattis problems for instance.

But secondly, while the problems in the beginning are trivial, you still have a huge advantage if you can immediately start programming without having to think, or having to look up a function in the stdlib. (Spoilers:) I probably wasted a minute today thinking about the solution of making a list for elves, and creating a new element whenever I saw an empty line. While a pro probably immediately did a split on "\n\n", summed and was done.

Re: Advent of Code 2022 is nigh

#65

Post by w-m »

If you want to join a small just-for-fun leaderboard to keep you motivated to go through to the last puzzle, feel free to join 194284-90c48b41

I've created this leaderboard last year and announced it here. It has a few people I know personally, and a few people from HN I've never met.

There's no community Discord or anything like that. The only form of communication are the stars appearing after other people's names after a while :)

Re: Advent of Code 2022 is nigh

#66

Post by nso95 »

pedrosorio wrote: post #58
Jtsummers wrote: post #49

Earlier quoted context omitted.

1. Have everything in place at the start (a script to grab the inputs helps). 2. Use a language that makes the edit/run cycle short. 3. Be very familiar with that language and how to process input in it. 4. Have solved many similar problems in the past so you can parse these problem statements quickly. 5. Have solved many similar problems in the past so you know exactly what code to write or use (if you have a utilit…

“Have solved many similar problems in the past so you can parse these problem statements quickly” This sounds like an advent of code specific thing. The problem itself is trivial, no need to have solved anything similar in the past, but the description is so verbose (because it’s a cute Santa-related story) that it becomes hard to parse the problem we’re trying to solve, quickly. I don’t think this is common in any o…

It helps if you sort of read the problem statement backwards

Re: Advent of Code 2022 is nigh

#68

Post by ofrzeta »

Jtsummers wrote: post #49
matsemann wrote: post #44

I wonder how the persons on the leader board manages it? First person today solved part1 after 39 seoncds, and was done with part2 after 53 seconds! I felt I was quite fast in being done with part1 after 3 minutes and part2 in 4 minutes (rank ~1000). I even tried to skimread the explanation to be fast, have a script to download the input and run it etc., but still no match!

1. Have everything in place at the start (a script to grab the inputs helps). 2. Use a language that makes the edit/run cycle short. 3. Be very familiar with that language and how to process input in it. 4. Have solved many similar problems in the past so you can parse these problem statements quickly. 5. Have solved many similar problems in the past so you know exactly what code to write or use (if you have a utilit…

> 1. Have everything in place at the start (a script to grab the inputs helps).

Definitely. A lot of it is always the same (open files, iterate over input, operating on arrays/lists etc.). Something like Peter Norvig has in Python:

https://github.com/norvig/pytudes/blob/main/ipynb/Advent%20o...

Re: Advent of Code 2022 is nigh

#69

Post by makeitdouble »

birdfood wrote: post #27
javajosh wrote: post #3

I did several of these last year and found them deeply unpleasant. I like puzzles but some puzzles are just...bad. Like the ones that require you to just brute force a solution, or write lots of detailed code to capture the requirements...that shouldn't be what this is. Just my opinion, of course, but like I said I found last year an unpleasant chore that felt more like work. Hope it's not like that this year.

I felt similar last year. I was trying to do them in Zig and Rust, and it felt like most of the challenge was just writing custom data parsers for poorly formatted data which felt too much like my day job.

> felt too much like my day job

Agree.

I think it helps to do it in a language you're not familiar with, or with some artificial limitation on how to solve it (no third party libraries for instance).

It takes more time, but felt less like a grind for me.

Re: Advent of Code 2022 is nigh

#70

Post by nikolay »

hubraumhugo wrote: post #36

I wonder how helpful GPT-3 and Copilot will be for this :) Not necessarily to solve the whole puzzle but rather to give hints when I'm stuck with a particular task.

Last year when I was doing AoC in Go, I was pleasantly surprised how good Copilot is and it felt like magic. Given most of the time I use Terraform, YAML, JSON, jq, and Bash these days, I can't really appreciate Copilot, but last year I did!
Post reply on HN