This raises an interesting question: should we cre...

# advent-of-code

[David Whittaker]

This raises an interesting question: should we create a solution that solves our input or a solution that solves any input?

[Kroppeb]

If your solution works for your input it will probably work for similar inputs. Although a just did an older problem as testing and for part 2 I decided to leave the possibility that my code wouldn't give the right result. It was based on the fact that I assumed that my target node had at least 3 children as the case for 2 children would be way more difficult and still had a 50% success rate in case that my target had 2 children in which case I could swap children in case of 2 and get the right result at the cost of 1 minute which would be shorter than actually coding the thing.

[Kroppeb]

Which would have gotten me 63th place on the leaderboard and even higher if I noticed there where , in input in part 1

[Kroppeb]

So If you are going for implementation speed just go for quick and dirty.

[todd.ginsberg]

I (try to) write solutions that will solve any input that the AoC author provides to participants. Meaning - if the instructions say we'll get two wires, I don't write code for three. Similar for null checking, etc. Anything specifically mentioned as an invariant in the problem description is fair game, anything not mentioned I ignore.

[stkent]

And then a 2018day15 comes along and nobody's solution seems to work for every input 😰