Clojurians Log v2

Maybe there is a way of transforming the set of points, into an equivalent one, with points closer and smaller radius

And then check the surfaces

I considered starting from a 'center of mass', maybe see if 'all nanobots weighted same' and 'all nanobots weighted by radius' made any sense

I solved it now, by first picking the best corner, then from there looping both logarithmic and linearly till it doens't improve anymore.

Turned out best corner had little value, just searching from the origin went almost the same time. 3 cycles more to find the correct answer, but not the 6000 evaluations to find the best corner.

In my case, there is no neighbor of the best corner that has is in range of the same or grater amount of nanobots, so a walk is not possible I think

It's no walk I do, it's about 13 cycles of jumping all over the place from 2^0-2^30 and that in all detections.

that's the line: https://github.com/akovantsev/adventofcode/blob/master/src/adventofcode/2018/day22.clj#L102

hey folks! beginner clojurist here. If someone could give me a few pointers on how to improve my day 9 code, I would be very thankful. I can't seem to find a fast enough approach. https://github.com/gdevincenzi/advent-of-code-2018/blob/puzzle09/p09/zipper.clj

@gdevincenzi "trick" to day9 was deque (double-ended queue) data structure

I'm not sure how to implement one in a functional way, but will try to find out more about it

someone has a deque implementation as a lib. I used mutable java one: https://github.com/akovantsev/adventofcode/blob/master/src/adventofcode/2018/day09.clj

thanks 👍

was a very good reminder how actually careful you should be with mutable things, god bless :rich:

You don't need deque for day 9, it's just pointers forward, and sometimes you need to know witch state you where in to get back.