thanks. it's been asked on SO I think, and there are a couple of other posts around the internet, but I find all of it very unsatisfactory. most answers are either "well, project your plane noise onto the sphere and deal with the artifacts" or "cut your sphere out of 3D noise", which is both quite suboptimal
(and I'm generally interested in data on the sphere since that's what I've been working on for the past couple of years in our cosmology department)
@trichoplax well, we need to go through commitment first
Well, I don't have any clue how to do that..... My solution: each elevator or floor object will have several values which will be modified any time an event occurs, so floor.buttonPushed will give what button was just pushed (if any).
and afaik it's actually an equal-area pixelisation for any number of samples (so it doesn't just tend to a uniform distribution as you increase the samples)
but you can really overthink this stuff. for circular zero (that game prototype, remember?) I needed to generate samples on the circle to monte carlo the score. so I thought I was really clever and did it in polar coordinates, with a non-uniform distribution in r to ensure a uniform distribution r dθ dr ...
well turns out, you're half an order of magnitude faster if you just generate samples on the square and discard those that lie outside the circle
so to generate uniform (random) samples on the sphere, I would maybe just generate samples on the cube, discard those outside the unit ball and then project them onto the sphere.
I wouldn't have guessed that circular zero used monte carlo for area. Not that I can think of a way of intersecting multiple arbitrary circles - it just hadn't occurred to me
@trichoplax yeah, I found this paper for intersecting 3 circles and gave up on it :D ... but I had the k-d tree already implemented, so testing whether a given point was coloured or not was pretty trivial (and not too expensive either)... so I decided to just monte carlo that, especially since I could run quite a lot of samples between when the mouse is released and when the wall has been completed (by doing only a bit each frame)
@feersum that's what I described above for the circle. in that case it turned out to be a lot more expensive than just sampling on a square and discarding.
I'm not sure how it would be if you need to project samples from the ball back onto the sphere though, since that is probably not exactly cheap either
I just realised that a sphere is only easy because projection from the ball to the sphere is always directly away from the centre. With any other implicit surface this isn't the case, so it made me wonder how on earth you'd distribute points evenly
GCD with strings
code-golf restricted-source
Objective:
Find GCD of two numbers, passed as strings.
Rules:
It must be a function with two arguments.
Both arguments are strings.
The return value must be a string as well, representing a positive number.
The strings may be assumed to be positi...
GCD with strings
code-golf restricted-source
Objective:
Find GCD of two numbers, passed as strings.
Rules:
It must be a function with two arguments.
Both arguments are strings.
The return value must be a string as well, representing a positive number.
The strings may be assumed to be positi...
meh... I just thought I had the best code golf idea... sort the digits of a number using only arithmetic... but then I (re)discovered this challenge from several months ago with the really broken scoring system: codegolf.stackexchange.com/q/25003/8478
but a handful of submissions ignored the weird bonuses and penalties and actually just golfed it without arrays and strings... those aren't competitive in that challenge at all, but I'm afraid that makes my idea a duplicate :/
@Lembik I think the trickier thing will be the actual enumeration and avoiding duplicates. you can rule out reflections by requiring that the snake starts to the right and first bends left, and that there are never more right than left turns. as for rotations I think you need a representation of the snake from its own point of view (in terms of turns) and be able to reverse that representation.
@aditsu "assuming some golfscript knowledge" :P
I also didn't expect ~ to be able to eval characters
and there's just no way to figure out what j does without reading the code ;)
CJam, 45
q~_,@m>0@{@(:T+@T*@+}/\;La{\)_@+@@md@@j@+}jp;
Finally I found a good use of j. But this is slower than the program below.
How it works
Long ArrayList Block j executes the block which takes an integer as the parameter, and Long j will call this block recursively in the block. It will...
@aditsu git is amazing. trust me, if you ever find the time to spend a week or two using it properly, you'll never want to go back to SVN or whatever you're currently using.
@MartinBüttner I get chat ping notifications on my phone, but I only see flags when I check the site (and I don't check meta flags very often), so if you want me, a chat ping would be faster :)
@MartinBüttner It's hard to tell what's competitive in that question - there appear to be open questions over the scoring of some of the better answers - but you probably already know that I'm of the view that being able to copy earlier answers with only minor changes makes it a duplicate whether or not the scoring system is strongly positively correlated.
GCD with strings
code-golf restricted-source
Objective:
Find GCD of two numbers, passed as strings.
Rules:
It must be a function with two arguments.
Both arguments are strings.
The return value must be a string as well, representing a positive number.
The strings may be assumed to be positi...
CJam, 45
q~_,@m>0@{@(:T+@T*@+}/\;La{\)_@+@@md@@j@+}jp;
Finally I found a good use of j. But this is slower than the program below.
How it works
Long ArrayList Block j executes the block which takes an integer as the parameter, and Long j will call this block recursively in the block. It will...
