@Lembik Input wouldn't be a flat map.

ok so for what sort of input is the output uniquely defined?

that's really what I am asking

I'm considering scrounging up the actual topographic/elevation data of the United States. :D

ok but you may find that posed as a CS problem there is not one unique solution

That's fine. It just has to be reasonably good.

as in a reasonable approximation to some solution you know already?

One I'll find, yeah.

will you measure the L_2 distance to your known solution?

It's already been done.

[shrug] I don't know yet.

it's been done as a coding problem or God has done it already with rain :) ?

both :P

I learned about it a couple weeks ago. It's a fascinating problem.

oh I just saw christianity.stackexchange.com/questions/642/… :)

ahh, lol

if you have a reference for where it has been done as a coding problem that would be very interesting

not so interested in God :)

I was introduced to it in a book named "Group Theory in the Bedroom, and Other Mathematical Diversions" by Brian Hayes. The author actually did write code to solve this problem. I can assure you it's been done.

He wasn't the only one either.

oh I see

@feersum Hmm. Let's say that task A is to do n = n*n % m10,000 times, given a seed of n=1 and m=395891 (a prime number) and that task B is to do the actual challenge. Let's further say that Python takes 1 second to do task A and the given solution takes 10 seconds to do task B whereas C takes 0.1 seconds to do task A and the given solution takes 5 seconds to do task B. Then Python's ratio is 10/1 = 10 and C's ratio is 5/0.1 = 50. Clearly, Python has the superior algorithm here.

americanscientist.org/issues/pub/dividing-the-continent

Your objection, as I understand it, is that if the metric were something else, we would have different ratios. I don't think that's a problem because Python should still beat C relatively.

@El'endiaStarman That's not clear at all. It's perfectly possible those times were achieved using an identical algorithm.

If feel people will need to read that article

@feersum Our investigations of sturdy squares have proven quite the opposite! :P

@El'endiaStarman ?

@Lembik That is exactly what was in my book. Brian Hayes collected his Scientific American articles and published it in that book.

@feersum Your C program beat the pants off of my Python program in every instance.

having read it.. I think the challenge is potentially great but is going to need a lot of careful thought and research

I would be more likely to answer a challenge that simply asked for the fastest code straight up, than one with a randomized scoring system.

"When I finally got a chance to write a program for this process, I found that the algorithm is exquisitely sensitive to the order of operations. Consider the situation just as the Pacific is about to reach the lowest point on the divide. If the Atlantic has not been raised in synchrony, then the Pacific waters will pour over the saddle point and flood part of the eastern basin, shifting the divide to an incorrect position. "

for example

@El'endiaStarman I used a faster language and a faster algorithm so I'm not sure what this is supposed to prove.

@feersum I thought you used a slower algorithm (at first, anyway).

@El'endiaStarman github.com/moovida/jgrasstools/blob/master/lesto/src/main/java/…

I'll stop now :)

You used an algorithm of filling the top and left sides first, then calculating the rest, right?

I'm fairly sure that's worse than my initial program.

@feersum Yes, you used a faster language. Python would never be able to win a fastest-code challenge if anyone posts with a faster language, like C, which is a great choice because it's so fast. I'm trying to figure out a way to compare languages fairly.

@feersum Didn't yours calculate all (n*m)! permutations?

No!!!!!!!

...I presume you're writing an explanation of what it does, then?

I can paste the original version's code if you want to see it.

Explain the idea.

I know you did some things with symmetries to get the number down.

The first version didn't use symmetries

But how did you not do O((n*m)!)?

I don't know what the time complexity would actually be

It fills in the cells in reading order

And goes back if it generates a 2x2 square that has the wrong sum.

I implemented also a version which uses 4 or 8 symmetries. I haven't made one which uses more symmetries.

There are over 10k symmetries for 6x6.

which problem are you discussing?

oeis.org/A263542

A263542 allocated for Lee Burnette ?

that's not very informative :)

Hmm that page used to be a lot more informative.

yeah... :P

I think you have to be logged in to see the draft.

Must be the wrong link

No you don't

nope

oeis.org/draft/A263542

It's probably a different URL to see the edit history etc

I can see lots of things there.

Like a history of 31 edits.

yes me too

This is the relevant bit:

> Number of MxN grids where 1<N<=M, all entries are distinct, all entries are at least 0 and at most M*N-1, and every 2x2 block of entries has the same sum.

@feersum you are really very good at combinatorics

Not really

I am not sure I have ever seen anyone better here at least

Example of harder combinatorics answer

oh I read that

Combo is hard...

Easier than other math

geometry is the worst

Competition geometry, I mean

I don't know how hard geometry gets at college, because I'm only three semesters into college math

three semesters... that's quite a long way

Geometry is the best way to feel like an idiot.

probability does that for me :)

That's another good one.

A guy made a fake proof that switching doors in the Monty Hall problem gives a 50% chance of winning and it took me around 20 minutes to figure out why it was wrong.

umm.. what is the correct answer?

1/3

actually the monty hall problem is a great example of why coding is great

err 2/3

it really helps to implement a simulation

I feel most people's confusion would leave them if they actually implemented it

@Lembik I once implemented it when I hadn't yet understood the proof. I then still didn't understand the proof, but believed in it a bit more.

@Pietu1998 :) That makes sense

hmm.. if there are no comments at all for meta.codegolf.stackexchange.com/questions/2140/… soon I may post it

@Lembik It's been in the Sandbox for only two hours...and those two hours are during PPCG's slowest times.

Give it a day.

@El'endiaStarman ah.. is there a graph for PPCG busyness by hour?

Look at the bars on the right side.

Those represent how many messages were made in this chat room over the course of a day.

that's great!

It is currently just about 12:00 pm UTC.

So yeah, like 5 people have seen your Meta post.

is chat room activity highly correlated with question/answer/comment activity?

Probably.

:)

it looks oddly bimodal

is that european evening then US evening?

That's only 1 day, not an average.

right

So it was probably an eruption of carrot spam or whatever.

oh I see.. I assume that European and US evening will be peaks though..no?

It's popular to use PPCG at work.

oh!

criminals :)

I got a comment!

although I didn't understand half of it :)

Carrot spam. :D

Spam for vegetarians.

what is the largest window/screen size that all the popular variants of BASIC will be able to support?

320/200 ?

@VoteToClose Hi. What frame/second do you need to make things smooth? 20?

@Lembik For your question, making the FPS range from 30fps up would be a good definition of "smooth". It should also have no visible flicker, so that's good enough.

20 fps is noticeably stuttery.

@VoteToClose ok.. does that then exclude lots of languages?

Not really. Some will do max framerate possible that is naturally supported, which for this situation could be several thousand frames.

I don't fully understand the use of fps here as I require that the sprite is shown at every position and I say how long it has to take to cross the screen

doesn't this already give an fps?

No - I could make it move across the screen in one jump. 1 frame per 2 seconds.

but then it isn't shown at every pixel position

hmm. what's a good way to phrase that?

I don't see anything saying that every pixel position should be used.

That's an unreasonable requirement anyway.

It needs to be shown at every pixel position? Okay, so, minimum fps rate would be several hundred.

@feersum it isn't there and should be.. but why is it unreasonable?

If the window is 2000 pixels wide and is crossed in 2 seconds, that requires 1000 FPS.

Monitors usually can do only 60.

^

hmm

But if it was a 200 pixel wide frame doing it in 5 seconds, that would be 40 fps, which would be fine.

(for monitors supporting that speed.

)

I have updated the quesiton

thank you voth

@Lembik As it stands, shouldn't the minimum be 40 fps for a 200 px wide screen?

You'd basically have to skip pixels to get lower.

@VoteToClose feersum persuaded me it shouldn't be shown at every pixel

Because 199/5 ~ 40.

Oh!

You might want to edit the comment stating "shouldn't jump pixels" then.

am I right in thinking that at 30fps say, crossing 2000 pixels in 4 seconds it will look smooth?

30 fps is smooth, alright. The eye won't notice the pixel jumps very well.

It'll be moving very quickly at that speed.

@VoteToClose done

@Lembik 16.666666667 pixels of movement per frame.

thanks very much for this

I set a minimum of 24fps :)

because I vaguely remember that cinema used to use that

Oh, okay. :P I'm sure you'll work out the details, I'm in English class right now, so I must go. :P Peace.

bye!

I agree with the comment about 200x200 being much too small. Especially if the image to move around is 207 pixels wide.

If you're trying to be inclusive of older languages and screen modes, the image should be much smaller, or even just a simple binary bitmap.

@Geobits ok... hmmm... 400 by 400 ?

I guess my real question is: Why is the image so big?

@Geobits can you find a small one?

with the right license

or can we just shrink that one somehow?