Hey, its been a long time.

philippeqc · Post by **philippeqc** » Sat Oct 20, 2007 10:03 pm

primes?

Fonkay · Post by **Fonkay** » Sun Oct 21, 2007 12:28 am

When I look at it I see some sort of picture. But I can't quite make out what it is. Looks almost like a map, or like a pile of leaves. My eyes start to ache before I can see the actual image.

Loor · Post by **Loor** » Sun Oct 21, 2007 6:05 am

To me it looks somewhat like a face...looking up towards the top left corner...

wrtlprnft · Post by **wrtlprnft** » Sun Oct 21, 2007 8:11 am

philippeqc wrote:primes?

Yeah, that was my first thought on why that one pixel (2nd from the top, 2nd from the left) might be special… It would also be an explanation why there are more white pixels in the top-left quadrant than in the others. But I still don't see the pattern.

Jonathan · Post by **Jonathan** » Sun Oct 21, 2007 8:37 am

It's indeed based on prime numbers. You might find a small amount of extra information by reading through the thread again.

Post by **Tank Program** » Sun Oct 21, 2007 12:37 pm

I see a quarter of a circle.

And sure, let's all hijack a thread!

Jonathan · Post by **Jonathan** » Sun Oct 21, 2007 5:51 pm

A circle?

I attached the largest version I could get under 2 MB. It happens to fill a small gap of uncertainty in the details of the solution.

Carnage · Post by **Carnage** » Sun Oct 21, 2007 10:52 pm

why do i see a dinosaur head on the smaller wen i de-focus?

im totally wrong, aint i.....

Jonathan · Post by **Jonathan** » Sun Oct 21, 2007 11:05 pm

Not wrong, just coincidental autosuggestion.

Post by **Tank Program** » Mon Oct 22, 2007 7:32 am

I see a headcrab now.

Jonathan · Post by **Jonathan** » Mon Oct 22, 2007 9:02 am

Heh.

Another hint: did you notice the change in overall brightness here?

Code: Select all

+--------------+
|              |
|              |
|=======       |
|              |
|              |
+--------------+

Doesn't matter which of the images. Also look for similarities between the two.

pavelo · Post by **pavelo** » Mon Oct 22, 2007 7:58 pm

attachment:
red) number of white pixels in jonathans image from 0 to x
blue) approximate number of primes from 0 to x according to the prime number theorem
green) the real number of primes from 0 to x

as you can see, green and red lines are nearly identical. could this be a coincidence ? i don't know
what i'm pretty sure is that the image isn't a (simple) encoding of primes, because it contains a lot of white-black-white-black-white subsequences, and it's impossible for those white pixels to all be primes. one thought that comes to mind is a random number generator that tries to emulate the distribution of prime numbers. but that too could be autosuggestion.

problem is that those functions are nearly linear, so if you placed the same number of white pixels completely randomly, the distribution would be nearly the same
but, then i tried to read the picture vertically, and the resulting curve was more linear than the previous (as it should be if my hypothesis is correct), so it just may be right. but then i don't know how does your hint fit into that

any comments ?

Jonathan · Post by **Jonathan** » Mon Oct 22, 2007 8:26 pm

My generator doesn't use random or semi-random numbers. It also doesn't approximate a distribution but is exact. As you can see the green and red lines match up exactly at the end, where ALL pixels have been taken into account without bias to some of them.

Website · Post by **Z-Man** » Mon Oct 22, 2007 11:17 pm

Wouldn't it be more coder-like to start the picture with 0 instead of 1?

Jonathan · Post by **Jonathan** » Tue Oct 23, 2007 12:12 am

With prime numbers there isn't really a 'right' place to start. Why not skip the still-useless one too? Only at two the pattern can begin to form. I did what made sense to me.