unforum

[lpobjoy]

Nice work JayJay, I have been looking at this card and i think that each of the numbers is actually a sub command/ program for the overall program, so we may have to work out each sub command before finally running the main program and then solving this card.

How does that sound??

check this out, it's a Peit interpreter.
http://www.bertnase.de/npiet/

Lew

[donxkey]

this may be meaningless but:

What is Code Obfuscation?
There are many things that obfuscation could be:

Encrypted code that is decrypted prior to runtime.
Compressed code that is decompressed prior to runtime.
Executables that contain Encrypted sections, and a simple decrypter.
Code instructions that are purposefully put in a hard-to read order (often at the expense of execution speed).

[lpobjoy]

I believe Piet is Obfuscate code

[doublecross]

[baf]

Yeah, I found the whole business of the CC difficult to understand at first too. It's all about what happens when you have more than one possible next block to choose from. This is probably easier to explain if you call the absolute directions "north", "east", etc., and only use "left" and "right" for directions relative to the DP. The DP points in an absolute direction, the CC points left or right.

Suppose you have red, yellow, and blue blocks set up like this:

RR
YB

If you're on the red block, and the DP is pointing south, where do you go next? That's what the CC determines. Say the CC is pointing left. Since the DP is pointing south, the direction to its left is east. So, given a choice, you pick the one furthest east.

Fortunately, the program in the large grid never actually needs the CC to choose the next block.

[baf]

As requested, here's a run-through of the first ten blocks to illustrate how Piet works. I've decided that I like using compass directions better than up/down/left/right, so that's how I'll describe it.

At the start, we are in the block in the northwest corner. This is colored red and has a value of 9. The DP is pointing east. (The CC is pointing left, but we will never need to use it.)

1) 9 red. We didn't come from another block, so there's no instruction to execute. Since the DP is pointing east, we choose the next block by proceeding through the eastmost edge of this one, which happens to be on the north row.

2) 1 dark red. This is the same hue as the previous block, but 1 shade darker. This color change indicates the instruction "push", so we push the value of the block we just left onto the stack. The stack now contains 9.

3) 1 dark blue. (NB: Dark blue and medium blue are pretty much indistinguishable by eye. This is where a good scan and a paint program come in handy.) Hue changes by 4 steps, darkness doesn't change. The instruction is "duplicate". The stack now contains: 9 9.

4) 2 magenta. (It's easier to see what the darkness is here, because it borders on a dark magenta block.) Hue changes by 1, darkness by 2 -- since there are only three darkness values, 2 darker is the same as 1 lighter. "Multiply": the stack now contains 81.

5) 1 dark magenta. Color change: Hue 0, darkness 1. Push again. Stack: 2 81.

6) 1 light magenta. Hue 0, darkness 1. Push. Stack: 1 2 81.

7) 1 green. Hue 3, darkness 1. Pointer. So we pop the 1 off the stack (leaving 2 81) and rotate 1 step to the right. The DP is now pointing south, and just in time, too: we've reached the edge of the grid.

8) 8 cyan. Hue 1, darkness 0. Add. Stack: 83. Remember, we go through the southmost edge of this block.

9) 1 dark cyan. Push. Notice how everything with a value greater than 1 is immediately followed by a push? That's a good indication that we're doing this right. Stack: 8 83.

10) 1 dark yellow. Duplicate. Stack: 8 8 83.

Feel free to run the rest of the program yourself and check my results. Just watch out for those colors: medium and dark versions are often hard to tell apart. When in doubt, I loaded a scan into a paint program and examined the RGB values.

[ramsfan]

Baf, agree with stratman. Would never have stumbled across this. Don't expect I'll get anywhere with this, but it will be fun trying, and give me a break from sailing.

[Stratman]

Thanks baf
...and thanks doublecross and ramsfan - at least I don't feel like I am the only kid in the class who stuck his hand up when the teacher asked if anyone didn't understand!
(Maybe it will get me nowhere but at least I now understand the first step!)

[ramsfan]

Tangentially it reminds me of the bit in Capote when he talks about the Mondrian's being hung upside down. And thanks to Google for reminding us that today is the anniversay of someone else's birth (Miro, but skinheads won't be celebrating).

[doublecross]

Thanks - I understand now.

[anansi]

Wow...

that's a pretty funky program. I'd love to help, I've got the card but i don't have a scanner or anything. Are people using the pics on page three to run it?

[baf]

It looks to me like the challengedmosaic.png picture is accurate for the main program, although it's still really hard to tell some of the shades apart by eye. For example, there are two different shades of blue in the vertical bar on the right, but I can only see the seam if I enlarge the picture a lot.

[lpobjoy]

I am trying to use the Piet interpreter, the important thing it to get the colours right!!!

[JayJay]

Right here is my translation in PNG format of the numbered tiles.
The colours used are those defined by
http://www.dangermouse.net/esoteric/piet.html
(the pallete is on the image!)

The image below is probably not 100% but you are all free to use the image. One thing I ask if you feel that the colours are wrong and it needs to be changed please then post your corrected version! We need to get this 100 % right. Cos debugging could be one hell of a problem especially when dealing with similar colours!

On the image is the pallete I worked from. I have not done the main image yet. But I strongly suggest this should be created from the pallete on the image.

I agree that these numbered tiles may relate to sub commands and that a big image should be constructed with these tiless in. When we have such an image then we should run the program to see what we get.

(Oh the image was quite painstaking! I must now rest my fragile little mind)

[Stratman]

Not getting anywhere but what I have been trying since Piet came to light...
I am trying various ways to sum the reds on the main grid and/or on the small number grids.
My feeling is that you move that number of places along the border of words until you reach the number and pick that word. (The answer i am imagining is a nine word phrase)
Where to start? I have been thinking that for 1,2,3,5,8 you start at the word to the right of the number and for 4,6,7 and 9 start from the word to the left of the number. (Because of the way the numbers are arranged in their little grids - there must be a reason for that). Clockwise/anticlockwise?...Count the blank squares in the corners? There are a lot of variables for possibilities.