You can't "plug in" a "certain value" anywhere, as there are no free variables. All this does is describe an infinitely large image full of utter garbage, and if you look in just the right location, you find the above image.

Conceptually, that thing is super simple and by far not as impressive as a quine. And school children can make quines.

But what's cool about it is that someone has actually found a combination of a pixel-formula and just the right location in its output, the one piece of garbage that looks like the original pixel-formula. You see, you could have a computer just scroll through random formulas until it detects an image that looks right, but if you have a look at the Y coordinate at which you find Tupper's graphic ... that thing has over 500 digits! Trying out so many combinations would take for-fucking-ever! I don't mean forever as in "at this altitude it takes forever to boil an egg". I mean forever as in "I wonder if this calculation has any chance of finishing before the cold death of the universe."

So ... how did Tupper do this? Does anyone know?

MrPumpernickel is wrong, it's not a true Klein Bottle because it intersects itself. Of course you have to imagine it as the shadow of a 4-dimensional object ...

Now it's blindingly obvious that the candy is falling out left and right, as this insufficient surface can't possibly contain them. Alright, so it's not obvious. But picture a few points (drawn as dots all the same size) on a piece of paper, and surrounding them with one other dot. You get the idea.

Beajerry is of course right, it _appears_ to be an ordinary bottle with the spout connected to the side. But if it really was, how in the world would they have gotten the M&Ms in? Huh? Huh?

It's not too difficult to find pictures of unfilled bottles that let you see the entire topology, so I'll leave this as an exercise to the reader.

I figured the dog wants to pee as high as other, taller dogs, in order to fully cover up their markings.