Hacker News new | ask | show | jobs
by GregBuchholz 3260 days ago
>If the tile is allowed to tile the plane both periodically and non-periodically, the solution would be obvious.

Maybe you can draw a sketch for those of us who aren't immediately seeing the obvious?

2 comments

Squares, but each row of squares has a random offset
That is a periodic tiling (just think about what happens if you follow the tiling along the rows instead of along the columns, you'll periodically see the same pattern).
No, you wouldn't necessarily
https://en.wikipedia.org/wiki/Euclidean_tilings_by_convex_re...

It's listed as a periodic tiling. If you move the whole plane one square to the left or right you'll find that everything fits into place nicely, making this a periodic tiling.

Hm, I'm not knowledgeable about this subject but this seems like the offset on each row is NOT random, but it's instead repeating. Imagine instead that each time I added a new row, I chose the offset to be, say, a completely new value.
Doesn't matter, because your new row will still be offset with the same random value to existing rows after moving one square to the left or right.
Or triangles, with similar "slide along the seam" changes.
If you allow reflections of the single tile, an easy counterexample is the pinwheel tiling[1], which is a non-periodic tiling composed entirely of isometric triangles.

An (imo less satisfying) example which does not require reflection of the tile would be as follows: Take a rectangle with width equal to twice the height. You can use this tile to create squares which are either split vertically or horizontally. Put a single horizontally split square at the origin, then tile the remainder of the plane with vertically split squares: this tiling is (rather trivially) not periodic.

[1]https://en.wikipedia.org/wiki/Pinwheel_tiling