The Tile That Refuses to Repeat

The hat monotile sketch renders the 2023 aperiodic monotile at adjustable subdivision levels. Pan, zoom, and raise the level slider to watch local clusters harden into larger metatiles that never settle into a repeating wallpaper unit.

Some mathematical objects become famous because they look impossible before they become obvious. The hat monotile is one of those. In March 2023, David Smith, Joseph Samuel Myers, Craig S. Kaplan, and Chaim Goodman-Strauss posted An Aperiodic Monotile and answered a question that had been hanging around tiling theory since the 1960s: can a single connected shape tile the plane, yet do so only aperiodically? Their answer was yes. The "hat," a 13-sided polykite, fills the plane without gaps or overlaps, but it never yields a translationally repeating pattern. Waterloo's writeup of the discovery emphasizes the long arc from Berger's 20,426-tile construction to Penrose's two-tile sets and then, finally, to one piece. Quanta's reporting catches the part mathematicians seemed to enjoy most: the thing was found by a retired print technician cutting paper shapes and noticing that the pattern would not calm down into repetition.

What makes the hat interesting is not only that it avoids periodicity, but how it avoids it. The arXiv abstract gives the clean version. The hat forms clusters called metatiles; those metatiles admit substitution rules; and hierarchical substitution forces nonperiodic structure at every scale. That is exactly what this gallery artifact chooses to foreground. It does not present the hat as a static curiosity. It presents the hat as a machine for building larger and larger supertiles out of four metatile families, H, T, P, and F. Once you move the level slider upward, the sketch stops looking like a decorative tessellation and starts looking like a proof diagram that happens to be beautiful.

The underlying code is unusually faithful to the mathematics. Instead of faking the effect with a precomputed background texture, the sketch builds the hierarchy directly from affine transforms on metatiles. The key substitution loop comes straight out of gallery/hatmonotile.html:

function buildTiling(level){
  const hats=makeHats();
  let tiles=buildInitMetatiles(hats);
  for(let i=1;i<level;i++){
    const patch=constructPatch(...tiles);
    tiles=constructMetatiles(patch);
  }
  return tiles;
}

That loop matters because it encodes the whole argument in miniature. A periodic tiling would eventually repeat a local arrangement by translation. This one keeps reexpressing local arrangements as ingredients in a larger nonmatching hierarchy. The recursion is the point. Each pass through constructPatch and constructMetatiles turns visible tiles into instructions for the next scale, which is why the image feels stable up close and unsettled in the large. The sketch's pan and zoom controls make that legible in a way a print diagram does not. You can drag across a patch that seems almost regular, zoom out, and watch that apparent regularity dissolve into a larger asymmetric scaffold.

The color palettes do more than decorate. They separate metatile roles that would otherwise collapse into one another, letting the substitution structure read as process instead of surface. The stroke logic helps too. Thin outlines preserve adjacency information, while the low-contrast background keeps the eye on how shapes braid into larger units. This is good creative coding because it chooses the right abstraction level. The sketch does not try to illustrate every technical wrinkle in the proof. It gives you the operational heart of the result: one shape, finite local rules, hierarchical growth, and no repeating translational unit waiting at the end.

Penrose tilings already taught us that nonrepetition can produce its own order. The hat changes the emotional scale of that lesson. Two tiles still feel like a system. One tile feels like a dare. The gallery piece captures that difference. It takes a theorem that made international news in 2023 and reframes it as an interactive argument about scale, recursion, and the thin line between ornament and proof. Spend a minute with it and the discovery stops sounding like a novelty headline about a funny shape. It starts to feel like what it was: a long mathematical search ending in a tile that keeps saying no to repetition, no matter how far you extend the plane.