DNA tiles drift in a test tube and stick wherever their edges match. Under the right conditions they grow a crystal that computes a pattern, tile by tile. Under the wrong ones they make nothing, or a heap of noise. The map shows every condition. Move the dot and watch what grows there.
Drag the dot and the crystal reacts as you move. Regrow starts a fresh one at the current spot.
Two things set the outcome: how hot the tube is and how many tiles float in it. Left to right gets colder, so matching edges hold tighter. Bottom to top gets scarcer, so tiles arrive less often.
The direction from the bottom-left corner sets the balance. Along the upper solid line a tile held by two matching edges just barely stays; above it nothing holds. Along the lower dashed line even a single matching edge is enough, so wrong tiles stick too; below it everything sticks. Ordered growth lives between the lines, and it is cleanest close to the upper one.
The distance from the corner sets the pace. Near the corner everything happens fast and the tiles can barely tell right from wrong. Far from it every choice is sharp, but the crystal takes days or years to grow. Life needs both: careful enough to keep its pattern, fast enough to finish.
Every arrival and departure of a tile is played out in random order with the correct odds (Gillespie’s method). A tile held by two matching edges stays far longer than one held by a single edge, and that difference is all the crystal has to tell right from wrong.
The map is measured, not drawn. Each of its 7,200 points is a separate run of this simulation, grown to 700 tiles, scored by its share of mismatched edges and by how long it took. Dark tiles carry a 0, light tiles a 1, and red tiles are mistakes.
The thresholds that separate the regions (1 mistake in 100 tiles, a thousand tiles within about three hours) are a choice; the gradients underneath them are not.