Notes · Journal ·
Logbook 2026-10-23
This morning was all about the mascot: four cells in a square.
I wondered whether the state could be told by where the cells are, not only by what they contain. Attached at rest. Slightly apart. Apart and turned 45° into diamonds. Diamonds going round. Back at 0°, attaching and detaching. And moving from one state to the next had to feel natural: rest to searching, searching to thinking, thinking to building, and so on.
A few rules came out of the first try. The square holding the cells always stays upright; only the cells rotate. Outside rest and conclusion, one cell disappears, which leaves three cells and a free slot to move into. Searching became an assembly line, each cell sliding into the free slot in turn, like a sliding puzzle. Thinking became three diamonds, one of them doing a half-turn at a time. I also dropped the check mark at the end, without replacing it: I didn’t like it much.
Then came the motion details. The way the fourth cell vanished wasn’t great, so now it slides behind a neighbour and comes back out from behind it. It has to be fully hidden, or fully back, before anything else moves. And cells should never go from attached to apart in the same movement: first they move closer or further, then they do the rest. One movement at a time.
I also tried giving every cell the same thin black border. It didn’t work very well, so we went back to the borders the illustrations use. The last pass made the gaps and the hiding faster, with some easing for a more organic feel: cells snap into contact, and turns overshoot slightly before settling.