seen: 2026-10-30, cubes, words, on a phone. selim: "when i go in the words, it shows a list. this is a violation… the words, the language of the application can also represent itself in the cubes… each of these boxes that you have as a list, they can be actually as its own pieces… in each of them, when i go in, it could have the descriptions… each of those words, they have an example."
picture: none kept; the screen was a column of word cards (a word, a paragraph, a "show me" button) and a two-column table of the code's older words.
breaks: 2 (the cube is the viewfinder, not the data), 16 (every piece has the same shape: the cards had no faces), 24 (a face is a complete square), 29 (a cube opens on its shape) and the screen shape of 05-patterns.md; it led to principle 30.
why it feels wrong: the words were already a cube in all but form: every word had a meaning, an example and (for some) a code name, the same shape for every piece. laying them out as cards and a table showed the data, not the cube, in the one place where the app explains itself.
how it could follow: words opens on its faces, names first: names, meanings, examples, code names, each seen from afar. a word opens close up on its meaning, with its example (a tap shows it in the cubes) and its code name (where the journey ends).
status: fixed the same day, cubes#20.