hanabi: rotate the pentagon top to match the seating #12

Merged
joeyr merged 1 commit from hanabi/mat-rotation into main 2026-08-16 23:08:19 -05:00
Owner

The image +y axis lands on TTS +z, so a mat drawn with its near edge at the bottom of the image arrives with that edge on the far side of the table. Every seat marker sat half a turn from the seat it named — the White marker across the table from the White player — and White faced a vertex instead of an edge, which is the one thing the five-sided shape was for.

Rotating the object rather than redrawing the art: the art geometry was already right (vertices at 54°, edge midpoints on the seat angles). Only the image-space to table-space mapping was wrong.

Worth noting how this presented: it looked like the deal was broken, because cards and markers disagreed. The deal was fine. The markers were lying.

The image +y axis lands on TTS +z, so a mat drawn with its near edge at the bottom of the image arrives with that edge on the **far** side of the table. Every seat marker sat half a turn from the seat it named — the White marker across the table from the White player — and White faced a vertex instead of an edge, which is the one thing the five-sided shape was for. Rotating the object rather than redrawing the art: the art geometry was already right (vertices at 54°, edge midpoints on the seat angles). Only the image-space to table-space mapping was wrong. Worth noting how this presented: it looked like the deal was broken, because cards and markers disagreed. The deal was fine. The markers were lying.
hanabi: rotate the pentagon top to match the seating
All checks were successful
lint / lua (pull_request) Successful in 18s
f669b3ca32
The image's +y axis lands on TTS +z, so a mat drawn with its near edge at the
bottom of the image arrives with that edge on the far side of the table. Every
seat marker sat half a turn from the seat it named -- the White marker across
the table from the White player -- and White faced a vertex instead of an edge,
which is the one thing the five-sided shape was for.

Rotating the object rather than redrawing the art, because the art's own
geometry was already correct: vertices at 54 degrees, edge midpoints on the seat
angles. Only the mapping from image space to table space was wrong.

Worth noting what this looked like from the outside: it read as the deal being
broken, because the cards and the markers disagreed. The deal was fine. The
markers were lying.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
joeyr merged commit 1bebf9252c into main 2026-08-16 23:08:19 -05:00
joeyr referenced this pull request from a commit 2026-08-16 23:08:21 -05:00
Sign in to join this conversation.
No reviewers
No labels
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
joeyr/tts-games!12
No description provided.