feat(aterm)!: retune the creature background to Kai's numbers

coilyco-ops commented

2026-08-29 15:58:56 +00:00

Owner

#1417 shipped 0.91 tint, 26% of the window, bottom right. Kai read those
against the tuner and picked 20% opacity, 60% width, top right, so the
creature stops being a watermark and becomes the thing the window is
built around.

What moved

creaturePresence 0.09 to 0.20 and creatureWidthShare 0.26 to 0.60 are
constants, and the plate cache name already hashes the geometry, so the
old plate cannot survive the change. The anchor was the one that was not
a number: composeCreaturePlate hardcoded the bottom, and the origin now
takes the top.

A square creature at 60% of a 16:10 width is 96% of that height, so 8% of
the plate will not fit above it and the top inset takes the 4% left. That
is a clamp rather than a setting, and it decides what a crop costs.

cscaled covers rather than fits, so a window wider than 16:10 crops a
band off each end: on 16:9 that is 5% a side. The inset absorbs the top,
and the feet lose the rest. Rendered at both aspects rather than reasoned
about: the head survives on 16:9 and the feet clip. Protecting the head is
deliberate, since a clipped head reads as broken where clipped feet read
as standing on the edge.

Neither a smaller share nor scaled in place of cscaled is taken here.
The first is the number Kai just chose, and the second ends the crop by
stretching the art 11% on 16:9, which is the one thing the seat-creature
work has spent the most effort not doing. docs/aterm.md carries both as
the exits if the crop turns out to matter on a real external display.

Verification

go test ./... green, with the plate geometry, the anchor, and the tint
assertions moved to the new numbers.
pre-commit run --all-files green, docs/aterm.md at 7225 of 8000.
The compositing rendered at 1440x900 and 1600x900 from the real baked
plate, which is how the 16:9 crop was measured rather than predicted.

Co-authored-by: Kai Siren coilysiren@gmail.com
Co-authored-by: Claude noreply@anthropic.com
Agent-Role: frontend

#1417 shipped 0.91 tint, 26% of the window, bottom right. Kai read those against the tuner and picked 20% opacity, 60% width, top right, so the creature stops being a watermark and becomes the thing the window is built around. ## What moved `creaturePresence` 0.09 to 0.20 and `creatureWidthShare` 0.26 to 0.60 are constants, and the plate cache name already hashes the geometry, so the old plate cannot survive the change. The anchor was the one that was not a number: `composeCreaturePlate` hardcoded the bottom, and the origin now takes the top. ## The inset stops being a choice at this share A square creature at 60% of a 16:10 width is 96% of that height, so 8% of the plate will not fit above it and the top inset takes the 4% left. That is a clamp rather than a setting, and it decides what a crop costs. `cscaled` covers rather than fits, so a window wider than 16:10 crops a band off each end: on 16:9 that is 5% a side. The inset absorbs the top, and the feet lose the rest. Rendered at both aspects rather than reasoned about: the head survives on 16:9 and the feet clip. Protecting the head is deliberate, since a clipped head reads as broken where clipped feet read as standing on the edge. Neither a smaller share nor `scaled` in place of `cscaled` is taken here. The first is the number Kai just chose, and the second ends the crop by stretching the art 11% on 16:9, which is the one thing the seat-creature work has spent the most effort not doing. docs/aterm.md carries both as the exits if the crop turns out to matter on a real external display. ## Verification - `go test ./...` green, with the plate geometry, the anchor, and the tint assertions moved to the new numbers. - `pre-commit run --all-files` green, `docs/aterm.md` at 7225 of 8000. - The compositing rendered at 1440x900 and 1600x900 from the real baked plate, which is how the 16:9 crop was measured rather than predicted. Co-authored-by: Kai Siren <coilysiren@gmail.com> Co-authored-by: Claude <noreply@anthropic.com> Agent-Role: frontend

coilyco-ops added 1 commit

2026-08-29 15:58:56 +00:00

feat(aterm)!: retune the creature background to Kai's numbers

ci / aos-cli-tests (pull_request) Successful in 32s

Details

ci / gate (pull_request) Successful in 53s

Details

28fe98387a

#1417 shipped 0.91 tint, 26% of the window, bottom right. Kai read those
against the tuner and picked 20% opacity, 60% width, top right, so the
creature stops being a watermark and becomes the thing the window is
built around.

## What moved

`creaturePresence` 0.09 to 0.20 and `creatureWidthShare` 0.26 to 0.60 are
constants, and the plate cache name already hashes the geometry, so the
old plate cannot survive the change. The anchor was the one that was not
a number: `composeCreaturePlate` hardcoded the bottom, and the origin now
takes the top.

## The inset stops being a choice at this share

A square creature at 60% of a 16:10 width is 96% of that height, so 8% of
the plate will not fit above it and the top inset takes the 4% left. That
is a clamp rather than a setting, and it decides what a crop costs.

`cscaled` covers rather than fits, so a window wider than 16:10 crops a
band off each end: on 16:9 that is 5% a side. The inset absorbs the top,
and the feet lose the rest. Rendered at both aspects rather than reasoned
about: the head survives on 16:9 and the feet clip. Protecting the head is
deliberate, since a clipped head reads as broken where clipped feet read
as standing on the edge.

Neither a smaller share nor `scaled` in place of `cscaled` is taken here.
The first is the number Kai just chose, and the second ends the crop by
stretching the art 11% on 16:9, which is the one thing the seat-creature
work has spent the most effort not doing. docs/aterm.md carries both as
the exits if the crop turns out to matter on a real external display.

## Verification

- `go test ./...` green, with the plate geometry, the anchor, and the tint
  assertions moved to the new numbers.
- `pre-commit run --all-files` green, `docs/aterm.md` at 7225 of 8000.
- The compositing rendered at 1440x900 and 1600x900 from the real baked
  plate, which is how the 16:9 crop was measured rather than predicted.

Co-authored-by: Kai Siren <coilysiren@gmail.com>
Co-authored-by: Claude <noreply@anthropic.com>
Agent-Role: frontend

coilyco-ops referenced this pull request

2026-08-29 16:00:37 +00:00

Confirm or retune the aterm creature background numbers against a live window #1418

coilyco-ops merged commit 8b1fc7f37c into main

2026-08-29 16:00:38 +00:00

coilyco-ops referenced this pull request from a commit

2026-08-29 16:00:39 +00:00

feat(aterm)!: retune the creature background to Kai's numbers (#1419)

coilyco-ops referenced this pull request

2026-08-29 16:06:10 +00:00

Confirm or retune the aterm creature background numbers against a live window #1418

Rows
Columns

feat(aterm)!: retune the creature background to Kai's numbers #1419

What moved

The inset stops being a choice at this share

Verification