The rim becomes a wall: half a millimetre of shoulder where a cliff dropped roamers - #214
Merged
Merged
Conversation
…ropped roamers Three instruments hit the same failure in one day (issue #211): a long undisturbed roam reaches the dish edge, and the wall -- which engaged only in its last 0.05 mm -- loses. Reproduced deterministically: placed brushing the rim at 70 degrees off radial, the animal spent 0.21 s in the zone and pushed straight through; free roamers on the default and bare plates died on it at 113-152 s, roughly once per two minutes of roaming. Fifty microns cannot turn a body. The zone widens to half a millimetre (World.WALL_ZONE, mirrored in the runtime): same stiffness, ten times the reach and ten times the peak restoring force -- and a soft shoulder is what a real dish edge is anyway. The validation cliff moves to 0.5 mm BEYOND the rim so the wall has somewhere to do its work; sample() already clips its cell indices, so a nose momentarily past the rim reads the rim cell rather than laundering garbage, and past the margin it is a genuine escape and still fails loudly. The reproduced trajectory is now the regression test: the same grazing approach must be turned with the whole body back inside the dish proper and the margin never doing the containing (measured: peak 44.65 mm against the old straight-through). Every pinned assay runs at the plate's centre; conformance reproduces bit-for-bit, and no determinism pin moved. Closes #211. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01CBroMXcYdfkyb5buJPcKvL
Two checks were calibrated to the old 0.05 mm lip. The out-of-dish sample test probed 0.01 mm past the rim, which is now a wall-contained nose reading the rim cell; it probes both sides of the real failure line (the 0.5 mm margin) instead. And the population suite's touch-is-private case relied on three radial pressers bouncing off the narrow lip at different times to keep their contacts distinct; against the smooth half-millimetre shoulder they settle into near-identical equilibrium forces and the distinctness guard trips vacuous. The three animals now press at different obliquities, which gives each its own sustained normal force by geometry (spread 86% of the largest, against the 10% the guard demands). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01CBroMXcYdfkyb5buJPcKvL
YesterdaysLemon
marked this pull request as ready for review
August 27, 2026 08:55
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #211.
The failure
Three instruments hit it in one day: a long undisturbed roam reaches the dish edge and dies with
DivergentSimulation— the wall engaged only in its last 0.05 mm, and fifty microns cannot turn a body. Reproduced deterministically: placed brushing the rim at 70° off radial, the animal spent 0.21 s in the zone and pushed straight through; free roamers died on the rim at 113–152 s, roughly once per two minutes of roaming.The fix
World.WALL_ZONE, mirrored in the WASM runtime). Same stiffness, ten times the reach and peak restoring force — and a soft shoulder is what a real dish edge is anyway.sample()already clips its cell indices, so a nose momentarily past the rim reads the rim cell rather than laundering garbage; past the margin it is a genuine escape and still fails loudly.Verification
test_the_dish_wall_turns_a_grazing_animal_instead_of_dropping_it): the same grazing approach must be turned with the whole body back inside the dish proper and the margin never doing the containing. Measured: peak 44.65 mm against the old straight-through.🤖 Generated with Claude Code
https://claude.ai/code/session_01CBroMXcYdfkyb5buJPcKvL
Generated by Claude Code