You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
feat(website): the stage rail is a completeness ledger (#1043)
* docs: spec — stage rail redesign as a completeness ledger
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* docs: plan — stage rail redesign as a completeness ledger
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* feat(website): stage rail copy becomes a four-claim ledger
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* feat(website): stage proof lines derived from the recording
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* feat(website): beat settle points, segment states, and the hold and closing cues
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* feat(website): stage publisher drives the segment bar and the checks
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* feat(website): the stage rail is a completeness ledger — segments, checks, one claim, one link, one proof
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* fix(website): stage proof counts the interrupts, pins every line, and traces the recording
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* refactor(website): beat map polish and a rewind test for the rail state
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* feat(website): stage stills carry the ledger
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* fix(website): stage rail — visible cue owns the pointer, segments are keyboardable, anchors resolve, hold and ledger layout
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* test(website): stage e2e follows the ledger rail
The scroll test now asserts the rail's anatomy: segment state and check
fill at the Tools boundary, the one hold line on its plateau, and at the
end the ledger at full opacity with four filled checks, the pointer, and
the install command the copy declares. A new test clicks the Persist
segment and polls act progress into the persist window.
The scroll-craft harness reported the hold line and the Render block
never reaching full opacity: their 30%-of-window default ramps left the
plateaus between the harness's eight sample points, and the ledger's
60% ramp-in only reached 0.43 at the last in-act sample (masked in the
report by a text-key collision with the Tools block). The hold line now
overshoots the threshold by 20% of the approve span with a 0.3/0.2 ramp,
the Render block arrives over the first tenth of its window, and the
ledger ramps in over a fifth of the render tail. A spec case models the
engine's ramps and requires every rail cue to sit on its plateau at one
of the harness's sample points.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* refactor(website): the ledger cue keeps real margin at the harness sample, and plateaus are asserted independently of it
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
---------
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
0 commit comments