Research — Auto walkthrough (draft, unvalidated)

Crosswalk timing: the long street, auto

Sit back and watch six crossings. A pedestrian steps off as the clearance countdown starts and the cars wait at the stop line. A narrow street is crossed with time to spare; the same signal on a wider street runs out with her still in the lane; a longer interval, a refuge island with its button, or a longer walk interval fixes each in turn; and a survey with missing measurements never reports a pass. Pause any time, drag the clock on the timeline, or pick a story.

Story

Geometry: illustrativev1 draft, unvalidated harnessWalking speed: fixed at 3.5 ft/s, no speed axis

Plan · one signalized crosswalk, to scale in feet; cars are held at the stop line until the clearance countdown ends

WALK, then the flashing hand with the countdown, then the solid hand when the cars go. The walker steps off as the clearance countdown starts: how the two intervals are sequenced on screen is a presentation choice, the harness compares each number to its own limit and does not model the sequence.

Coloured dots are the exported stops, each read from a data cell. The small grey dots are the need table, with no verdict and never joined. The line between the dots is not drawn: the file holds nothing between two stops.

Check being read · click a card or a table row to look at it on the drawing

Actual

Required

Margin

How the harness read it

    Every check at this step · a failing row is highlighted

    There is deliberately no single crossing verdict. The accessible walk indication is set equal to the walk interval in this export, so it can only report not measured when the walk interval is.

    Every check · the steps this story compares, named

    Columns are the steps the data marks for this story, in story order and labelled with their own step names. Not measured means the survey has not reached the dimension, never a pass. Not applicable means there is no island in the scene.

    Precomputed results:every verdict, required value and margin on this page was computed by the real Python harness at export time and is played back from a data file. The page runs no harness logic and holds no standards tables, and it never interpolates between exported values. Not a live API call. The ada-harness is a v1 draft: no ADA consultant or accessibility specialist has reviewed it, and nothing here is a compliance determination. Every verdict, required value, margin, citation and reasoning line was computed by calling check_crosswalk directly at export time (harness commit a3895f4) and is played back from a data file. The function has one fixed maximum walking speed (3.5 ft/s, read from the data) and no speed parameter. The needed time, the speed a walker would need and the phase total are the exporter's arithmetic over the function's inputs and that constant; they decide nothing. Street widths, lane counts and every clearance and walk value are illustrative scene choices. The animation is a presentation: the walker steps off as the clearance countdown starts, the clock runs faster than real time, and the harness does not model the sequence. With a refuge island the function checks only the one timed distance; it does not evaluate a leg beyond the island. There is deliberately no whole-crossing verdict.Real function this maps to:check_crosswalk.Not modeled: pedestrian detection and call logic, signal phasing and coordination with the cars, a second leg beyond a refuge island, mid-block crossings, a different walking speed, accessible pedestrian signal sound and vibration, and any real intersection. The drawing is schematic.See also:Crosswalk timing, interactive with free play · One corner, auto · Curb ramp, auto sweep