Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
39 changes: 26 additions & 13 deletions ACCESSIBILITY.md
Original file line number Diff line number Diff line change
Expand Up @@ -5,7 +5,7 @@ and a set of live charts. That makes some of it hard or impossible to use withou
and some of it hard to use without a mouse. This page says what works today, what
doesn't, and which results are available without the visual app.

Last reviewed: November 2026, version 0.5.1.
Last reviewed: November 2026, for the browser version and the release after 0.6.0.

## Results available as text

Expand All @@ -18,14 +18,17 @@ cargo run --release --bin bench sensors all 20 4 # sensor study
cargo run --release --example corridor # corridor study
cargo run --release --example lpi # pedestrian safety treatments
cargo run --release --example cooperative # signal-free cooperative crossing
cargo run --release --example failures # failures at the cooperative crossing
```

These commands print the average and 95th-percentile delay, throughput, stops, person
delay, pedestrian waits and near misses, a corridor's end-to-end trips and the safety
counts. They don't print everything the app shows. Some numbers appear only in the app,
such as CO₂ per vehicle, the level of service, the breakdowns by approach, movement and
vehicle class, the charts and a selected vehicle's own figures, and only the
`cooperative` example prints the maximum delay.
counts. The `failures` example prints how often vehicles touched after a failure and how
fast, the emergency stops, the crossings withdrawn and the delay a failure adds. They
don't print everything the app shows. Some numbers appear only in the app, such as CO₂
per vehicle, the level of service, the breakdowns by approach, movement and vehicle
class, the charts and a selected vehicle's own figures, and only the `cooperative`
example prints the maximum delay.

The [README](README.md#results) and [docs/insights.md](docs/insights.md) report the
findings as text and tables, and every image in them has a text description.
Expand All @@ -40,9 +43,11 @@ findings as text and tables, and every image in them has a text description.
* **Zoom:** in the desktop app, Cmd or Ctrl with +, − and 0 enlarge, shrink and reset
the whole interface. In the browser, the browser's own zoom does the same.
* **Motion:** the simulation can be paused at any time, and its speed set from ×0.5 to
×100. Flashing signals (flashing yellow arrows, flashing DON'T WALK, turn signals) blink
×100. Flashing lights (flashing yellow arrows, flashing DON'T WALK, turn signals, the
light bars of emergency vehicles and the hazard lights of a disabled vehicle) blink
about twice a second at every simulation speed, below the three-flashes-per-second
limit for photosensitive seizures.
limit for photosensitive seizures. The red ring around a disabled vehicle pulses more
slowly, a little under once a second.
* **Screen readers, desktop only:** the desktop app exposes its controls to the
operating system's accessibility interface. On macOS it presents buttons, checkboxes,
sliders, dropdowns and the text of the results panel, and dropdowns and number fields
Expand All @@ -56,23 +61,31 @@ findings as text and tables, and every image in them has a text description.
* **The map and charts are images only.** A screen reader can't describe the vehicles,
signals, queues, signal timing diagram, line charts, or the cooperative crossing's
timeline of crossings and histogram of crossing gaps. The numbers in the results panel
are readable.
are readable, and so is the sentence it gives for each failure at the cooperative
crossing.
* **No screen-reader support in the browser version.** It draws everything on a single
canvas, so assistive technology finds nothing to read. Use the desktop app or the
command-line tools instead.
* **Pointer-only tasks:** panning and zooming the map, selecting a vehicle, placing a
pedestrian at a crosswalk corner and the charts' hover readouts need a pointer.
pedestrian at a crosswalk corner and the charts' hover readouts need a pointer. So
does choosing which vehicle of the cooperative crossing fails; the *Stall one in the
box* and *Stall one at the worst moment* buttons make one fail from the keyboard, and
what follows is reported as text in the results panel.
* **No shortcuts between panels:** reaching the settings means tabbing through the whole
top toolbar, about two dozen controls.
* **Color:** the *Color by delay* and *Color by speed* map views use color alone. The
side-by-side table marks the better result in green, though the numbers are always
shown.
* **Color:** the *Color by delay* and *Color by speed* map views use color alone. So do
the dots on the vehicles of the cooperative crossing, which show whether a vehicle has
its trajectory, is waiting for one, is disabled or has stopped for a disabled vehicle;
a disabled vehicle also has a ring and hazard lights. The side-by-side table marks the
better result in green, though the numbers are always shown.
* **Dark theme only**, and the app doesn't follow the system's light or high-contrast
settings.
* **Large screens:** the desktop window needs at least 1100 × 700 points, and the
browser version isn't designed for phones.
* **Fast changes at high speed:** at ×30 to ×100, small signal lights and detector
markers change color several times a second.
markers change color several times a second, and with *random stalls / h* turned up,
the rings that mark disabled vehicles and contacts at the cooperative crossing can
come and go as fast.

## Supported environments

Expand Down
5 changes: 4 additions & 1 deletion CONTRIBUTING.md
Original file line number Diff line number Diff line change
Expand Up @@ -106,7 +106,10 @@ pass that file's path in place of the two folders.
clock or another source.
* **Safety comes first.** The regression tests require zero vehicle overlaps, and most
arrivals served, on every layout, control mode and controller. A change that
introduces an overlap isn't ready, however much it helps delay.
introduces an overlap isn't ready, however much it helps delay. The one exception is
made on purpose: when a vehicle of the cooperative crossing is made to fail, vehicles
that can't stop in time touch, and each contact is recorded with its speed instead of
counting as an overlap.
* **Add tests with your change.** New functionality, such as a controller, a safety rule,
a change to the reward or new simulation behavior, comes with automated tests that
check it, and a bug fix comes with a test that fails without the fix. Put tests that
Expand Down
5 changes: 4 additions & 1 deletion GOVERNANCE.md
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,10 @@ This document describes how decisions are made and who does what.
* **By evidence.** Questions about realism are settled by what real traffic does, citing
sources such as the *Highway Capacity Manual*, the *MUTCD* or published studies where
possible, and questions about performance by benchmark results across several seeds.
A change that lets vehicles collide isn't accepted, however much it helps delay.
A change that lets vehicles collide isn't accepted, however much it helps delay. The
one exception is made on purpose: when a vehicle of the cooperative crossing is made to
fail, vehicles that can't stop in time touch, and each contact is recorded with its
speed.
* **Through pull requests.** Every change to `main`, including the maintainer's own,
goes through a pull request whose required checks pass: the tests on Linux, macOS and
Windows, test coverage of at least 80% of lines, the browser build, rustfmt and
Expand Down
Loading
Loading