1
0
This repository has been archived on 2026-08-25. You can view files and clone it. You cannot open issues or pull requests or push a commit.
karti 85be0b13f8 feat: boards load before you ask for them
The owner asked for one map, with detail arriving in the background rather than
behind a wait. The interface already said one place (4bd8481: one ladder, no tab
strip, no boot card). This is the world catching up: on a desktop both metros are
now resident within seconds of landing, and the switch that used to be a build
becomes a fog dip over a cache hit.

**Measured, three runs each, California -> The Bay after an 8 s dwell:**
the longest blocking task falls from **521 ms to 58 ms** (before 700/501/521,
after 58/58/61) and long tasks from 2 to 1.

**And the cost, stated rather than omitted.** Over ten seconds sitting on the
state board: 601 frames -> ~571, p95 frame interval 18.0 -> 18.7 ms, and **one
long task of ~500 ms that did not exist before**. The build is not chunked yet,
so this moves a freeze from the moment of interaction to a moment nobody asked
for. That is a real trade and the follow-on that removes it is chunking the
build. `PREFETCH_ENABLED` in main.ts turns the whole lane off in one line.

A parked board draws nothing: 371 draws per frame before the lane lands a board
and 371 after, patched at the GL entry points. The performance budget reproduces
every cell to the digit.

**Three defects found on the way, two of them latent for a while.**

1. `environmentRig` had no idea what off-stage meant. `createScene` calls
   `apply()` unconditionally, two hundred lines above the `present` check, and
   the rig holds one probe per kind: on a key miss it convolves a new one,
   disposes the one the visible board is using, and repoints every applied scene
   at the replacement. A board built ahead of the camera is observed at its own
   centre — about forty minutes of apparent solar time across the state — so its
   key differs by construction. The board on screen would have changed colour
   because something invisible finished loading. Harmless until today only
   because every build was followed within ~800 ms by the swap that presented it.
   Fixed with `ApplyOptions.offstage`; the regression test was written first and
   failed first.

2. `mounting` was assigned and never cleared. Harmless while the only reader was
   the next `mountCity` wanting something to abort. It stopped being harmless the
   moment a second lane asked "is a foreground build in flight?", because the
   honest answer after the first mount of the session was permanently yes — the
   background lane armed exactly never, and the only symptom was a feature that
   silently did nothing. Found by measuring, not by reading.

3. `buildBoard`'s `onProgress` ends in `bootProgress`, which raises the switch
   pill. A prefetch would have put "Building The Bay Area... terrain 42%" over a
   visitor who did nothing — the exact chrome that removing the tab strip was
   for. `quiet` closes it.

**The policy is pure and lives with the eviction policy it has to agree with.**
`prefetchTarget` and `canAdmit` in boards.ts take scalars, never a camera, so
boards.ts keeps CONTRACT section 1's no-DOM/no-WebGL/no-three promise and
prefetchPolicy.test.ts can assert against the real pack bounds. Two regimes,
because collapsing them was the first version's mistake: with free room the gate
is simply "is there a slot", since the common path is to land on the state at
1,551 km and click a metro, and on that path a proximity trigger fires never;
with the cache full — the handheld case, capacity two with California pinned —
proximity is the only thing that justifies an eviction. The discs are asserted
non-overlapping against the shipped bounds (92 + 146 km of reach across a 314 km
gap), which is the covering-set argument residentCapacity already rests on.

**Also fixed, and separately load-bearing: reconcile's flag parser lied.** It
prefix-matched, so `?reconcile=palette` selected nothing and produced an empty
set — indistinguishable from the flag being absent. Every photograph taken to
judge a rule could have been a photograph of the unreconciled board with no way
to tell. Exact names now, with a warning that says "the flag is NOT off".

And two consumers were reading the raw packs beside a reconciled World:
`buildLadder` derives every rung's stand-off from `focus`, and `createMinimap`
was handed `entry.city` next to `handle.world`. Both now read `world.city`.
Latent with the flag off — `reconciledCity` returns by identity — which is what
kept it alive: it corrupts the measurement rather than announcing itself.

**What I did NOT ship, having tried it.** `roads` on by default. The rule exists
to stop California drawing 1,919 m freeways, and `city.roads` is not what draws
them: California is the one board with `roadTraffic`, so `scene.ts:816` takes the
`createFreewayWorld` branch and `createRoads` — the only reader of `Road.width` —
is never called for it. The visible corridor is a deliberate atlas glyph sized so
DRIVE mode can drive down it. Photographed at three chapters the rule moved
Downtown LA not at all (empty diff bbox), FiDi by RMSE 0.0006, California by
0.001% of pixels; the only measurable effect anywhere was -112 triangles on
Southern California. See DEFAULT_RULES for the whole argument.

1,688 + 295 tests, every gate, chapter-identity OK against the unmodified
fixture, ui-smoke PASS, budget PASS with private-request checks clean on all ten
cells.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-23 13:15:37 -07:00
2026-08-19 03:06:01 -07:00

Tera

The map view of Lumbridge Simulate — California from above, in three.js. Its other half, Spaces, is the offices you walk into: one engine and one asset library, seen from outside and from inside.

Apache 2.0. Runs at tera.lumbridgecorp.com.

San Francisco


What it is

An engine plus data packs. The default board joins Los Angeles and San Francisco with live deterministic traffic on US-101 and on the honest I-5 → I-580 → I-80 approach. Choose either route chapter to follow the procedural black Model X.

The engine renders terrain, coastline, built cities on authored street grids, bridges, roads, markers, road traffic and air traffic. A city pack is pure data — coastlines, hills, districts, landmarks, camera chapters — so adding a city is a data contribution anyone can review, not a fork.

California, the detailed Bay Area, and Los Angeles / Orange County / Riverside ship today. The corridor is intentionally sparse; detailed cities remain their own boards rather than forcing a 600 km world into one full-resolution mesh.

A plan view sits top right: the board drawn flat, with the footprint of the camera's own frustum on it, so you can see where you are looking from outside the shot. Click or drag it to move the camera; scroll it to dolly. It is a 2D canvas rather than a second WebGL context, drawn from the same city pack, and it follows the sun into the night along with everything else.

Who sees what

Three tiers, resolved once at boot by src/access.ts:

anonymous signed in admin
the map, the plan view, the named chapters
observed weather and live aircraft
the office public depth — shell, furniture, viewpoints, nobody home full depth, with presence full depth
the marker feed per TERA_MARKERS_ACCESS
the godmode panel (G) — date, season, weather override, counters, pose editor

The sky is public on purpose. Cloud cover over San Francisco is a government sensor reading, and the aircraft are broadcasting their positions unencrypted to anyone with a receiver; neither is something an account can grant you access to. Gating them cost the only moment that makes this project land — real fog rolling off the Pacific onto a city you recognise, at the real time of day, on a first visit.

The markers are the one feed that can carry something private, so the server decides. TERA_MARKERS_ACCESS is members by default and an operator has to say public out loud, which /api/v1/health then announces in degraded[]. The default is the safe answer rather than the common one, because the failure mode is silent: nothing errors, nothing looks broken, the data is just readable by the internet.

These are drawing decisions, not a security boundary, and src/access.ts says so at length. Live data and office presence are withheld by the API, from a caller it does not recognise; the client tier stops the app asking for something it will not get. Admin is granted only by TERA_ADMIN_SUBJECTS on the server — never inferred in the browser, and never from an API that failed to answer. A deployment with no API at all is open, because "clone it and it works" is the promise; it is not "clone it and you are an administrator".

Quick start

npm install
npm run dev

Play controls

The bottom mode dock is the local-player source of truth: View, Drive, Explore, Fly, or office Walk. A transition clears stale held input and atomically hands the follow camera to one subsystem. WASD is movement; Q/E is vertical or yaw, I/K pitches the crow, Space is the primary action, G glides, P resumes assistance, R resets, and C switches the driving camera. A standard gamepad maps both sticks, triggers, shoulders, and rising-edge action buttons.

Touch play uses a pointer-ID analogue stick at lower left and only the actions that apply to the current mode at lower right. The Map button remains available during possession. Touch, keyboard, and gamepad state are independent, so a released or cancelled finger cannot clear another source that is still held. The UI and follow camera are presentation adapters only; they never enter Arena observations, rewards, snapshots, traces, or simulator hashes.

Headless RL environments

Tera also exports a versioned, renderer-independent Arena contract with five deterministic environments: US-101/I-5 driving, Frontier Valley office navigation, seeded SF/LA office robot jobs, crow waypoint flight, and California electric-aircraft flight. They share the client controllers and office plan, but require no canvas, DOM, Three.js scene, network service, or new runtime dependency.

Import them from @lumbridge/tera/arena. Seeded train/dev scenarios, component rewards, safety terminals, maximum steps, snapshots, checksummed traces, exact replay and executable inaction/scripted baseline proofs are documented in ARENA.md.

The visible SF and LA office robots use that same fixed-step job state. Their patrol, parcel, inspection, and charging loops are authored demonstration scenarios—not presence, telemetry, or evidence of real company work—and the UI labels them as a seeded simulation.

Using the engine

import { createScene } from "@lumbridge/tera/engine/scene.ts";
import { createStage } from "@lumbridge/tera/engine/stage.ts";
import SAN_FRANCISCO from "@lumbridge/tera/cities/sf.ts";

// One stage per canvas, for the life of the page. Cities are put on it and
// taken off again; a renderer per city leaks its shadow map on every switch.
const stage = createStage(canvas);

const scene = await createScene(stage, {
  city: SAN_FRANCISCO,
  markerPalette: { hiring: 0x4ade80, closed: 0xef4444 },
});

scene?.setMarkers([
  { id: "1", lat: 37.7765, lng: -122.4241, label: "Somewhere", colorKey: "hiring" },
]);

createScene is async because the heightfield is built in a Worker — half a million samples, about 730 ms on the Bay Area, and not on the main thread. It resolves to null if the build was abandoned through options.signal, which is what makes switching city mid-build cheap.

The engine renders Marker[] and looks colours up by colorKey in a palette you supply. It does not know what your markers mean — that mapping lives in your adapter. This is what lets one renderer serve a private map coloured by one scheme and a public map coloured by another, without either being a fork.

Adding a city

Write src/cities/<id>.ts exporting a City. Trace the coastline and parks by hand, place hills as radial peaks, and give each district its street bearing.

Two rules, and they are not stylistic:

  • Do not import geometry from OpenStreetMap. OSM and Nominatim output is ODbL — share-alike, and incompatible with this repo's licence.
  • Do not commit logos or brand assets. They are trademarks, not code.

See ARCHITECTURE.md §3 for the full reasoning, and NOTICE for the attribution and data-provenance statement, and PROVENANCE.json for the machine-checked shipped-artifact and original procedural-lineage ledger. Run npm run provenance, npm run licenses, and npm run sbom before accepting assets or dependencies.

Aircraft

The engine takes a FlightSource. Two ship here: SimulatedFlights (original, flies real approach and departure corridors) and AdsbFlights (open community ADS-B feeds such as adsb.lol).

FlightRadar24 is deliberately absent — their terms forbid scraping and forbid redistributing their data, so a client for it cannot live in an Apache-2.0 repository. Commercial sources belong in private deployments. The best long-term answer is an RTL-SDR receiver: first-party data with nothing to comply with.

Layout

src/engine/    renderer — terrain, blocks, structures, markers, flights, scene, minimap
src/cities/    data packs — pure geography, no code
src/transport/ serializable route packs and renderer-independent simulation
src/arena/     versioned headless RL contract, scenarios, traces and environments
src/assets/    original procedural asset library
src/adapters/  where outside data plugs in
src/tools/     instruments — god-only, dynamically imported, never statically

engine never imports cities; neither imports adapters.

Nothing under src/tools/ may be reached by a static import from the app. It is loaded by one await import() behind access.can.debug, so a visitor who is not an admin does not download the code at all — which is the strongest available reading of "nothing here runs for a non-god visitor": not a hidden panel, not a disabled panel, no panel. src/tools/index.ts states the rule and what silently undoes it.

Licence

Apache License 2.0 — see LICENSE and NOTICE.

The ordered build plan and parallel work lanes live in BUILD_PLAN.md.

S
Description
Immutable Apache-2.0 Tera baseline through 2026-08-24; current development is proprietary.
Readme Apache-2.0 9 MiB
Languages
TypeScript 93.3%
JavaScript 5.3%
HTML 1.3%