- Sea, land, waves, clouds, cables, tubes and packets are drawn by the
graphics card; the land is geometry built once, so panning and zooming
repaint nothing.
- Cities, labels and markers are painted once into a sprite atlas and drawn
in one call (with drawImage in 2D, which speeds up the 2D map too).
- Settings > Map drawing: Auto or Always 2D, per device. Browsers without
WebGL2, a lost context and puzzles keep the 2D map.
- npm run bench:gpu compares 2D and WebGL in a real browser on a real GPU;
?fps shows a frame meter.
The MCP server gains a second session, --browser: the game in the player's tab, reached through the
agent relay npm run dev hosts at /agent (loopback unless --host; Host and Origin checked, no token).
The player turns it on in Settings → Agent API, a section only the development page has, and picks the
clock: a pace of 30 s to 5 min a day that nobody can pause, or the agent's own. Headless gets the
running clock too (--clock realtime --pace MS) and both get a wait tool.
All of it lives in src/agent/ (development page only). The game offers HOOKS, UI.clockLocked,
setSpeed() and runClock(dt), the frame's day block moved out unchanged, and nothing more: the page
players get grows by under 1 KB gzipped and runs exactly as before.
scripts/mcp.js registers the same tools on @modelcontextprotocol/server (2.4.0, with
@modelcontextprotocol/node 2.1.2 for node:http; dev dependencies, pinned) instead of its own JSON-RPC
loop. stdio works as before; --http [HOST:]PORT serves Streamable HTTP at /mcp, on loopback unless
told otherwise.
Over HTTP every request's Host must be an address of this machine (or --allow-host), an Origin other
than localhost is refused, and a bearer token (SOT_MCP_TOKEN or --token-file) is optional; bound
beyond loopback without one, the server warns. load_save and export_save take JSON only there, or files
under --save-dir and nowhere else. Calls from every client run one at a time, in order.
scripts/mcp.js speaks MCP over stdio and loads the game from src/ in jsdom,
seeded, like the tests. Tools: new_game, load_save, export_save, status,
advance (with per-day timings), network, city, find, traceroute,
peering_check, hot_spots, actions, act, screen, panel, click, open_link.
Actions go through ACT as buttons do; files, slots, settings and the clock
stay out of reach, and nothing evaluates code in the page. Nothing is
hosted and the page is unchanged. See docs/MCP.md.
npm test ran every file one after another (--test-concurrency=1, from when there were three files). It now runs as many files at once as there are cores (TEST_JOBS=1 for the old way), and --shard i/n picks one of n parts of about equal time, balanced by tests/durations.json (npm run test:timings). CI runs four parts and the build side by side; a final 'test' job is green only when all are, so the check keeps its name. A newer push to a PR cancels the older run.
- ACT is split by area into game/90-actions/*.js (const ACT in the first file,
Object.assign in the rest)
- CITY_LL, CITY_NUDGE and LANDS_LL have one entry per line; the id-keyed ones
are sorted
- site/i18n/*.js are sorted, one entry per line; i18n-check enforces it and
--fix (npm run i18n:fix) restores it; scripts/i18n-add.js adds, removes
and renames entries in place
- scripts/legacy/port-branch.js moves a branch's site/index.html edits into
src/ during a merge with main
- perf workflow and PR template build the base page with scripts/lib/page.js
site/index.html (900 KB of HTML, CSS and JS in one file) becomes src/: a shell
(src/index.html), styles/ and game/, concatenated in path order by
scripts/lib/page.js. The assembled page is byte-identical to the old file, so
load time and in-game performance are unchanged.
- scripts/build.js replaces build.sh (same env vars, same output)
- npm run dev: a zero-dependency dev server that assembles on every request
- tests assemble the page in memory; npm test runs every tests/*.test.js
- perf-bench takes --base-ref to build a git ref's page
- scripts/legacy/split-index-html.js is the splitter, kept for older branches
hwMul() scales new purchases of routers (routerCost, swapCost), optics
(optCost, on top of tariffs) and own builds (dfBuildCost, ownCabCost):
free for 180 days, then +4%/year, capped at +30%. Monthly costs of what
you own do not move.
Shortages are a new random event after day 180, never two at once:
optics +50%, routers +40% or builds +30%, for 45-75 days. Counter-moves:
stock spare optics from a dark fiber link at today's price, or buy a
365-day price lock from the NOC (80k + 30k per POP, double during a
shortage), which freezes today's trend and ignores shortages.
Measured on a 16-POP European network ($530k/mo revenue): the curve
adds ~5% of revenue a year in year 1 and ~23% by year 5 if every dollar
is reinvested in hardware; a lock costs about a month of revenue.
Each transit flow's backhaul round trip over your own links (2× the path
latency to the exit) now affects service: free up to 20 ms (~1,500 km),
up to 20% worse at 50 ms, and half the service value past 50 ms. sfOf()
multiplies delivery by it, so revenue, broadband growth and business
customers all feel it. Peering, PNIs and caches are exempt.
City panel chip and Transit-tab warning name the far exit and the round
trip; the NOC lists slow cities. Two tests updated: Moscow on Frankfurt
transit is still fully delivered but now slow, and the Red Sea test pins
Asian demand that would otherwise churn away on Marseille transit.
Peering inbox (Networks tab): requests from small and made-up networks
arrive with every IX port, max-prefix alerts take bilateral sessions
down for a few weeks unless answered (raise the limit for growth, keep
it down for a leak; raising it on a full table is a route leak).
Unanswered mail costs reputation, which makes selective networks up to
10% pickier. The Large Peering Model agent costs $20M a year and gets
about a third of its decisions wrong, with a log to prove it.
PNIs are ordered, not instant: 2 to 6 weeks for small networks, 6 to 12
months for big content networks, Tier 1s and big eyeballs, for new PNIs
and augments alike (halved during a conference). Content networks and
Tier 1s start at 100G; Doitsche Telekomm sells 1G. Microsnoft refuses
PNIs and IX sessions in Ashburn, New York, Boston, Atlanta and Miami.
Only delivered PNIs carry traffic, bill, or count for goals and Tier 1.
npm run perf -- --browser times the same workloads in Chromium as a desktop
and as a phone (390x844 at 3x, CPU slowed 4x), one browser per version. Each
run collects garbage first and ends by forcing layout and the map canvas, so
deferred work counts. The jsdom DOM is about 30x slower than a browser's and
has no layout, so the interface rows need this.
Shared machines swing single runs by 30% on identical code, CPU time
included. A row now counts as slower or faster only when it moved 15% and a
Mann-Whitney U test gives p < 0.01 (15 runs a side by default). Identical
code: no flags in three full runs; a routing slowdown planted in a copy is
flagged in jsdom and on the phone.
The panel workload renders both selections every run instead of alternating,
which split its samples in two. CI drives the runner's Chrome through
playwright-core, no browser download.
scripts/perf-bench.js times routing, a game day, an action, panels, the
NOC dashboard and the map on tests/fixtures/big-network.json (104 POPs,
950 PNIs, 832 caches), base and head in alternating turns, and says
whether routing results changed. perf.yml runs it on every PR against
the base; perf-comment.yml posts the table as one updated comment.
New performance skill, CLAUDE.md rule, PR template and docs.
A network with a hundred POPs and a thousand peering ports froze the game for
most of a second every day and after every click, because routing checked all
~1,800 exits for each of ~14,000 flows. Now:
- compute(): exits are asked once per network which networks they reach, each
flow scans only those, and the first ECMP group is found without sorting. The
graph runs on city indices with a typed-array Dijkstra (same heap moves, same
ties), routers and per-link traffic are kept in typed arrays, the diverse-path
check reuses one live-link graph, and R.ef is put together when a link is read.
Results are bit-identical to before.
- Panels and the NOC dashboard are patched in place (morph) instead of being
rebuilt, so the browser no longer restyles 1,500 rows every tick. Buttons whose
action changed are swapped out, keeping #44's tap protection.
- The Networks tab works out each IX port's sessions once instead of once per
network.
- draw() reuses screen-point buffers, a per-frame pair index and cached label
widths, strokes the dotted fiber line on the path it already has, and uses
native roundRect: about a third fewer canvas calls per frame, same output.