# TDK Pinball Machine for modern PCs This is a native Rust reconstruction of the 1995 Windows game **TDK Pinball Machine**. It uses the extracted original artwork and sound, recreates the playfield as a fixed-step simulation, and runs from the same source on Linux, macOS, and Windows. The game opens at the original 640x460 canvas size and presents every artwork pixel one-for-one. The fixed, non-high-DPI client prevents desktop scaling from distorting or vertically offsetting the pixel-art presentation. The original program is not required at runtime. All required game assets are embedded in the executable at build time. ## Run Install a current stable Rust toolchain, then run: ```sh cargo run ``` For a distributable optimized binary: ```sh cargo build --profile production ``` The binary is written below `target/production/`. On Linux, Macroquad's native development packages are also required (X11, OpenGL, and ALSA). Windows needs no extra runtime installation; macOS builds with the normal Apple developer command-line tools. ## Browser version Build and serve the WASM website locally with: ```sh just web-serve ``` Then open . The browser build keeps the original 640x460 presentation centered on a black page and stores settings and a local fallback table in browser storage unless the optional same-origin `/api/highscores` service is deployed. See [web/README.md](web/README.md) and [highscore-server/README.md](highscore-server/README.md) for deployment details. ## Deterministic mechanics validation Named scenarios can be advanced without waiting in real time. The simulator uses exact 120 Hz steps, prints its final state as JSON, and can write both the complete step trace and the original-size 640x460 framebuffer: ```sh cargo run -- --simulate claw-6 --at 0.15 \ --trace /tmp/claw-6.json \ --screenshot /tmp/claw-6.png ``` Use `--step N` instead of `--at SECONDS` to reproduce one exact update. The available scenarios are `loading`, `attract`, `highscores`, `name-entry`, `effect-7`, `autoplay`, `launcher`, `flippers`, `panel`, `targets`, `claw-1`, `claw-6`, `claw-7`, and `claw-18`; `--seed N` fixes random choices. `autoplay` charges each ball and operates the flippers from live ball position for long end-to-end validation runs. ## Original and modern controls | Action | Original key | Additional modern key | | --- | --- | --- | | Start game / add up to 4 players | Keypad `+` or the main `+`/`]` position | `=` is also accepted | | Charge launcher | Hold Down arrow, release to launch | - | | Left flipper | Ctrl (left or right) | `A` or Left arrow | | Right flipper | Enter (main or keypad) | `D` or Right arrow | | Nudge | Space, Left Shift, keypad `3` | Right Shift aliases keypad `3` | | Help | F1 | Enter/Escape returns | | Settings | `TDKPIN.INI` before launch | F10 opens the portable settings screen | | High scores | Automatic when a player finishes | F9 opens the portable table viewer | | Sound | F12 | - | The five original speed choices remain available in settings. Physics now uses the original invariant 100 Hz millipixel substep, batched as 5/4/3/2/1 steps on the recovered 50/40/30/20/10 ms callbacks. Claw animation, flipper edges, and visible state publication follow the same selected callback cadence. The help screen is the original artwork in English, German, French, Italian, or Spanish. As instructed on that screen, double-clicking its upper-left exit box closes the program. ## Saved data Native settings and the ten-entry high-score table are stored as `save.json` in the platform's normal per-user application-data directory. Browser settings and the fallback table use `localStorage`; a deployed browser build uses the same-origin high-score service as its shared table when available. On first run, native settings are imported from an adjacent original INI (or the embedded distributed TDKPIN.INI), and the table is imported from the original `HISCORES.DAT`. ## Reconstruction status This repository distinguishes exact recovered material from behavioral reimplementation. See [RECONSTRUCTION.md](RECONSTRUCTION.md) for the evidence ledger, known inference boundaries, and validation performed.