ange-yaghi's engine-sim looks like an engineering tool and is emphatically not one. Its README says so in capital letters: it "is NOT a scientific tool and cannot be expected to provide accurate figures for the purposes of engineering or engine tuning." The stated purpose is to produce engine audio and simulate engine response, meaning the target is what the engine sounds and feels like at the throttle, not what a dyno would actually read. It was built for a YouTube demonstration, and the fidelity it chases is perceptual, not physical. That inversion is the interesting design choice, not a limitation. A simulator has a fidelity budget, and where you spend it decides what the artifact is for. Spend it on numerical accuracy and you get a tool engineers can trust and nobody enjoys. Spend it on believable sound and immediate response and you get something 9,500 people starred, because it delivers the experience of an engine. The move worth stealing is declaring the target out loud. By stating up front that the numbers are not to be trusted, the project sets the right expectation and frees itself to optimize for feel without pretending to a rigor it never wanted. **Cross-Domain Connections**: - [[Dual-Purpose Memory Writes Turn 16 Bytes Into a Fractal-Sound Generator]] — another case where the goal is a compelling generated sound, and the cleverness serves the experience rather than correctness. - [[Honestly Weak Beats Silently Absent as a Scoping Choice]] — declaring "this is not accurate" up front is the same honesty move: name the boundary instead of letting the user assume rigor that is not there. ## Source - [[20260802 Engine Simulator]] — ange-yaghi (GitHub) — https://github.com/ange-yaghi/engine-sim