A physics education that doesn’t depend on where you are

Most people never get to play with physics. They get equations to memorise and problems to grind, and if the intuition arrives at all it arrives late. Meanwhile the tools that would build that intuition (run the experiment, change one number, watch what happens) have been possible in a browser for fifteen years and are still scattered, uneven, and mostly unexplained.

This is an attempt at the version that isn’t: simulations that are actually correct, wrapped in lessons that know which specific wrong idea they are trying to dislodge, free to anyone with a browser, with no account and no tracking.

How the simulations work

Each one integrates the real differential equation of the system with a named numerical method: fourth-order Runge–Kutta where accuracy matters most, velocity Verlet where long-run energy conservation matters more. No game physics engine is involved anywhere. Engines like Rapier and Cannon are built to make motion look plausible under contact constraints, and they freely apply position corrections that violate the equations of motion. That is exactly the wrong trade for teaching: the whole point here is that a student who works the problem by hand gets the number the screen is showing.

The integrators are checked in continuous integration against closed-form solutions: projectile range against v₀²sin2θ/g, the oscillator period against 2π√(m/k), orbital closure and energy conservation over hundreds of periods. If the numerics ever drift away from the algebra, the build fails.

Everything runs in your browser at a fixed physics timestep, so a 144 Hz monitor and a 60 Hz laptop produce identical results. Nothing is sent to a server, because there is no server to send it to.

How the lessons are built

Each lesson asks you to commit to a prediction before it shows you the result, and only at the end names the intuitions it took on. That ordering is the mechanism: committing to a prediction is what turns a surprising result into a reason to revise a belief instead of a fact to file alongside it.

Every design decision traces to published work, and the research page lists each source with what it found and what it changed here. Where the evidence is thin, that page says so.

What this borrows from

The idea that a simulation can be a primary teaching instrument rather than a garnish comes from PhET Interactive Simulations at the University of Colorado Boulder, whose two decades of design research underpin most of what is good here. The ambition to cover an entire curriculum with simulations rather than a handful of showpieces comes from 3JCN Physics, Thomas Nguyen’s library of over 350 browser simulations.

Reusing it

All content is licensed CC BY 4.0. Use it in a class, assign it, remix it, translate it, put it behind your own front page. Attribution is the only condition.

Every simulation can be linked at a specific configuration: open it in the sandbox, set it up, and use Share setup to copy a URL that reproduces exactly that state. It is the fastest way to hand a student the situation you want them to look at.

Status

Six units are here: classical mechanics, waves and sound, fluids and thermal physics, electricity and magnetism, optics, and modern physics, the ground a three-semester introductory sequence usually covers. Every lesson is written and every simulation runs its own equations.

What that does not mean is that it is finished. No student has used this yet, and student confusion is the only test that finds the things an author cannot see. If you teach with it and something misleads someone, that is the most useful thing you could tell us.