The post is a weekend-built interactive simulator for explaining why orbits exist, not just showing planets circling. It walks from Newtonian ideas like equal and opposite force, inertia, and "falling but always missing" into real solar-system data, a toggleable N-body integrator, and finally a visual introduction to Einstein’s curved spacetime. The author says the physics runs in real units, with only the display scale remapped so planets are visible, and that the whole thing is client-side in TypeScript and Three.js.
People were broadly enthusiastic because the project is approachable, fast, and genuinely useful as a teaching tool. The strongest feedback was not about polish but about scientific framing. Several readers said the simulator blurs the line between a conceptual animation and a physically correct model. The biggest issue was the Einstein ending. It presents relativity visually, but the actual simulation is still Newtonian, which can mislead viewers into thinking the tool has switched theories when it has not. Others flagged specific inaccuracies, including an animation that appears to show Earth’s axis precessing over a day instead of staying fixed in space through the orbit, orbital speed numbers that look far off for Earth, and at least one panel where force and velocity seem frozen when they should change.
The useful takeaway from the comments is that educational physics software gets judged on two things at once. It has to feel intuitive for beginners, and it has to avoid planting the wrong mental model in exactly the places experts know learners get confused. Readers appreciated that the author responded fast, shipped fixes during the conversation, and was explicit that the project was partly a way to confront their own understanding with reality. That openness landed well and made the inaccuracies feel fixable rather than fatal.
If you build educational sims, expect users to stress-test both the UI and the underlying model. The fastest way to earn trust is to label approximations clearly, separate conceptual visuals from actual simulation code, and patch obvious inaccuracies quickly.
Positive and constructive. People liked the simulator’s clarity, speed, and educational intent, but many immediately started auditing the physics and visuals for places where a beginner could walk away with the wrong model.
Key insights
01
Einstein slide needs a cleaner boundary
The simulator’s relativity section works as a visual metaphor, but it muddies the relationship between Newtonian gravity and general relativity. Newtonian gravity is the weak-field, low-speed limit of Einstein’s theory, not a separate rival model, and the current handoff makes that continuity harder to see. That matters more because the author confirmed the N-body code is still purely Newtonian, with no post-Newtonian correction such as Mercury’s perihelion precession.
If you present multiple physics models in one product, mark exactly where the math changes and where it does not. A simple label like "conceptual visualization" versus "simulated dynamics" would prevent most of this confusion.
The day-and-seasons section appears to animate Earth’s axis in a way that looks like precession, even though the intended concept is axial parallelism. Over the span of a year, the axis should stay pointed in nearly the same direction in space, which is why the tilt toward or away from the Sun changes with Earth’s position in orbit. Precession is real, but it happens over roughly 26,000 years, so showing it in a one-day or one-year context plants the wrong idea.
When teaching astronomy, animation choices can be more misleading than text. Review any motion you add for visual appeal and ask whether it implies a physical timescale you did not mean to show.
The explanation leans too hard on the neat beginner story that there is one magic speed for orbit, slower means crash and faster means escape. That is only true at a given radius for a circular orbit. In practice, speeding up often gives you a larger ellipse and slowing down gives you a smaller one. Commenters pointed out that falling into the Sun from Earth’s orbit is especially hard because you first have to cancel a huge sideways velocity of about 30 kilometers per second.
If you teach orbital mechanics, show circular, elliptical, and escape cases side by side. That gives learners a model they can actually use for spacecraft intuition instead of a binary crash-or-fly-away picture.
For readers who want to go beyond an educational demo, REBOUND came up as a production-grade N-body package used in planetary formation research. The point was not that the Show HN should use it, but that there is a well-known reference implementation for people who want stronger numerics, standard integrators, and examples that connect toy simulations to scientific practice.
If this project grows past education into analysis, compare its outputs against REBOUND early. That gives you a sanity check on integrator behavior before you add more features.
A useful reminder from orbital mechanics practitioners was that the clean Kepler picture is only the starting point. Real trajectories need perturbation calculations because every body influences every other body, and small effects accumulate over huge distances and long time spans. That is why hand-derived ephemerides were so demanding and why even the solar system is not perfectly, forever stable.
Treat long-run orbital accuracy as a different problem from short-run visualization. Once you move from teaching intuition to forecasting trajectories, perturbations and numerical error control become first-class features.
Most attention went to mistakes, but one reader called out that the solar-system-through-space view avoids the viral "vortex" animation that wrongly turns planetary motion into a trailing spiral. Showing the planets as a helix relative to the Sun’s motion is a rare place where the project is more correct than popular science content people may already have seen.
Keep this section and consider explaining explicitly why the older vortex animation is wrong. It is a chance to replace a common internet misconception with a better mental picture.
qunabu.github.io
And Yet It Moves, Quite A Lot Like That