Numbers like 100 or 1000 are arbitrary in one sense — the universe does not care whether a collection of simulations ends in a round digit. But milestones matter for a different reason: they create moments of reflection. Looking back across one thousand interactive models, from a simple pendulum and an SPH fluid in 2024 to cosmic ray showers, Boolean networks and 3D cellular automata today, the project has evolved in ways that were impossible to predict at the start. This devlog is as much a celebration as a technical update, and we make no apology for the enthusiastic tone.
mysimulator.uk now hosts 1,000 interactive science simulations, all running natively in the browser with no plugins, no downloads and no sign-in required. Every simulation is available in English, Ukrainian and Polish.
Wave 109 Recap: Twenty Physics Simulations
Devlog #101 covered Wave 109 in detail, but it deserves a summary here because it set the stage for the milestone. Twenty simulations were released, each chosen to make a specific equation visible at interactive speed rather than simply animate a phenomenon. The principle behind the project has always been: a simulation should answer the question “what does this formula actually look like in motion?” Wave 109 applied that principle to some of the hardest questions in contemporary physics.
- Debye LengthPlasma shielding heatmap, λD = √(ε0kBT/nq²)
- Topological Defects2D XY model, BKT vortex unbinding
- Stochastic ResonanceLangevin bistable, SNR bell curve
- Granular JammingBidisperse MD, φJ ≈ 0.84
- Bragg DiffractionFCC/BCC powder pattern, nλ = 2d sinθ
- Spring NetworkMaxwell criterion zc ≈ 4
- Quantum Dot EmissionBrus equation, CdSe colour palette
- Spin LatticeHeisenberg Metropolis MC, Tc
- Airy DiskBessel J1, Rayleigh criterion, ImageData
- Reaction FrontFisher-KPP PDE, v = 2√(Dr)
- Entropy MixingΔSmix particle simulation
- Soap BubbleGauss-Seidel Laplace relaxation
- Cellular Automaton 3DThree.js InstancedMesh, 20³ grid
- Boolean NetworkKauffman NK, K = 2 critical phase
- Population Wave2D Fisher-KPP, Allee effect
- Network PercolationErdős-Rényi, Union-Find
- Random GrowthEden + DLA fractal, D ≈ 1.71
- Hall ThrusterE×B drift, Isp, thrust
- Cosmic Ray ShowerHeitler cascade, NKG profile
- Kirchhoff's LawsMNA matrix, Gaussian elimination
That wave brought the total to 940 simulations. Sixty remained before the milestone. Wave 110 was designed to cross the line in a single release, and we chose breadth over depth: sixty simulations across eight thematic batches, ensuring that no major domain of science or engineering was left without new material.
Wave 110: The Sixty That Made 1000
Planning Wave 110 was an exercise in prioritisation. The simulation catalogue already covered most of classical physics and much of modern physics, but several fast-moving fields — signal processing, computational neuroscience, generative mathematics, and contemporary electronics — were underrepresented relative to their importance. Wave 110 corrected that imbalance with eight thematic batches.
Fluid Dynamics
Fluid simulations are among the most computationally demanding and visually rewarding
on the platform. Wave 110's fluid batch added models for compressible shock waves,
Kelvin-Helmholtz instability at a shear layer, Rayleigh-Taylor driven mixing,
rotating convection in a spherical shell (a desktop model of planetary interiors),
and a Hele-Shaw viscous fingering simulation that reproduces the branching patterns
seen when oil displaces water in a narrow cell. Each is implemented on a 2D grid using
finite differences or spectral methods, with GPU-accelerated rendering via the
ImageData API wherever frame rate is the limiting constraint.
Life Science
Biology has always been underserved by browser simulation relative to its scientific importance. Wave 110's life-science batch addressed this directly. New additions include a mitosis visualiser with phase-by-phase chromosome separation, an action potential propagation model along a myelinated axon (including the Hodgkin-Huxley equations with Ranvier node saltatory conduction), a gene regulatory network showing toggle-switch bistability and oscillator circuits, and a predator-prey metapopulation model on a spatial grid that generates the classic Lotka-Volterra cycles enriched by dispersal dynamics. These simulations are intended to serve secondary and undergraduate biology education as directly as the physics models serve physics courses.
Generative Art & Mathematics
The generative art category has been one of the most popular on the platform since its introduction, and Wave 110 expanded it substantially. Highlights include a space-filling curve explorer covering Hilbert, Peano, Gosper and Moore curves at user-selectable iteration depth; a parametric surface gallery (torus, Klein bottle, Boy surface, Steiner surface) rendered with WebGL Phong shading and normal mapping; a Voronoi diagram builder with Lloyd relaxation; and a tiling explorer covering all seventeen wallpaper symmetry groups. The mathematical depth of this batch is genuine: the wallpaper group simulation labels each tiling with its orbifold notation and shows the fundamental domain explicitly.
Electronics & Signal Processing
Following the Kirchhoff MNA solver introduced in Wave 109, Wave 110 added a family of electronics simulations covering digital logic gates with timing diagrams, a transistor characteristic curve tracer (BJT IC–VCE families), a phase-locked loop with lock-range visualisation, and an ADC/DAC pipeline demonstrating quantisation noise and aliasing simultaneously. The signal processing batch added simulations for matched filtering, spread-spectrum communications (DSSS and FHSS), wavelet decomposition on user-supplied images, and a spectrogram analyser with adjustable window functions (Hann, Blackman, Kaiser).
Chemistry
Wave 110's chemistry additions focus on reaction mechanisms at scales ranging from molecular to reactor. New simulations include a molecular orbital viewer for diatomic molecules (H2, N2, O2), a polymer chain conformation explorer using Markov chain Monte Carlo, a surface adsorption isotherm simulation (Langmuir, Freundlich, BET), and a packed-bed reactor model showing concentration profiles along the reactor length as a function of residence time and Damköhler number. These complement the Belousov-Zhabotinsky and Fisher-KPP reaction-diffusion simulations already in the catalogue.
Space Science
Building on the Hall thruster and cosmic ray shower models from Wave 109, the space science additions in Wave 110 include a magnetospheric dynamics visualiser (magnetic field topology, magnetopause, radiation belts), a cometary tail simulation driven by solar wind pressure and radiation, an atmospheric re-entry heat shield simulation using the Fay-Riddell stagnation-point heating formula, and a neutron star equation of state explorer comparing polytropic, realistic nuclear-matter, and quark-matter models at densities up to 1018 kg m−3. The neutron star simulation is one of the most technically demanding on the platform: it integrates the Tolman-Oppenheimer-Volkoff equation in real time across a slider sweep of central pressure.
Signals & Control
Modern engineering is inseparable from feedback control and signal analysis. Wave 110 introduced a root locus plotter that updates in real time as loop gain changes, a Bode plot generator for user-defined transfer functions, a Z-transform visualiser for digital filters, and a model predictive control (MPC) sandbox where users set horizon length and cost weights for a double-integrator plant. These are tools that undergraduate students in control and signal processing courses can use directly alongside their coursework.
Neuroscience
The neuroscience batch is arguably the most ambitious in Wave 110. It includes a spiking neural network visualiser (leaky integrate-and-fire neurons with spike-timing-dependent plasticity), a Hopfield network associative memory demonstration, a visual cortex orientation selectivity model based on difference-of-Gaussians receptive fields, and a decision-making model based on the drift-diffusion process with adjustable signal-to-noise ratio and reaction-time distribution output. Together they give the platform a serious computational neuroscience wing for the first time.
Try It Yourself
With a thousand simulations to choose from, it helps to have a starting point. Here are six from the two waves that represent the breadth of what has just arrived:
The full catalogue is browsable via the category index, which
now lists over forty science and engineering categories. The search page
accepts natural-language queries and returns results ranked by conceptual relevance.
Every simulation is available in English (/), Ukrainian (/uk/)
and Polish (/pl/).
What a Thousand Simulations Actually Means
It is worth being specific about what the milestone represents and what it does not. A thousand simulations does not mean a thousand one-dimensional animations with a play button. The criterion for a simulation to appear on this platform has always been: does it respond meaningfully to user input, and does that response reflect the underlying physics or mathematics rather than pre-recorded data? Every model on the site is either analytically computed or numerically integrated at runtime; none are video clips or look-up-table animations.
The breadth is genuine. The catalogue spans plasma physics, condensed matter, quantum mechanics, optics, acoustics, fluid dynamics, thermodynamics, chemistry, reaction-diffusion systems, mathematical models of ecology and epidemiology, algorithms, cryptography, machine learning, electronics, control theory, signal processing, computational neuroscience, space science, generative art and mathematical visualisation. That range is deliberate: science does not divide itself into silos, and neither should a simulation platform designed to support curiosity-driven learning.
If you are a teacher or lecturer looking for specific simulations to use in a class, the contact page is the best route to getting topic-specific recommendations. Several of the simulations in Waves 109 and 110 were designed specifically in response to requests from educators who wanted interactive tools for concepts that are hard to convey on a whiteboard.
What Comes Next
Reaching 1000 does not mean stopping at 1000. The backlog of simulation ideas is longer now than it was at the start of the project, because each new simulation reveals adjacent concepts that deserve their own interactive treatment. Wave 111 is already in planning.
The near-term priorities are quality improvements to existing simulations: better mobile performance, more detailed explanatory text, additional parameter controls, and deeper integration with the Ukrainian and Polish language versions. Several simulations in the early waves were built before the current technical standards were established and will be rebuilt rather than simply translated.
Longer term, the platform is exploring richer interoperability: the ability to link two simulations together so that the output of one becomes the input of another. A Fourier transform simulation feeding a convolution simulation; a Navier-Stokes solver feeding a passive scalar advection model; a gene regulatory network feeding a population genetics model. The underlying architecture supports this, and the first composite simulations may appear in Wave 112 or 113.
Thank you to everyone who has used the platform, suggested simulations, reported bugs, translated content or simply shared a link. A milestone like this is not a solo achievement. The simulations exist because they are needed and used.