I'm a bit late with this post, as finishing Graviton and my paper had happened a few weeks ago already... It doesn't matter though, here are some of my results.
New features include:
- Fourth order Runge-Kutta integrator
- 3D mathematics instead of 2D
- Minimal mode with no graphics
- Orbit display
- More bodies and accurate data from HORIZONS in the default solar_system file
- A user module and a primitive event system, where the user may override various event handlers for information extraction and custom reports
- Enhanced reports (ETA calculation, important dates (beginning, end), timers, extensible via the user module)
- Improved control over the simulation (via CLI arguments)
- Tested saving methods
In addition to all these new features, I've ran a lot of test for an astronomy project of mine. One of the tests I've ran was simulating from the year 2006, 23 years, into the year 2029, in order to compare my results (the distance between Earth and the asteroid 99942 Apophis) with NASA's. In short, it seems like my work is pretty solid, as I've my results deviate from NASA's only by 0.505%.
The code is a complete mess, though, and has to be reworked. I blame Python.