v1.9 Recombination Epoch

After several months of a rather dormant state, I have new solver methods, quality-of-life improvements and demo scenarios. Windows binaries included as usual.

In my opinion, the most prominent new feature should be the adaptive solver methods, automatically adjusting time step size for a given approximate error accumulation tolerance, rather than providing a fixed time step size. It does not work the same way, for example, RK89 does; which uses solvers of different orders. The current adaptive time-step solvers instead use half-length steps to compare the error accumulation against a full-length step. If a bad step is taken, it will be repeated with a smaller step size. If there is room to enlarge the step size, a larger step size will be attempted in order to cut down on overall simulation execution time.

Another very important improvement is the implementation of an automated orbit propagator validation test. It constructs an n-body model of the Solar System (with Sol and all 8 major planet barycenters) using JPL Horizons API and compares the propagation result against a later state. This way, manual labor is eliminated, and different solver setups can be validated quickly (without doing a journal paper's worth of work every time). Now, maybe you can write that thesis of yours with the time you have saved.

Finally, you will now also find an Earth-Mars transfer scenario in the /scenarios folder, creatively titled mars_transfer.osf.

Full Changelog: v1.8...v1.9