Skip to main content

September

Participated in the 2024 Cylon GameJam for all of September up to the 24th.

Tuesday 24th​

  • Expanding out jitter dynamics scene in Ephemeris project
  • Debugging torque in jitter
  • Compared torques:
    • Standard Unity physics
    • Unity.Physics package
    • Jitter
    • A quick prototype using Myriad components
  • Jitter just seems to have the wrong inertia tensor?

Wednesday 25th​

  • Properly integrating rotations in Myriad
  • Experimented with Z3 for solving thruster firing
    • Z3 probably isn't feasible to use in-game, just PoC
  • Cleaned up various physics related things in Ephemeris into a single Dynamics namespace
  • Removed Jitter physics

Thursday 26th​

  • Moving ApplyImpulse method to Myriad.Dynamics
  • Added new GetComponentRef overloads to Myriad which retrieve many components all at once
  • Fixed Myriad.ECS GitHub build action failure
  • Updated Ephemeris to use new Myriad.Dynamics components, instead of ad-hoc mass and orientation components that existed before
  • Fixed up docking sim to compile and load properly with new stuff
  • Experimented with a basic "aspect" systems for physics
  • Replaced default camera blend with "cut" (nothing else really makes sense at this scale)
  • Tested Z3 in Unity. It required a native DLL but it does work!

Friday 27th​

  • Experimented more with Zen in Unity.
    • Issues with creating Real: casting -4 acts different to new Real(-4)
    • Timeout failed to prevent a hang when using Minimize!
  • Spawning player pod and station near each other in docking sim
  • Debugging issues with orbital paths
  • Modified NBody integrator to work for entities with no mass (just assume mass=1)
  • Setup a test harness for calculating thruster activations
  • Fixed various physics issues with torque
  • Found article about thruster torques with iterative solving (2D): https://blog.cosmoteer.net/2011/01/starship-builder-idea-thruster-algorithm.html
  • Experimented with determining thruster throttle values using matrix inversion (for sets of three thrusters acting as a "virtual thruster gimbal")

Monday 30th​

  • Experimenting more with matrix inversion for torques (i.e. including thruster offset as well as direction).
  • Built solvers for linear force and angular torque, they seem to work well independently
  • Fixed up initial scenario so it always spawns a sensible distance away by inspecting orbital velocity and delaying by a suitable amount of time