@El'endiaStarman Hi! I'm sorry, I disappeared for a while

I've made a double-pendulum simulation before, but that's all

Are you still working on the pendulum? I might have some ideas

@EricTressler As a matter of fact, I had a crucial realization this morning. I had been treating $\theta$ and $\phi$ as completely independent, but I realized this morning that a pendulum moving through 3D space with no external forces will trace out a great circle, and that has $\dot{\theta}$ and $\dot{\phi}$ varying in time, which my simulation was not doing.

It's possible the only problem my code had was treating the combination of $\dot{\theta}$ and $\dot{\phi}$ correctly for determining initial velocity.

i've been working on a fluid simulation, but i'm beginning to think this is a gigantic waste of time

Why so?

i'm just implementing a lattice boltzmann algorithm that i found elsewhere, it's too slow to be interesting to watch in real-time unless the resolution is tiny, and i don't understand what's going on well enough to extend it from 2d to 3d properly

Ah, that's a bummer.

i've spent a lot of time on this for very little gain, versus things like my graph layout code

where at least i know what i'm doing and i get reasonable returns on my time invested

i think even my watercolor trees were a better idea, since there nothing substantial was going on at all and it was just aesthetic. anyway.... not sure where that leaves me now for projects

are you going to keep working on the pendulum code? When I did it, I used a 4th order Runge-Kutta approximation to the PDE. it worked really well

Definitely going to keep working on the pendulum.

Probably won't go so far as Runge-Kutta though.

Simply doing vel += accel then pos += vel works really well for how little effort it is.

i guess it's probably not a noticeable difference unless you compare directly; it might be easier to move to things like the spinning chains if you set it up as a PDE, though