Oh, goody! Now, we are putting in "particle physicist" as a requirement for being a game developer.
why sound soo strange?
“game developer” is something extremely generic and require all the possible skill you can immagine.
you need people good at music for soundtrack
you need good artist for artwork
game designer need to be good at math and spreadsheet.
you need to have no selfesteen and selfpreservation for be a community manager.
Physics/software ignoramus spotted. There is a huge tradeoff between accuracy and computational expense for any physics model. This is both in terms of the numerical approximation to the model, as well as to the model's accuracy and robustness at various scales of analysis. Think about the flow of a water through a pipe. There are many models you can use to describe this phenomenon:
- A low-resolution model which only considers the change in flow behavior along one direction, the axis of the pipe, and models the pressure drop as a function of averaged friction.
- A macroscopic model of the fluid flow which assumes the flow to be a continuous media in 3D, but which makes simplifying assumptions (incompressible fluid, no slip boundary condition at the pipe walls, laminar flow, etc)
- A more complete macroscopic model which includes more complex effects (compressibility, interaction at the solid-fluid interface, turbulence, etc)
- A nanoscale model which treats the fluid as H20 particles with classical potentials
- A quantum-level model which solves the Schrödinger equation to obtain more accurate molecular potentials
If it's not obvious, moving from one model resolution to the next incurs
massive computational expense. It is virtually impossible to model a macroscopic phenomenon like this with the tools of molecular dynamics, unless you make clever simplifications, have access to a supercomputer, and are willing to wait weeks to obtain microsecond-length simulations. And you're expecting this shit to make it into games programming?
The only ubiquitous physics simulation in games currently is ragdoll physics, which treats objects as rigid bodies with constraints (pin joints, friction, collisions, etc). Developers are still working out the kinks there, because obtaining robust, realistic models that aren't numerically unstable and don't melt your processor is a difficult task. The second you start asking for flexible body dynamics, or realistic water physics (not just water surface animations that look pretty), let alone real-time simulations of biological tissue failure, you're basically asking for a totally new paradigm either in physics modeling or in CPU/GPU architecture.
EDIT: what I totally forgot to reiterate here is that most physics simulation isn't worth the effort. Clever animations, hacky math, and dice rolls are almost entirely sufficient to capture most of what you'd want out of elaborate simulations. This is as a person who cares deeply about physics simulation as a field and who loves to see it in games -- developers should get away with as many shortcuts as possible.
