It is common knowledge that hair simulation and rendering is particularly challenging and in fact rare are the examples in games that propose believable hairs. It is pretty easy to imagine how hard it is to actually simulate the high amount of fibres and to simulate the various scattering events that may happen between so many strands, let alone the fact that each of them is translucent.
Can someone formalize what are the main challenges in hair rendering, especially for real time scenarios? How these are overcome by state-of-art techniques? Are there some de-facto standard theoretical models used?
To clarify, for real time I don't necessarily mean in a game context.