In the academic applied probability/math finance community, Backwards Stochastic Differential Equations (BSDE's) are extremely popular, and they provide a single framework for several different problems, notably hedging and utility maximization, where models of market imperfections might stop older models. The basic setup in a high level sense, is that you specify a terminal condition (i.e. what you might like to hedge at the end of a trading period), and the dynamics in time backwards from that point. For those interested, they are not equivalent to forward SDE's precisely because there is a filtration.
I'd like to know if people are using these devices in practice, and for what purposes, and basically, what is the state of the art?