Our research project aims at collecting common sense knowledge about scripts
and mining the collected data into processable script representations.

This talk will briefly outline the first step (knowledge acquisition) and give
a more detailed description of the second step. The talk is about ongoing
research; our approaches for mining script instances are still in the
development phase.  The goal is to generate a unique script representation
(comparable to an automaton) from equivalent script instances entered into a
web form.