In this talk, I will briefly present a range of extensions to the fault-tree notation. These have been designed to enhance Fault trees as a means of representing the events leading to major accidents. The second half of the talk will be a brainstorming session. I would like to gather the group's ideas about techniques for assessing the `usability' and `utility' of Fault trees. In particular, I am concerned to identify any ways in which work from the psychology of programming might be used to assess this graphical notation.