In this presentation I will summarise some of the work I have done towards
increasing the robustness and efficiency of deep parsing and then talk about
some of the questions that have arisen from this research, focusing particularly
on meaningful parser evaluation. I will outline some ideas I've had for
evaluating parses based on information content, both syntactic and semantic, and
also how this might relate to application-based evaluation. Feedback and
suggestions in this area would be very welcome.