I think we should continue to work on lots of things. There are a number of issues on which I have opinions. I think that we should restore some long-term senior positions. I think we should go after people, very interesting, accomplished senior people or middle-aged people as well as young people, for sabbatical appointments. I think that we should have some work that's data-rich, that involves putting in the results of a large number of historical accidents, rather than dealing with abstractions that might be associated with any planet in the universe, or no planet at all. I think—but of course, there's a trade off there: the more specific you make the model, the more data rich, the more historical accidents you include; while the verisimilitude of the model may increase, the transparency goes down. So what are the very simple ones good for? Do you just make simple models and then compare them with the—simple computer models, for example, then compare them with the facts? Well that's not such a great idea usually because even if they agreed it would be peculiar; since the models are so far from reality, why should they agree with the facts? However, you can sometimes find properties, like scaling properties, but there're probably many others as well, that can be traced continuously from real systems all the way down to very skeletal systems with just a few simple principles. When that's true one might be able to learn from the simple systems, the simple models rather, something about how these mathematical properties occur in the… in the much more complicated cases. And I think that's very helpful and has been done in a number of situations.