Three posts in one day. And all of such high quality. You lucky people. Pascal has a go at explaining the std.nutters. Some of it is the usual correct stuff, but some of it is wrong: What you need, over and above all that, is constant social interaction with other practising scientists. Oral tradition and daily exposure to other scientists' everyday decisions are indispensable. That sounds fairly plausible, doesn't it? Until you think of Newton. Or indeed, of Einstein.