It seems like we're always linking to something that Leila (author of How To Worry Friends and Inconvenience People) has done. This time, though, I want to link to a story her beau, Tom, has written for the Guardian. It's here. It's a piece about a young computer-science whiz. Now I may be doing you all a disservice, but I always tend to imagine that the sort of people who'd read a publishing blog spent more time in the English Department than the Science Block at school. But even if that's true, maybe you wondered what they got up to over there. One of the nice things about the story Tom is picking up on is that it's also about Stephen Wolfram - and Stephen Wolfram thinks that a good deal of the science that came out of the twentieth-century was headed into a cul-de-sac, and that it would never be able to explain most of the things we see in nature. If nothing else, that certainly justifies any time you spent reading novels when you could have been playing with bunsen burners.

Wolfram, because he's so modest, suggests an alternative starting point for science. His idea is that lots of simple things out there in the world can act like very primitive computers. Remember all that stuff from the Eighties about Chaos Theory? People thought that simple things behaved in simple ways, and then it turned out that even working out when the next drip was going to fall from a dripping tap could be a complex mathematical undertaking. Well, Wolfram realised that lots of things in nature could be so finely balanced that they were very sensitive to their environment, reacting differently depending on what was going on around them. And if you thought of the environment as 'input' and you thought of what happened next as 'the output', then what you'd got in a lot of natural systems were very simple calculating machines. And if nature was full of ultra-simple calculating machines, it made sense to do a bit more research on what super-simple computers were capable of. At which point Wolfram realised hardly anyone had ever bothered with that research because they hadn't expected to find anything interesting.

So Wolfram invented lots of imaginary and very basic computers - computers so simple you wouldn't even call them computers - because all they did was follow incredibly simple rules. And he found all sorts of extraordinary things. If you take the output from one of these primitive computers and feed it back in, then you could generate astoundingly complex behaviour from a set of rules that weren't much more complicated than the rules for noughts and crosses.

Why is any of that interesting? Well, if nature is full of tiny calculating machines - and twentieth-century science has been ignoring the study of ultra-simple computers in order to do lots of complicated differential calculus, then to a large extent science has been missing the point of... well, of science. And the proof mentioned in Tom's article - if it holds up - demonstrates that it's possible to create a fully functional computer, one that could in principle run Excel and Tomb Raider, from a design that's much simpler than any computer anyone else has ever conceived of. If something that acts like a computer can be made from simpler components than anyone thought possible, it might be a step towards proving that nature really is full of tiny, super-simple computers. Which means that your grandkids might be ignoring a completely different kind of science as they bypass the Science Block to head to the English Department.