Sir

Negotiating Stephen Wolfram's vast book A New Kind of Science, on which you report in the News Feature “What kind of science is this?” (Nature 417, 216–218; 2002), reminded me of Huckleberry Finn's epic journey drifting down the Mississippi on a raft — not just the length, but the scenery.

On the left bank of the river we pass a long series of wondrously complex and often very beautiful pictures, the amazing products of his simple automata. On the right bank we drift by his conjectures and speculations about the meaning of these gorgeous images. He 'explains' nearly all the real world's puzzling phenomena, from particle physics to evolution of species.

How might we cross safely from the left bank of the river to the right? I found myself yearning for a solid bridge to link at least one of Wolfram's pictures to the real world, but unfortunately, all he offered was a leap of faith.

Which leads me to wonder: what did we really see on this long and dazzling journey? If it was 'science', even of a new kind, it should fulfil at least two criteria. First, it should be falsifiable: it should be possible to design an experiment to disprove a proposed link between the model and the real world. And second, it should generate new hypotheses about the real world which scientists can test in their laboratories. Perhaps Wolfram could enlighten us: how might his computer models fulfil either of these criteria in order to qualify as science?