This idea that every connection, so to speak, in the nervous system would be encoded by a specific molecular interaction gave rise to a lot of problems. Jacobsen wrote a paper in which he said it was impossible for the simple reason that there weren't enough nucleotide base pairs in the DNA to code. That is, he calculated the number of synapses in the nervous system and showed they were vastly in excess of the number of base pairs in the DNA let alone the genes. But that was naive because it strongly underestimated the power of combinatorics. That is that if you take a number of elementary processes and repeat them in various ways you can get enormous computation power of it. And in taking one of these I was able to show, just as an exercise to suit myself, that I could specify a retina... with about 10 million specific chemicals or combination of chemicals, by using just 112 genes. I didn't believe in it, because it assumed the prime numbers existed in nature, that's one of the reasons I didn't believe in it. But it was an interesting exercise, because another reason is I calculated... because it involved one cell telling another cell to turn on a gene, and then it would tell its neighbour... the point is you could calculate how long that would take to do, and would take too long. So it was I think implausible. And it is that that I think is so important in trying to make theories, is that they should always contain some biochemical plausibility — some biological plausibility — estimate. That is, you can do things in a very sophisticated mathematical way but if in fact it takes 70 years to calculate in a biological system it's of no relevance. It's the difference between theories being correct and theories being true... which is... many theoreticians don't make that diff... difference, but in fact many theories are correct in a logical sense but they're untrue because they don't refer to the natural thing we're all interested in. Well, if that was so then you could argue that you would knock-out... that you would be able to knock out a given synapse by a single gene, and again many people felt that that was implausible, and it's very clear that what you had to have was a construction paradigm that could build all of these connections.