On Sun, 9 Jan 2005, Miranda van de Heijning wrote:
Perhaps I don't understand Moore's Law properly, but aren't we actually well behind on its schedule? If we published our 10,000th book at the end of 2003, doubling in 1.5 years means 20,000 books in mid-2005. At the current rate we are not likely to reach that figure--anywhere between 16k and 17k seems more realistic.
When you've been watching a growth curve for a long time, the changes don't seem as drastic as when you watch for a short time. Any growth rate prediction has to have a starting point, we have always used 1990, though we could restart with other years, and, yes, given that there are ups and downs in any real growth curve, you could always pick either the highs or lows [as they do in government statistics] and then skew the results massively as they change exponentially. Since we grew so rapidly in some periods, sometimes in excess of TWICE the Moore's Law predicitons, you could always start with those years as a baseline to demonstrate that we are no longer growing at TWICE the Moore's Law rate. However, 1990 is the first year we actually made totally consistent additions every month to the collection, so it's the best start point. When you map things out over very long periods, all the bumps in the road seem to flatten out. . .I can send you a graph from 1990 to 2005 if you like, so you can see that unless you use a much larger graph than I can include in this format, it looks very smooth, even though it switches from a resolution of years to months at the top. mh