Moore's Law, Moron's Law, and the blogsphere

The reason why the ABC article attempting to apply Moore's Law to the blogsphere bothered me is that it fundamentally misinterprets Moore's Law to fit some preconceived explanation as to why the blogsphere is expanding. This is the same reason why I get irritated when I see the "Long Tail" meme attached to so many powerpoint slides that it doesn't belong on.

I'm not a hardware engineer, but I have read a lot of material on Moore's Law in an attempt to understand it better and it's clear that Moore was as much concerned with the cost of manufacturing integrated electronics as he was the performance potential.

What is often simply referred to as adding more transistors to a chip as a function of processing potential is actually a pretty rational opening statement in the argument for increasing yields in semiconductor manufacturing, given that the more transistors you can fit on a square inch of silicon the more will fit on a wafer by decreasing the size the more dies you will fit on the wafer itself. This is important because assuming the defect rate remains constant, increasing the density of components ensures that the yields increases disproportionately thereby realizing the observation that:

Reduced cost is one of the big attractions of integrated electronics, and the cost advantage continues to increase as the technology evolves toward the production of larger and larger circuit functions on a single semiconductor substrate.

There is one other really important part of Moore's Law that conveniently gets left out of the media accounts and cocktail party conversation, which is that there is a point of imbalance whereby low yield costs are offset by high packaging costs. In other words, as more components are added to each die the packaging costs decrease, but at some point the the defect density mentioned above kicks in a yield costs increase. What Moore was really getting to here was the notion that the evolution of chips was entirely dependent on the industrys ability to build new state of the art fabs every year that increase wafer size, decrease component size, and decrease defect density.

There's obviously a lot more to Moore's Law than simply manufacturing costs, which is why the original article by Moore (and Englebart's work before him) is timeless. What Moore was able to do was package together a collection of observations and very unique insights to explain how a set of converging trends would change the world as we knew it. Moore was right, but I am quite sure that even he could not have predicted how many explanations the media would come up in their attempt to cover what he actually said, and more importantly what he actually meant. So the next time you hear "Moore's Law" (or "network effect" or "long tail") attached to some trend, ask the critical question "why?".