As I suggested in my previous blog on this series of commentaries on future trends in the IT industry, we could probably spend several years improving things even if computer hardware didn’t change at all. But the fact of the matter is that hardware has changed, and will probably continue to do so for the next several years.

Indeed, this has been true since the introduction of the first behemoths at the end of World War II, and certainly all through the 1950s. But much of our thinking about this phenomenon can be traced back to a paper that appeared in Electronics magazine, on April 19, 1965 — which was titled, “Cramming More Components Onto Integrated Circuits,” written by a 36 year old Intel engineer named Gordon Moore. The basic thesis of the article was that a trend had been underway since the invention of the integrated circuit (IC) in 1958, and would probably continue for quite some time, in which the number of transistors that could be placed on an IC would double approximately every … well, 18 months or 20 months or 24 months, depending on whose recollection you listen to.

In the paper, Moore predicted that the “doubling” trend might well continue for another ten years; but it has actually continued to the current day, and a vintage-2005 paper (“New Life for Moore’s Law“)suggests that it may continue for another ten years, i.e., until the year 2015. In 2008, Intel predicted that the trend would actually continue through 2029 (see “Moore’s Law: ‘We See No End in Sight,’ says Intel’s Pat Gelsinger“). And futurists like Ray Kurzweil have suggested that even if improvements to transistor-based technology does hit physical limits, technology will continue advancing with new approaches such as optical or quantum computers.

Whether improvements do indeed continue that far into the future is beyond my ability to predict; at least for now, I’m content with the idea of Moore’s Law continuing for another five or ten years. If indeed the technology doubles every 1.5 years, then we get three consecutive doublings — i.e., a factor of 8 improvement — in 4.5 years, or (in very rough terms) a 10-fold improvement in 5 years. By the same logic, we’ll have a 100-fold improvement in 10 years.

There’s an aspect of this that I’ll discuss in more detail in future postings, but I’ll summarize it here: a 10% improvement is “incremental” and doesn’t motivate us to make fundamental changes in the way we do things (unless it continues year after year, and leads to a “compounded interest” effect). But an order-of-magnitude improvement is a qualitative change.

For example, if I told you that the new car you’re going to buy next year will cruise at a 10% faster speed than this year’s, or that its average gas mileage will be 10% better, chances are you would just yawn. But if I told you that instead of cruising at 60 mph, you could cruise at 600 mph, you’d agree that it’s a whole new ballgame (the fact that our existing highway infrastructure couldn’t handle cars traveling at 600 mph is a limitation that we’ll discuss later).

It’s also important to remember that we see these improvement in several different “dimensions”: CPU speed, memory capacity, size (footprint), network capacity, and cost. Some of these dimensions are combined — i.e., today’s computers are dramatically faster and smaller and cheaper, and with dramatically larger amounts of storage capacity. But each dimension on its own is important, and needs to be considered on its own.

We’ll begin doing that in the next installment of this blog series — by looking at the consequences of computing technology that’s an order of magnitude faster in five years, and 100 times faster a decade from now.