As I discussed in yesterday’s blog posting, Moore’s Law tells us that we should expect to see computer technology becoming ten times more powerful in the next five years, and 100 times more powerful a decade from now — in such key dimensions as speed, cost, size, and storage capacity.

Let’s start with the dimension of speed (I’ll discuss the other dimensions in separate blog postings in the next few days). I just peeked at the details of the desktop Mac that I’m using to write this blog; it says I’ve got a 2.8 GHz Intel Core Duo processor. If Moore’s Law means that in the year 2015 I’ll be using a 28 GHz Intel “Core 20″ processor, does that mean Microsoft Word will run 10 times faster, or that Excel will “crunch” the numbers in my expense-account spreadsheets 10 times faster? And if it did, would I care? Would I even notice?

There is considerable skepticism on this point, and it’s something that most of us have observed since the introduction of the PC and the growth of the business/consumer software industry. We often refer to it as “bloatware” — i.e., the tendency of software product companies to add enough new features and GUI “bells and whistles” to soak up all of the additional CPU power in the latest laptop and desktop computers. Indeed, just as there is Moore’s Law to provide a long-term, optimistic view of the future of hardware technology, there is also Wirth’s Law, formulated by software guru Niklaus Wirth in 1995, which says, “Software is getting slower more rapidly than hardware becomes faster.” And in 2009, Google’s Larry Page suggested a variation called Page’s Law, the tendency of software to get twice as slow every 18 months.”

While there are undoubtedly some exceptions (none of which I can think of right now), I think it’s fair to say that very few consumer applications are CPU-bound. No, I don’t need Microsoft Word to run 10 times faster; and no, I don’t need to have Google retrieve a million “hits” in 13 microseconds for one of my search queries.

On the other hand, business applications are CPU-bound … at least in some cases. That’s probably an over-simplification, since a typical business application uses a combination of CPU number-crunching, as well as RAM memory, disk storage, transmission of data through I/O channels, transmission of data across telecommunications networks, and goodness knows what else. But to the extent that CPU power is the critical resource, we can appreciate how important it might be for tomorrow’s faster processors to help businesses keep up with the massive volume of transactions that must be processed in “batch mode” during each 24-hour day, as well as helping businesses process “real-time” transactions with an acceptable response time.

Even though this may be important, I can’t help feeling that it represents the kind of “incremental” improvement that won’t represent a dramatic change. On the other hand, I can imagine the possibility of tomorrow’s entrepreneurs looking for new business models to take advantage of the tenfold-increase in computing power. We’ve already seen one example of this in the music business: in the “old days,” I would usually hesitate before deciding to spend $14.99 to purchase an entire album of music (regardless of whether the album was “burned” onto vinyl, or a tape cassette, or a CD). But with digital music and iTunes, I hardly hesitate at all to spend $0.99 to download an individual song; indeed, the “resistance” to such a transaction is so minor that it wouldn’t surprise me if I carried out 20 such transactions, and spent $19.80 to buy 20 individual songs.

What if a future business organization came along and offered me the chance to download and play a song once for $0.10 … or maybe even $0.01? If I became obsessed with the song for a day or two (as indeed I do, sometimes), I might be tempted to click a button, time after time after time, on my future iPod, and play the song over and over and over again. If you scale this up to a marketplace of, say, 100 million music-loving teenagers, then you’d better have a lot of CPU power to handle that kind of transaction load.

In addition to future scenarios where a moderate amount of computing power has to be applied to a qualitatively greater number of transactions, we’ve got a list of more “traditional” compute-intensive problems that would be greatly assisted by 10-fold and 100-fold improvements in CPU power. These include such things as weather forecasting — wouldn’t it be nice if weather forecasters really could tell us whether it’s going to rain next weekend? More important, wouldn’t it be nice if weather forecasters could provide an accurate, credible warning about the path of an approaching hurricane, or the arrival of a deadly tornado, a day or two in advance?

In addition to weather forecasting, there are also compute-intensive problems in biomedical research, computational genetics, artificial intelligence, as well as grim areas like nuclear weapons research. As an InfoWebLinks website observes, one good place to start looking for compute-intensive problems is to look at the various supercomputer centers around the country — and they’ve got links to six such centers, so that’s a good place to start.

Not too long ago, the modest computer on my desk would have been considered a supercomputer. So, one of the important “dimensions” of future advances is that computing problems that used to require special, multi-million dollar supercomputers — of which there were perhaps only a dozen in the world — can now be carried out on a modest budget, on anyone’s home computer.

But those supercomputer centers are still out there, and they’ve still got multi-million dollar machines that none of us could afford on our own. And there are still problems out there that we simply can’t solve today, because we don’t have enough CPU cycles. Sometimes we can combine the computing power of thousands of separate machines, as is being done with the SETI Project; but there’s no doubt in my mind that we’ll find useful things to do with all of the CPU/speed improvements promised by Moore’s Law for at least the next decade, and probably far beyond that.