Most of us who've been analyzing the computer industry for any length of time have been burnt by the "who needs that?" column. My first experience with the phenomena came way back in the '80s when I publicly decreed that the newly launched 386 really wasn't necessary. A few months later I got my hands on an advance copy of what would become Windows 3.0 and immediately began clamoring for Intel to release the 486.

All that eating crow has resulted in an virtual taboo against technological satisfaction, at least among us writers. I've learned not to dismiss faster, bigger, more powerful, because inevitably, more has never been enough, at least not for long. I'm becoming convinced however that we are nearing the time when the taboo must be broken.

Of course, there are always going to be some who simply can't get enough speed, storage, or resolution. However, those people don't drive this industry, despite what you may have heard. In truth, the "sweet spot", that "bang for the buck" prime spot that every technology buyer knows, resides just a bit down the slope from the newest, hottest, "revolutionary" levels. This is where most purchases are made--one or even two steps above the average machine already on the street, but nearly that far below the best currently on the market.

Conventional wisdom insists that early adopters, who are willing to spend freely for the latest-greatest, fund the development and eventual mainstreaming of new technology. As is often the case, conventional wisdom is wrong. It is the promise of a mass market that drives the high end. For example, Intel's P4 is selling at relatively high prices, but in very little volume with extremely high overhead, resulting in little, if any profit. However, the mere presence of the P4 drops the P3 into the sweet spot. Intel's fab lines are optimized for P3 production and much of the R&D costs for the chip have been written off. That is where the money comes from.

The P4 will get all the headlines for a while, until a new generation is promoted, then the masses, who've been slobbering for the much hyped P4 will suddenly discover that it is within their budget and they'll create a profitable mass market. In other words, in this industry, the cart does indeed push the horse. Let's assume that the P4 doesn't get good press (remember the Pentium Pro?). Would Intel be hurt by slower P4 adoption? Sure, but not nearly as much as it will suffer from the resulting loss of P3 sales. Buyers look at what they want, then buy what they can afford, usually attempting to reconcile the two as much as possible.

It might not be strictly logical, but if in our example, AMD offers a "better" P4-level processor, its mainstream sales of the Athlon would almost certainly surge--even if the Intel mainstream offering was better or cheaper. It's akin to stock car results--we know that the cars on the dealer showroom floor have little in common with those NASCAR beasts, but if Ford beats Chevy on TV, we just might believe Ford's consumer cars will best Chevy's as well.

OK, but so what, who really cares which Intel-architechure CPU wins? It doesn't really affect the industry, right? Well, it doesn't matter, as long as someone wins, but we're quickly coming to the point where the model begins to break down. Its been a long time, in computer terms, since a CPU became truly obsolete. Even an antique CPU such as the Pentium at 233MHz is still capable of running almost every application on the market. Only a few esoteric programs will refuse to run on a Pentium II. Most retail software, even many creative applications call for minimum configurations as low as a 133MHz Pentium--a CPU that had its heydey in the early 90s.

Its becoming obvious that even if you buy last year's Pentium III it will be a viable machine long after it has been depreciated. Sure, a 1Ghz system is faster than an 800Mhz version of the same chip, but the difference is increasingly becoming a factor only to benchmark technicians and a continually shrinking pool of specialists. Once, a new CPU, with its next level of speed, enabled important advances in what was practical. A 286 was fast enough to handle the keyboard input from the fastest human typist--as long as the text display was ASCII. However, a graphical interface, required for showing true on-the-fly WYSIWYG text editing overwhelmed even its successor, the 386.

The 486-era brought in the advent of multimedia, and after a slow start, video, animation, stereo sound, and 3D became possible. The Pentium improved on the quality of multimedia that could be used. However, at about the 200MHz Pentium level, a funny thing happened. Multimedia hit a plateau--for the most part, a 233MHz Pentium didn't "add" capabilities or increase quality.

3D quickly became the next hurdle. Games such as Quake did run better and with more detail on a faster PC. For the first time, CPU vendors began touting the "gaming market" as an important segment. Unfortunately for the CPU guys, graphics card makers quickly began supplying 3D accelerators that enabled a 266MHz Pentium II to run most 3D games at frame rates that were near the limits of human distinction. Sure, a 400Mhz PII could generate more frames per minute, but the eye could barely tell the difference.

Now with the budget, sub-$1000, market flooded with 500- and even 600Mhz systems, the market for faster and faster processors seems to be slowing down. Other than the much hyped 1Ghz barrier, few people outside of a few hardcore renderers even care about the speed of the current crop of new CPUs. People are becoming more concerned with the peripherals--having a large monitor, a photo-quality printer, enough RAM. I even have a friend that drives an hour to a computer superstore to buy older computers than he could buy locally, even though the price is virtually the same. His rationale is that his current crop of 400Mhz computers are plenty fast and if he brings in a new 500Mhz system, employees will fight over it, even though it delivers no real advantage running the word processing and Internet applications used exclusively in his office.

It isn't just the CPU that is outpacing demand. It wasn't long ago that it was a tired cliché that no matter how big your hard drive, you'd fill it up in a fraction of the time you'd expect. It just isn't true anymore. With 30GB hard drives quickly becoming a mass market commodity, few people worry about how much space a couple images off the Internet might consume.

The taboo is strong, but it is also doomed. Without a doubt, we will need the next generation CPU someday. We'll also certainly look back someday and laugh that we could live with a measly 30-40GB of local storage space. However, hardware is improving at a pace that is outrunning demand and the killer apps that will require those new levels of hardware are years away.

For creative professionals (obviously, my target audience), faster and more powerful systems are still in demand, at least for the 3D animation and special effects segments. Unfortunately, with the cart pushing less and less on the horse, the horse will eventually balk. In other words, if the mainstream isn't convinced that it must have the next level of CPU, the chip vendors are going to think twice about throwing billions into R&D to create the P5 and P6. If the chips only appeal to the relatively small market for supercomputers, the prices for these next generation chips will mirror the prices of supercomputers.

With all due respect to the makers of supercomputers, it isn't advanced technology that justifies their $100,000 and higher price tags. It is simply that the market is so small that each unit carries a significant R&D debt. The traditional mainstream PC market is on the verge of entering that same economic brick wall. Of course, this isn't necessarily a bad thing, if the energy and focus is quickly shifted from faster, faster to better, better. The automobile industry made a similar transition in the early part of this century (for you mathematically challenged, the current century ends this winter). Engines got bigger and more powerful, then, in the 30s, after a brief fling with 12- and even 16-cylinder monsters, efforts were shifted to improvements. Hopefully, the computer industry will make as smooth a transition as the auto industry accomplished.

top      home      search      user forum      subscribe      media kit      contact      webmaster@digitalmedianet.com