No one wants to reinvent the wheel. The PC is another story.
It’s not so much that we want to see bigger, flashier computers. Technology typically starts out with an enormous form factor, after which vendors are pushed to reduce the footprint to the point of near-invisibility. We’re seen it
with servers, which have been trimmed down to blades. We’ve seen it with laptops, which are finally thin and light enough to take on trips without having to contact a chiropractor afterwards. So how did the traditional PC get left behind?
The answer, according to research firm iSuppli Corp., is the slow transition from ATX motherboards to something that can take advantage of newer interface connections like Serial ATA and PCI Express. In a study released Monday, iSuppli said those improvements are finally here and could usher in a slew of smaller, more efficient PC designs throughout 2004.
Intel has been working on Balanced Technology Extended (BTX) interfaces since at least last year, which give developers options to balance thermal management, acoustics, system performance, and size.
ATX has been the standard among motherboard makers like Via and Asus practically since it was first developed in 1990. That seems like a long time for an industry that’s supposed to thrive on change, but in some respects PC evolution presents a chicken-and-egg conundrum. Do you develop the motherboard first and then try to get the interfaces to work around it, or do you hammer out the interfaces and recreate the motherboard to take advantage of them?
The latter approach has been the favoured one, and its only shortfall is that it can inflate expectations around the adoption of new interfaces. Things like Universal Serial Bus (USB) 2 or Serial ATA may not seem particularly compelling without the motherboard to back them up. It’s kind of like when Intel releases a faster processor without first making sure software developers have applications that require faster clock speeds.
With the BTX specification now available, however, it will seem more natural that manufacturers exploit interface enhancements. PCI Express not only promises to shrink the size of circuit boards and ease the way network cards are installed, but also simplify the motherboard design process.
Serial ATA may also benefit, but you never really know what will happen to these specs once they’re out in the real world. When the first versions of USB came out in 1998, everyone talked about the possibility of “”daisy-chaining”” peripherals from one to the other, but few OEMs went to the expense of turning their scanners and printers into a hub for other devices.
The end result of all this work should be smaller, quieter PCs at a time when some organizations no longer need them. If more desk space is all that important, why not simply move to desktop-replacement notebooks? There may be a better chance for these kinds of systems in the home networking market, where style and footprint are more attractive selling points, but enterprises have proven over the last two years that hardware is not something for which they are willing to pay a premium.