The "good enough" PC
PC World has a treatise on “good enough” computing. This isn’t actually a new trend but it’s never stood still for as long as it has now.Jerry Pournelle used to describe cheap CPUs from Cyrix and IDT in the late 1990s as “good enough.” Running at 166 and 200 MHz, they ran Windows 95 and NT4 and Office 97 just fine. They weren’t good gaming CPUs, but for everything else, they were great, and you could build a computer with one of those and save $100 or more over using a comparable Intel CPU.
Trouble was, the mainstream moved. Intel knocked off all the upstarts by starting a megahertz war, and AMD came back from a near-death experience to compete. The requirements to run Windows increased nearly as rapidly, and it wasn’t all that long before 900 MHz was pretty much the bare minimum to run Windows comfortably.
But chips kept getting cheaper, and today you can buy a 2 GHz CPU for pretty close to what a Cyrix or WinChip CPU cost. But you get more than 10 times the power for that money. And Windows XP runs perfectly comfortably on a 2 GHz CPU, whether it’s a new Intel Atom or Celeron or a 5-year-old corporate discard. So does Office 2003, which is the very last version of Office that any sane person would want to use.*
*Besides being the evil spawn of Windows Vista and Microsoft Bob, Office 2007 also crashes more often than Windows 3.0 did. The only way I can go a week without losing work from Office 2007 crashing is to go on vacation.
The PC World author claims that Linux and Open Office running on Intel Atom CPUs will be the undoing of Microsoft. I think that’s a bit of a stretch. Netbooks running Linux got returned to the vendor a lot. I suspect the biggest reason is because they probably couldn’t figure out how to get their USB mobile broadband cards–I’m talking the stuff that cellphone vendors offer for 50 bucks a month–working in Linux. That, and they probably couldn’t get Flash working so they couldn’t see Facebook and other popular sites the way they could on their regular PCs.
Frankly, the two things that keep me from buying a $200 Dell Vostro netbook this weekend are the price of mobile broadband ($50 a month), and my concerns about the reliability of anything sold by Dell in the last 5-6 years. I work with a lot of Dell equipment, and once the warranty goes, their machines do not age gracefully at all. But I think Dell will sell a lot of these units, because the price is absurdly low, they weigh two pounds, and they run anything but 3D games and intensive graphics apps nice and fast. Sure, a dual-core system with its memory maxed out and a solid state disk will outrun it, sometimes even running circles around it, but that system will also cost 10 times as much.
I do think Office 2007 is the best thing that ever happened to Open Office. Open Office’s interface is a lot more familiar and doesn’t hide anything, and while it may not be as fast as Office 2003, it’s certainly faster at most things than Office 2007 is.
Linux has been usable for basic computing for a very long time, but getting it installed and configured remains a challenge at times. A netbook that connects painlessly to the wireless networks in restaurants and to cellphone makers’ mobile broadband cards while running Linux probably stands a chance. Giving some automated, easy means to synchronize application data and web bookmarks between the netbook and a desktop PC would probably help a lot too–something that does the same thing that Activesync does for moving data between Windows PCs and Windows Mobile PDAs. Will these things happen?
But I do think an era of “good enough” is upon us. There was a time when the top-of-the-line PC would be entry level within a year or two, and that’s not really true anymore. The entry-level PC of today is comparable to the mid-range PC of five years ago. For most of my lifetime, basic computing on a five-year-old PC was always painful, no matter how good that PC was when it was new. That’s not the case today.
Graphic designers, video producers, and scientists will always need ever-more powerful systems for their work, so they’ll continue to drive the cutting edge. But everyday computing is stabilizing. I don’t think Intel wants the future of everyday computing to be the cheap Atom CPU, but at this point it may be impossible to avoid it. If Intel decides to quit playing in this space, AMD can design something comparable to replace it in the marketplace. The Geode won’t cut it, but something based on the Athlon XP architecture and built using a modern process certainly would.
And frankly I’m glad about this development. It’s been nice not having to buy a new computer every three years or so.