Desktop Linux soothsayings

LINDOWS CEO Michael Robertson has ventured some interesting predictions about the near future of Desktop PC market evolutions.

Writing in his company website's "Michael's Minutes" column, Mr. Robertson predicts the following developments will occur next year:

1) No-Cost PCs -- Service providers will offer "free" PCs to new customers who commit to long-term contracts, like some cellular telecom companies provide free or heavily subsidized cellphones, and broadband services offer DSL and cable modems free to customers. He speculates that AOL might be tempted to jumpstart it's lagging rate of new customer takeup. He also hints that Lindows might offer free PCs to punters... bundled with it's Click-N-Run applications software download service.

This doesn't seem too far-fetched, actually. WalMart sells MicroTel PCs (with Via C3 CPUs inside) for $200 (Lindows) or $300 (Windows XP). Such low-end boxes are certainly fast enough for email and messaging, surfing the web, and playing MP3s, etc. One is tempted to ask "What about the monitor?" But WalMart also sells the same MicroTel system, with a 15" monitor included (and Windows XP preloaded)... for $400. Why WalMart isn't already selling the same systems with Lindows preloaded for only $300 is a mystery, but it shows that such low prices are possible. The WalMart online store web-page advertising these low cost PC offerings is here.

2) Windows XPLite -- Microsoft will be forced to offer a stripped down version of Windows XP at less than half the cost in late 2003. His theory is that the PC OEMs will balk at continuing to pay the Microsoft tax (which he estimates at upwards of $100 per PC including overhead) on dirt-cheap hardware. Yet even this fighting retrenchment by Microsoft won't derail the gathering storm of Linux on the Desktop, according to Robertson, so Linux will keep gaining market share throughout 2003.

One might like to see this as more than just wishful thinking, but it's hard. The major PC manufacturers have been in Microsoft's thrall for a long time, and it's not likely -- the eventual Antitrust case remedies notwithstanding -- that Microsoft will diminish its preload cash machine without a protracted struggle that makes the WWI Battle of Verdun look like a drunken fistfight at a picnic (metaphorically, of course). This will take significant market share gains by the white box vendors, plus defection from Microsoft by a second-tier vendor (Micron, or Gateway?) to force Microsoft's hand. But... stranger things happen in markets.

3) Linux Desktop 10% Market Share -- This is self-explanatory, as Robertson reasons from over 50% Linux penetration in the server space to anticipate growth of Linux on Desktops, albeit taking up more slowly.

Since he's not specific about whether he means 10% of the installed base or 10% of new installs, this prediction is difficult to evaluate. If he is looking for 10% of the installed base, sheer inertia suggests that he is dreaming (or, to put it more charitably, spinning for investors).

But, if he's talking about new PC Desktop purchases, perhaps he's right that 10% or more of the buyers of new PCs will be increasingly choosing Linux. It's already the choice for many build-your-own PC cognoscenti (except for hardcore gaming addicts), so it's possible that Linux will soon become the smarter choice for cost-conscious mainstream PC buyers. One suspects that penetration of the midrange+ PC markets ($600 and up) will be needed to support this scenario. See above comments on (2).

However, Robertson might be right, too. After all, the International Herald Tribune recently quoted IT market analysts IDC to the effect that Linux has already overtaken Apple, with 3.9% of Linux on Desktops surpassing 3.1% using Macs. If this is true, 150% growth for Linux on Desktops suddenly doesn't look so quixotic as a prediction, in a year. The International Herald Tribune article mentioned is available here.

4) Windows Wars Return -- This is where Robertson's predictions get all too interesting. He extrapolates from his earlier prophecies, and suggests that the PC OEMs will dump Microsoft Windows and develop their own branded desktop window managers... all based upon Linux (and KDE or Gnome, presumably). His argument is based upon an analogy with automobile manufacturers, and he asks: "Can you imagine a car company that didn't design the interior of cars they manufactured? Worse, what if all cars had identical interiors and there was only one supplier who marked them up so that they made more profits than all the actual car companies combined?" Well yes, that's ridiculous. Sounds plausible.

But I'm not buying it. This particular analogy is flawed. No coachwork company can offer a car interior that installs into a vehicle by reading a CD, and costs far less than an automaker can design and install it.

It simply strains credibility that PC OEM's will rebuild software staffs to compete with Linux distributions, and Microsoft, on user interfaces. Unless PC manufacturers start partnering with major Linux distributions and the KDE and Gnome teams, it won't happen. But, I could be wrong.

Robertson does make the point that Microsoft has succeeded in its quest to crush the OEMs through what economists call "collateral competition" in which one player leverages its power to take the majority of profits away from suppliers providing the other components of a product. His analogy that Microsoft has cornered all the major PC OEMs into selling razors, while Microsoft sells the high-margin blades, is so accurate.

He's right that the PC OEMs and Microsoft have mirror-image profits and costs for PC systems: manufacturers scrape by with single digit profits and 90%+ marginal costs, while Microsoft enjoys very low marginal costs with huge profits. But, I don't see PC OEMs stirring to change this, at least not next year. They are too dependent on Microsoft to dare it, I think. However, about this prediction, I'll be happy to be in error.

5) No All-In-One PCs -- Companies will make lots of specialized PC-based appliances for in-home use... gaming (Xbox), television (Tivo, ReplayTV) and music (Rio, SonicBlue), etc. (even kitchen, home finance, according to Robertson), but won't develop and sell all-in-one PCs that can do all these things. According to Robertson, consumers want simple interfaces... and PC-based appliances running Linux underneath do it.

I can't think of anything to call this assessment in question just now. It's canonical that many people who own VCRs can not figure out how to change the time display (the "blinking 12:00 problem" it's been called) and one imagines this general level of cluelessness about consumer tech will only persist. And custom Linux will surely drive these devices.

Mr. Robertson's prophecies for fierce competition during this next year in the Desktop PC market are here. µ