In its context at Camp Shelby the expression summarised an attitude to received wisdom: if it doesn't work, stop doing it.
So how's that look-like-Windows thing working for the Linux community? Is the wave of desktop adoption far ahead of where it was in 2001 and 2002 when this started? And, if not, why don't we stop doing it? Is it because we're stuck on stupid?
Sun's Java Desktop System was intended to look and act as much like Windows as possible without incurring legal liabilities for copying and yet run on Linux. It replicated Microsoft's x86 based client server architecture, and delivered the Windows experience, minus the frequent failures and data losses, on Linux.
Lindows was going to do that too, at the consumer level. Wal-Mart saw itself delivering American-made white boxes with American Linux software to American customers. Linux uber alles, and all that. So what happened?
Quite a lot of things: SuSe fell, the Chinese accepted Sun's software and expertise but not their invoices, IBM continued to consolidate its hold on Linux, Lindows became Linspire, Wal-Mart faltered, and overall desktop Linux seems to have pretty much stalled.
This approach seems to have failed. We're not only no further ahead now, we may actually be further behind because we seem to have lost momentum.
Microsoft employee and fellow blogger John Carroll says dropping this approach will mean the end of Linux but ask yourself: what would total success make Linux? Right, a copy of Windows, and that's not really the point, at least for us, is it?
What really scares Microsoft is innovation, not copy-cat desktops. If you want to see the kind of visceral hatred fear inspires, check out Ballmer's antics over Google. Or, for a more subdued example, read Doug Barney's put down of the Unix business architecture (big servers, smart display desktops) in a recent RedmondMag Editorial.
Here's a bit from his comment on Oracle:
Larry Ellison boldly promoted the Network Computer (NC), a zero-intelligence Internet screen scraper that would sell for as little as $200. Uneven Internet bandwidth was one problem. Citrix and Microsoft thin clients were another, but the real killer was full-powered PCs that gave NCs a beating worthy of Bruce Lee. This vision died a painful, public death and Ellison has been uncharacteristically quiet about his pronouncements ever since
Of course he attacks Sun too, but what he's trying to do is cement a lie in the reader's mind: the network computer is dead. But it's not - it's making a resurgence as people get over the stuck on stupid business of copying Microsoft's client-server architecture to go after the real desktop productivity killers: uncertain reliability, insecurity, and the limitations the fear of failure puts on user desktop experimentation.