Anatomy of a Falsehood: Windows 2008 20 Percent "Faster" than Vista?
There's a bit of baloney floating around the Internet ether right now, stating that Windows Server 2008 is somehow magically "20 percent faster" than Windows Vista and is, thus, the ultimate workstation OS. A story making this claim, from Blorge, was quoted everywhere online so I decided to see how this event unfurled. Sure enough, the title of the Blorge article, "Windows Server 2008 is 20% faster than Vista," seems unequivocal. But jump into this so-called story and you'll see the waffling begin. After configuring Windows 2008 properly (a process that "isn't easy" according to Blorge), the new OS "can be 20 percent faster than Vista." But it doesn't say how this claim was arrived at, linking instead to an Information Week article declaring that Windows 2008 is the "speediest and most secure version of Windows to come along in a decade." But how does InformationWeek determine this? Instead of actually testing the performance of the system itself, this site links to a blog whose tests show that Windows 2008 is "almost 20 percent faster ... in some tests" than Vista. Ah. "Almost." "In some tests." It turns out that the "almost 20 percent" gain occurred in only one performance benchmark, and that test was performed on just a single machine. How much testing was done overall? They ran three benchmarks. Three. And the average performance delta between Windows 2008 and Vista was actually 14 percent, not "almost 20 percent." As for that one test that was "almost 20 percent," the actual performance delta was 16 percent, not 20. But hey, what the heck: There's no reason you can't turn one test, by one blog, performed once on one machine, info definitive proof that Windows 2008 is 20 percent "faster" than Windows Vista. I mean, after all, I'm sure the test was done fairly and correctly the only time it was performed. By whoever did it. But we do know why they did it: A Microsoft blogger, they say, shows us how to turn Windows 2008 into a Vista knockoff, "one that's faster and more scalable than the original." Except that that's not what this blogger wrote at all: Instead, he wrote only that you could make Windows Server 2008 "look and feel" like Vista and made absolutely no performance claims at all. In fact, implementing some of his dubious advice (enabling Hyper-V) will actually turn off power management completely. I wonder if that accounts for that 20 percent performance boost? Sorry, I don't mean to over-think this. Unlike some other people.
Follow the progression, my friends, and discover how myth becomes reality on the Internet:
Culprit #1: Activewin and dozens of other sites where no juicy headline need be scrutinized
http://www.activewin.com/awin/commen...=43039&Group=1
Culprit #2: Blorg, where "almost 20%" becomes a definitive 20 percent
http://vista.blorge.com/2008/03/11/w...er-than-vista/
Culprit #3: InformationWeek, where it must be true if someone else says they tested it
http://www.informationweek.com/blog/...oft_relea.html
Culprit #4: xpnet blog, where misquoting a blog leads to 3 tests on a single machine
http://exo-blog.blogspot.com/2008/03...one-right.html
Misquoted originator of this mess: Vijayshinva Karnure, who may very well have valid reasons for running Windows 2008 as a workstation OS
http://blogs.msdn.com/vijaysk/archiv...esktop-os.aspx
Read 'em and weep, both for the future of journalism and for the continued right-wing conspiracy (for lack of a better term) against Windows Vista.
