Saying that Windows "Workstation" 2008 is faster than Vista doesn't make it so

I thought I'd left this whole Windows "Workstation" 2008 stuff behind, but it seems that I'm back for more ...

A few eagle-eyed readers sent me a link to a piece written by Randall C. Kennedy on InfoWorldwho seems to take exception to the fact that I had the audacity to disagree with him over the claims that Windows Server 2008 can be transformed into a better desktop OS than Vista. Apparently, anyone who disagrees with him is suffering from Technology Attention Deficit Disorder (TADD).

In my response here I'm going to ignore Kennedy's childish ad hominem attacks (and the fact that he seems to think that writing about this topic is somehow "ripping off" his ideas) and stick to the facts.

Let's take a look at the evidence that Kennedy presents.

Never mind the fact that I already wrote about the detailed benchmark results proving that Windows Server 2008 outperforms Windows Vista on identical hardware. That was months ago, and we all know that the TADD crowd has trouble paying attention during those boring test results discussions.

OK, so what Mr Kennedy is brining to the table is a study done by exo.performance.network. In fact, Mr Kennedy is so sure that this data is 100% reliable that he mentions it twice:

Vista vs. XP – Windows XP is ~40% faster than Windows Vista on identical hardware. This isn’t an estimate. It’s the raw performance delta that engineers over at the exo.performance.network (www.xpnet.com) measured while testing Windows XP (SP3) vs. Windows Vista (SP1).

OK, I hate to have to say this but since when did the DMS Clarity Tracker Agent (the tool used by exo.performance.network to collect the data) become an industry standard benchmark? As far as I can see about the only people claiming that it is are the folks running the site. In fact, it's hard to find a benchmark tool that officially supports server OSes (for example, Futuremark benchmarks don't support Windows Server 2008). One that I've used that does is PassMark PerformanceTest 6.1 and on the systems that I tried out, I saw nothing of interest. Some tests favored Vista, others Server 2008, but overall the results were too close to call.

When Kennedy pleads with us to "please, people, get the facts," I'm staggered that the only source he can quote is exo.performance.network. Maybe it's him who is suffering from TADD, and that is preventing him from bothering to do any testing of his own.

Bottom line, if you are going to present data to support your case, either get that from a reliable source or do the legwork yourself. Relying on a single shaky source is one way to set yourself up for an epic FAIL. His piece is nothing more than the  “my wife makes the best pumpkin pie” or “I’m a great driver” stuff that I warned about in my previous post on the Windows "Workstation" 2008 myth. Without robust data to back it, it's just bunk.

Sorry Mr Kennedy, but it sounds to me like you really WANT TO BELIEVE that Windows "Workstation" 2008 is better than Vista. Well, whatever floats your twig ...

Thoughts?