Update: David Richards is back at it again with totally inaccurate and bogus information. His earlier article information was according to a 'Microsoft insider' and now he attributes the statements to Raymond Vardanega (Acer Austrailia Marketing Director) who was told these things by an unidentified person at Microsoft. Net result: total hearsay and David is admitting to not verifying information and sources. Frankly, it would be nice for David or Raymond to identify who at Microsoft told them this. Anyway, in this latest article David states "A major problem for Microsoft is that the current version of Media Centre (MCE) is clumsy and prone to crashing. It contains code which is not productive in delivering an entertainment experience for consumers." Totally, unequivocally false. Read on to learn more...

I've been reading what Robert, Sean, Loren and the XBox team have already said about the now infamous '60% Of Windows Vista Code To Be Rewritten' article by David Richards. For the record, this 'story' is absolute poppycock. The only reason I'm blogging this is because I got word of mouth some of our MVPs didn't know what to think -- and I want all of them and our current / future customers to be 100% confident we are on track to deliver a safe, secure and stable operating system with cool features (like Windows Media Center, Windows Media Player, Windows Sidebar, etcetera, etcetera, etcetera) with Windows Vista.

Robert put it this way...

Even the evidence denies this story. At Mix06 last week we had Media Center PCs for people to use, running, gasp, Windows Vista. An entire keynote (damn cool demos too) ran on Windows Vista and it didn’t crash the entire time. That doesn’t sound like something that needs a 60% rewrite. Or something that isn’t on schedule to ship.

Let me give you the backstory for Roberts comment (he hasn't heard this until now and it will independently corroborate his observation)...

I was responsible for getting six Media Center PCs built from scratch in one evening (Sunday night) at Mix06. It took Ernie Booth and myself about 6 hours, mostly because we had to share two installation DVDs and a single USB key among the six machines and they were spread out over four locations and two floors. Setup went without a hitch and the only driver we had to update post-setup was the sound card driver (using the USB key) -- every other device on these Dell enterprise machines (read: not originally designed to run Windows Vista or Windows Media Center) used the out of box Windows Vista drivers. The machines ran *great* for the duration of the conference. Clemens Vasters of the Indigo team even sat down at one of these boxes and watched football (soccer for us Americans) streamed over the net from his home in Germany one evening.

In addition, I was responsible for the primary and backup Windows Vista machines for the Joe Belfiore keynote. Joe rehearsed on Monday night, and towards the tail end of the run through using this machine we noticed the album art wasn't loading quite as fast as expected and the audio took a while to start playing. The reason: We had been running the machine through it's paces, adding new content and syncing devices to it non-stop for about 8 hours with no reboot. After 8 hours of *actively* running a beta OS with Visual Studio 2005, Q podcast application using the Windows Media Center Presentation Layer, Apple iTunes installed (wouldn't you like to know why and FWIW was later uninstalled, which typically torques even my most stable Windows XP build) and greater than 10,000 tracks in the Windows Media Player Library it was performing pretty much up to snuff. A quick reboot (which I *always* do before an on-stage practice or live demo, but forgot to that evening for some reason) resolved all issues. The machine performed flawlessly after the reboot, again early the next morning during practice and again for Joe's keynote in front of 1000+ individuals (including Robert). We later used the exact machine on stage for our 'Developing For...' presentation at Mix06. It was a stock HP zd8000 laptop (again, read: not designed for Windows Vista).

So, you would think I was running the February CTP or some other build which had been vetted and throughly tested for consumption by the masses and appropriateness to use with a high profile keynote and the demo machines.

Nope.

I used the latest build from a development branch of the eHome source code tree which contained some functionality we needed to make the Q:Helix perform as expected. It was a random build which hadn't been through Build Verification Tests (aka BVT, the most basic of tests to qualify a build before broader release to other teams). It wasn't even from the main branch for Windows Vista (aka WinMain) which is what typically gets posted for beta testers. Generally speaking, if you select one of these builds you typically expect things to NOT work since the regression rate (regression=bug fix causes other things to break) can be pretty high. I had confidence any build I picked would work just fine.

And now you know...

...the rest of the story.