Here's something I'd like to get off my chest. Whenever you sell something you NEVER bash it, make it appear inferior than other products or destroy the consumers' confidence in your product, your engineering ability or ability as a manufacturer. If you do so people will not buy your stuff, even if it fits their needs like a glove.
I'm 100% sure that Intel knew that prescott wasnt getting anywhere way before they launched it. They knew tejas was worthless way before they announced it was getting canned--they probably started moving staff off it right after they figured trying to tame its heat was going nowhere, but they had to time their announcements right in order to achieve two somewhat opposed goald: a) not damage consumer confidence and b) show shareholders that they had a plan besides just throwing megahertz at the problem. Business consumers are very coy about any decision that might involve risk since everyone in management is trained to avoid it. If Intel had canceled tejas before or right after Prescott came out they would have looked for alternatives or just held back on their purchases of new equipment. So instead Intel had to shout to the world that Prescott was great, that it would scale magnificently, etc. Then, when the prescott's penetration was enough they had to appease the shareholders that didnt see much of a future for Intel's netburst. So I would say that even if conroe wasnt officially called a "desktop" chip it last year, it was being developed with the desktop as its target.