Unless we see >70% of devices running touch before Haswell, I fear even Haswell might not be enough to stop the decline let alone increase PC sales. That means we won't see something happening until at least 2014, presumably when Windows 9 launches.
Haswell will do what it was designed to do, which is create the perfect balance of a "PC-first" device. It's a PC, first, but it will also have a tablet mode, carry perhaps extra weight and thickness, but that's mostly driven by more features, a larger display, more ports, etc. At 10W, you won't be able to do fanless, but you'll probably have excellent battery life with a 50 Whr battery, along with the weight and the thickness that larger batteries drive.
No, Haswell is not a magic solution for building an iPad-like device, but Clover Trail is. You are so down on Atom, even though Atom is the RIGHT solution for competing head on, without the impacts to weight, thickness, or battery life. Atom can do fanless with 1.7W of TDP. It was designed to have the same power as ARM SOC's, a competitive cost structure, and leadership performance. Let me know when you find a tablet app that can't run on Clover Trail - including the 3D games. An Imagination SGX 545 core may be somewhat old, but reviewers claim it keeps up with all the latest tablet-optimized 3D games, so that's a checkbox feature, even if the benchmarks are lower. As for CPU performance, it delivers a better experience than even quad core ARM A9 cores, and probably close enough to dual core ARM A15 when compared on the shelf.
So why argue that Microsoft needs to change? Even with today's Windows 8, Clover Trail makes the best mainstream tablet that anyone can buy today. It has more features, and more capabilities, at a competitive price point, once you normalize the storage and keyboard options.
I'm sure Haswell will make excellent top-tier super-tablets and convertible devices that excel in a PC-first experience, but if people want tablet-first next year, they are only going to get it from Clover Trail.