While it does that, it should make Windows 8 the first release that breaks with the past by moving all legacy technologies into a sandbox a la what OS X originally did.
Doing so would likely break many of the duct-tape and hot-glue bodge-jobs holding together a slew of mission-critical legacy applications. For example, I've worked at four very big financial institutions and interacted on a daily basis with various systems of a dozen others. And I've noticed something they all have in common. Their "legacy" systems were built decades ago and connected to via an assortment of terminal emulators such as Attachmate, Hummingbird, and others I only dimly remember by their gaudy splash-screens. In fact, the workstation image contains several different terminal emulators as some systems only work properly with a specific version or brand. Some even have funny little wrapper-programs that run on top of these terminal emulators.
And then there are some other applications developed in-house or by a hired third party that present a whole new interface, connect to different databases, yet still interact with the terminal emulators in unclear and often bizarre ways.
And what web-apps there are only work properly in IE6. Maybe IE7 if they're lucky.
And then there are the Excel macros . . .
How would you sandbox these applications without breaking all the tenuous and shoddy linkages that hold the whole house of cards together?
It's easy to say "screw 'em, shoulda coded right in the first place" but a major financial institution isn't going to roll out a new operating system that forces them to redevelop large swaths of their legacy application base without a damn good reason. These same banks have been "working on" migrating away from legacy apps for the better part of two decades and all they have to show for it are a couple of web apps and thin graphical veneers shoving commands through a terminal emulator and ridiculous piles of shockingly retarded Excel macros. They'd just as soon stick with what they have.
If a researcher discovers something surprising, the next steps are confirming their results and measurements were accurate and are repeatable. Then experiments can be devised to test why this might be so.
Nobody should do much believing in science. String Theory, Dark Matter and Dark Energy aren't things to be believed. They're just potential and incomplete explanations for what might be going on. The next step is trying to devise experiments to detect these things and/or test the implications.