I have taken Microsoft to task for being too concerned about compatibility at the expense of making long overdue fundamental changes in Windows. But the business concerns that lead Microsoft to be excessively cautious about changing Windows don’t apply at Microsoft’s research arm, which is developing a radically new OS called Singularity.
The primary design objective of Singularity, which Microsoft made available to researchers for free download on March 4, is to increase the security and reliability of computing. One way it does this is by imposing much tighter restrictions on how running applications can communicate and otherwise affect each other using a new approach Security Isolated Processes.
Don't expect this technology to turn up in a version of windows anytime soon. Microsoft makes it clear that singularity as it exists today is a prototype that is a very long way from becoming a real, usable operating system. For one thing, the current version is designed to run only as a virtual machine on a Windows system running Microsoft's Virtual PC software.
Still, these research projects can be very important in the long run. For 1985 to 1994, a team of computer scientists at Carnegie Mellon University developed a core operating system called Mach. Two of the leaders of the Mach project were Avie Tevanian and Rick Rashid. Tevanian took his work to Steve Job's NeXT, where it became the core of the NeXTSTEP operating system. Tevanian followed Jobs to Apple, where the Mach-based NeXTSTEP evolved into Mac OS X. Mach has also had a profound effect on the development of a number of other contemporary operating systems.
As for Rashid, he too left academia for the corporate world, where he is now senior vice president in charge of all of Microsoft's research efforts, including the singularity project.