WinHEC 2007: Time for Vista to Deliver the Goods

3. Making input meaningful again. For decades, Microsoft has had an interest in getting a handle on the interactive tool that may inevitably render the mouse obsolete. Microsoft Research has actually made some serious headway over the years in natural language recognition. But time after time, the company's ability to realize the vision of its research staff has fallen far short of its lofty goals. This year at WinHEC, we're told we should see some actual breakthroughs in the field of next-generation input. What we're expecting is that the company made some headway in gesture recognition and speech analysis. Granted, these are the basic two stumbling blocks, and we've stumbled over them for so many years now that even the pulp science monthlies find them uninteresting any more. But if Microsoft has at last made a leap over these stumbling blocks, then we may get our first look at the next, next generation of Microsoft Office.

2. The changes in the system kernels for Vista and Longhorn. The chief reason Microsoft's kernel design for Windows was changed was to introduce a new security model at the lowest level. In fact, that low-level security change is so drastic as to have caused controversy among security vendors who now no longer can protect the system kernel the way they used to. But Windows Server "Longhorn" is where these kernel changes may matter the most, because although any network attack may take place on the client, it will target the server. We have been told to expect a new order of beast with regard to Longhorn, which has just now entered its general first public beta round, but even newer builds of which may be on display at WinHEC.

1. Virtualization, virtualization, virtualization. For every practical reason you can think of - security, practicality, cost savings, ease of administration, sheer coolness - the ability to virtualize Windows Server within a host whose own hardware knows how to keep it virtualized, is easily the greatest advance in Windows technology in a decade. Creating an intentional barrier of abstraction between the low level system hardware and the high level operating system will radically refine how servers are designed and built, far beyond just implanting them with smarter BIOSes. It also opens the door to the Trusted Computing Platform - perhaps the most effective means to secure the integrity of system hardware thus far developed, yet controversial because its trust can be leveraged to deploy DRM on a massive scale.

For weeks, Microsoft was touting its best batters at the plate at WinHEC, ready to hit several pitches clear out of the park. But then the company may have hit a foul ball, choosing not to deliver on several key features of Windows Server Virtualization in time for Longhorn's RTM. Soon after that decision was made, several virtualization demos Microsoft had planned at WinHEC were cancelled - demos that would also have been showcases for recent advances by partners Intel and AMD, who now may have to go it alone.

So a huge question system builders and admins alike will be asking all this week is, what's up (or down) with virtualization? And if Microsoft can't deliver it, perhaps there's someone else who can. We will keep our eyes and ears wide open this week for instances of the word "VMware."

BetaNews will be filing reports throughout the day from the Windows Hardware Engineering Conference in Los Angeles, all week long. Stay with us for some hot topics, perhaps a few solid answers, and a whole lot better questions, as we find out how -- or whether -- the personal computer and the network server are ready to make some big changes.