At Microsofts Developers Conference in Los Angeles, Bill Gates laid out the companys premises for Web services success.
Gates asserted that .Net is built on open standards. The trinity of TCP/IP for transport, HTTP for interaction and XML for meaning has become his belief system. Meanwhile, IBMs involvement has largely allayed developers concerns that XML-based SOAP would "embrace and extend" XML into a Windows-specific distributed computing protocol. Microsoft must understand that the industry remains predisposed to detect and rebel against any such maneuver.
Gates also said .Net is platform-neutral, inspiring laughter. Microsoft Group Vice President Bob Muglia added that core .Net services would be available to users free of charge—"certainly on Windows." Does that mean uncertainly elsewhere? The industry knows all about "runs better on Windows"; enterprise developers, we hope, have learned to avoid the self-inflicted lock-in of uncritical API adoption.
Gates promised that .Net will "preserve and connect" existing Web content, rather than impose the costs of "rip and replace" that have plagued Java developers. Microsofts long- awaited Visual Studio .Net development suite deftly directs the ubiquitous skills of Visual Basic and Visual C++ developers toward new .Net efforts; Microsoft engineers eagerly show the ease of augmenting Web sites with service-oriented shells.
In the same way, therefore, that Windows 3.0 preserved the value of DOS applications while also enabling new GUI designs, .Net appears to enable migration to service-based strategies at the convenience of enterprise planners—rather than on a timetable written in Redmond. Microsoft should complete that thought by supporting Windows 2000 for as long as enterprise buyers prefer to deploy it.
Gates lauded the power of .Nets loosely coupled model, with its prospect of inherent scalability and robustness in an imperfect network environment. Microsoft needs to prove that its security and fault tolerance infrastructures can carry the resulting weight of users expectations.
Lastly, Microsofts hopes for .Net rely on the traditional positive-feedback loop of broad industry support. The company may underestimate users aversion, though, to a continuing conversation between their PCs and the Microsoft mother ship that occurs in the HailStorm (now .Net MyServices), Passport and Product Activation technologies; the company should prepare to adjust its expectations of how much control PC users, and enterprise buyers, will readily yield.
These are the promises that must be kept if .Net is going to transform the digital world rather than merely giving Microsoft a handle to spin that world in its own direction.