October 28, 2003 9:22 AM PST
Longhorn: Hurry up and wait
Microsoft has not said when the final release of the operating system will come, but it has offered some hints of the path it plans to take. "I think the time line from when we go to beta until when we're done will be very consistent with other OS releases," said Joe Peterson, a Microsoft vice president.
Given the company's track record with Longhorn's predecessors, Windows XP and Windows 2000, that means that two to three years is a likely time frame, meaning that Longhorn would debut in late 2005 to early 2006.
"We're currently planning three betas," or test releases, of Longhorn, Peterson said.
Microsoft is hoping that it will have most of the hard work done by the time it releases the first full test version of Longhorn next year. "A lot of the hard work will be getting to that first beta," Peterson said. "We don't expect that tail end of the project to go any longer" than typical Windows releases do, even though this upgrade is more involved.
Given how ambitious Longhorn is, a delivery date three years away is realistic, said Stephen O'Grady, an analyst at research company RedMonk.
"This is really a significant revamp of basically every level of the operating system--everything from the file system to the presentation and pretty much everything in between," he said. "There's a realization of the enormity of what (Microsoft) is doing."
O'Grady said the timetable for Longhorn will have an impact on Microsoft's application divisions. The company is planning Longhorn versions of its different products, but in some cases, such as its server applications, Microsoft will need to weigh whether it should hold off on a release to synchronize with Longhorn, he said. "It'll be interesting to see the trickle-down effect and what other impacts this time frame may have," he added.
Indeed, many elements of Longhorn are still not in place. Microsoft has not yet decided which technologies will be in the final Longhorn release. And while Microsoft executive Hillel Cooperman on Monday showed off what Longhorn's desktop will look like, that code is not part of the Longhorn code Microsoft gave developers.
"The particular look and feel that Hillel showed is not on that" developer preview, Peterson said. However, he said, the features and programming hooks that are needed to enable such interfaces are available to developers.
As for the Longhorn user interface itself, "that's something we will release when we get to the formal beta of the product," Hillel said.
Also, not all of the developer interfaces, such as WinFX, are ready. And even if they were, and developers built programs, no one would have the OS to run them on.
"Certainly if you write to WinFX today--the new API (application programming interface) set--that is unique to Longhorn," Peterson said.
Given the early nature of Longhorn, what does Microsoft suggest that developers do?
Peterson and other Microsoft executives said developers should start creating prototype applications for the new OS. In doing so, they can give feedback to Microsoft about what is good and what isn't and thereby have a say in what actually goes into Longhorn.
And although it may be too early for many developers to start creating Longhorn programs, executives say that by writing managed code and developing Web services, programmers will find themselves in good shape when Longhorn finally debuts.
Developers "can start moving to using the .Net framework and start writing managed code," Peterson said. "Making those two moves today will position you very well for Longhorn."
Adam Sohn, a product manager in Microsoft's .Net platform strategy group, said, "If all you do coming out of this (Professional Developers Conference) is realize that you should take advantage of managed code on the (.Net) Framework, you are going to be ahead of the game for Longhorn."
But given that Longhorn is still years away, Microsoft knows it runs the risk of pushing developers too soon and potentially creating a backlash, particularly if the company's schedule slips further.
"We certainly thought about that," Sohn said. Nonetheless, he said, the benefits outweighed the risk. "First of all, we needed the feedback."
How seriously programmers actually start writing for Longhorn depends on how much lead time they have, he said.
"For some of these guys, who are on really short release cycles, it's a planning exercise now," Sohn said. "For people who are on longer release cycles, they might have more to do now."
CNET News.com's Martin LaMonica contributed to this report.