In late August and early September, I started hearing rumors that the announced November launch of SQL Server 2005 might be delayed yet again. I couldn't imagine that, with a 5-year gap since the last release, Microsoft could afford another delay, but something had to be behind the rumors. Then, during a keynote at the Professional Developers Conference (PDC) in mid-September, I received an emailed update on SQL Server. In that update, Paul Flessner, senior vice president of server applications (which include, of course, SQL Server), said, "We believe our Database Mirroring feature needs more time in the hands of customers before we make it generally available for production use...and \[we\] will release it for general use as soon as you tell us it is ready."
Ah! That explained the rumors. The feature was pulled so that the release didn't have to be postponed yet again. Smart move.
Although database mirroring was one of more exciting features of SQL Server 2005, most people would rather see database mirroring done right and would also rather have SQL Server 2005 without database mirroring than experience another delay. And Microsoft has decided it can't afford to have product schedules that slip for years. Holding back features that don't meet a quality bar is necessary, but products have to go out more than once every 5 years.
So here we are. It's November, and unless something drastic happens after this column goes to press, SQL Server 2005 will debut on the 7th, without database mirroring. (See Michael Otey's Web exclusive article, "SQL Server 2005: An Evolutionary Release," InstantDoc ID 47974, for an overview of the product.)
Maybe the quality was worth the wait and Flessner's update emphasized the commitment not to ship until the quality was assured. But at the same time, the pressure within Microsoft to ship products on a regular timetable is evident. Our sister publication, SQL Server Magazine, publishes a special SQL Server 2005 and Visual Studio (VS) 2005 launch issue this month, and for that issue, I interviewed Flessner and his VS counterpart, S. Somasegar. (See SQL Server Magazine, "Value in Integration: SQL Server 2005 and Visual Studio 2005," November 2005, InstantDoc 47876.) In our conversation, both Flessner and Somasegar stressed the need to ship products every 2 years. Flessner said, "I'm back on an every-24-months release cycle," and Somasegar said, "...we absolutely want to get to an 18- to 24-months cadence."
But SQL Server is only the latest in a string of late-to-market Microsoft products, including Longhorn client (now called Windows Vista) and Longhorn server and going all the way back to NT 5.0—I mean Windows 2000. With Microsoft's recent VP-level reorganization (and undoubtedly more personnel adjustments are coming), the company is scrambling to address its seeming inability to quickly respond to market needs and get products out the door on schedule.
What's MSN Got to Do with It?
A first glance at Microsoft's new organization could leave you feeling like something is wrong with the picture. How did MSN, which had been in its own separate division, get lumped with the Platform Products and Services Division? (Hint: That's the division that includes the products with the worst shipping records.) Steve Ballmer answered that question in the announcement of the reorganization: "Our MSN organization has great expertise in innovating quickly and delivering software-based services at scale."
Case in point: Desktop Search. WinFS was supposed to be the foundation for Vista's key innovation, desktop search. However, WinFS kept failing to materialize, which left the Vista team seeking alternative technologies to drive desktop search, which kept delaying the launch. Meanwhile, MSN (which has been considered something of a low-status stepchild within some Microsoft divisions) was quietly, quickly, and successfully launching MSN Desktop Search to counter Google's advances. Not coincidentally, MSN Desktop Search provides a lot of the functionality Vista and WinFS had promised. While the Windows Big Men on Campus (BMOC) were falling behind, the scrawny MSN guys on the other side of the highway were kicking BMOC butts.
Will bringing MSN into the mainstream help Microsoft innovate and become more agile? Will the BMOC be humble and open enough to learn from those nimble MSN guys? Or will Windows impede MSN's agility? Driving change is intensely hard—especially at Microsoft. But accepting change in the trenches is even harder. And, as much as they might deny it, Microsoft's rank and file are as reluctant to change as are members of any big, entrenched bureaucracy. MSN had the advantage of being underestimated and left to its own devices. I hope MSN's underdog drive can survive being thrown into the Microsoft mainstream and that SQL Server 2005 will be the last product with a 5-year release cycle.