JANUARY 2006 :: BIG BUSINESS

Rebooting Windows
Competition Forces Microsoft to Change the Way It Builds Software

By Robert A. Guth
Staff Reporter of The Wall Street Journal

Jim Allchin, a senior Microsoft executive, walked into Bill Gates's office one day last year to deliver a bombshell about the next version of Windows: "It's not going to work." The new version, code-named Longhorn, was so complex, its writers would never be able to make it run properly, Mr. Allchin said. Microsoft would have to start over.

Mr. Allchin's warning led to what he and others call a transformation for Microsoft's most important product, at a critical time for the company. Microsoft, though dominant in computer software, faces a growing threat from rivals such as Google, Apple Computer and makers of the free Linux operating system. In recent years, these companies have been dashing out some software innovations faster than Microsoft. Google is especially good at introducing new programs such as email and instant messaging over the Internet, watching how they perform and replacing them with improved versions.

Microsoft can't entirely replicate that approach with Windows, since the software is by its nature a massive program overseeing all of a computer's functions. But Microsoft is now racing to move in that direction: developing a solid core for Windows onto which new features can be added one by one over time.

As always, Microsoft's great fear is that it will lose its near-monopoly on computer operating systems and basic office software. In the short term, there is little danger of that. But the more Google and other software makers encroach on Microsoft's turf, the greater the chance that someday computer users will wake up and find Windows superfluous.

"What happened when the American car companies failed to update their manufacturing lines? There was a more efficient way to bring cars to market for a lower price and they lost their market," says Microsoft Vice President Chris Jones. "We're in a little bit of a different industry, but it's the same thing."

Mr. Allchin's reforms address a problem dating back to Microsoft's early years. Traditional computer science called for methodical coding practices to ensure that the large computers used by banks, governments and scientists wouldn't crash. But as PCs took off in the 1980s, companies like Microsoft didn't have time for that. PC users wanted cool and useful features quickly. They tolerated-or didn't notice-the bugs riddling the software.

Fast and Loose

Mr. Allchin says he never liked the fast-and-loose culture of PC software. But there was little he could do. As soon as Microsoft finished one version of a product, it moved on to the next.

When the project to build a successor to Windows XP got started, teams of engineers set off to develop it as they always had. Mr. Gates was especially eager for them to add a fundamental change to Windows called WinFS that would let PC users search for and organize information better.

Mr. Allchin says he soon saw his worst fears realized. In making large software programs, engineers regularly bring together all the new unfinished features into a single "build," to test how they work together. But with 4,000 engineers writing code each day, testing the build became a mammoth task. When a bug popped up, trouble-shooters would often have to manually search through thousands of lines of code to find the problem.

Meanwhile, top engineers were being diverted to the effort to fix a rash of security holes in existing versions of Windows. "The ship was just crashing to the ground," Mr. Allchin says.

In late 2003, Mr. Allchin called on two men to help. The first was Brian Valentine, known for his ability to turn around troubled software projects. The second was Amitabh Srivastava, a fellow purist among computer scientists. New to the Windows group, Mr. Srivastava had his team draw up a map of how Windows' pieces fit together. It was 8 feet tall and 11 feet wide and looked like a train map with hundreds of tracks crisscrossing. That was just the opposite of how Microsoft's new rivals worked. Their best programs were like Lego blocks-they had a single function and were designed to be connected onto a larger whole.

While Windows itself couldn't be a single module, it could be designed so that Microsoft could easily plug in or pull out new features without disrupting the whole system. That was a cornerstone of a plan Messrs. Srivastava and Valentine proposed to their boss, Mr. Allchin. Microsoft would have to throw out years of work on Longhorn and start out fresh. Some features, including Mr. Gates's WinFS, would have to wait.

Reset

On Aug. 27, 2004, Microsoft said it would ship Longhorn in the second half of 2006-at least a year late, and without WinFS. Mr. Allchin announced to hundreds of Windows engineers that they would "reset" Longhorn using a clean base of code. By late October, Mr. Srivastava's team was beginning to automate the testing that had been done by hand. If a feature had too many bugs, software "gates" rejected it from being used in Longhorn. If engineers had too many bugs, they were banned from writing new code.

The quality of the code flowing into Longhorn soon began to improve. The time to create a new "build" fell to just a few days, allowing a faster cycle of writing and testing. Last winter, the Windows group was able to install a workable version on their PCs.

On July 27, 2005, Microsoft shipped the beta of Longhorn-now called Vista-to 500,000 customers for testing. Experience had told the Windows team to expect tens of thousands of reported problems from customers. Instead, there were a couple thousand.

And in August, Microsoft delivered a test version of Mr. Gates's WinFS idea-not as a part of Longhorn but as a planned add-on feature. Microsoft recently said it would issue monthly test versions of Windows Vista, a sign of the group's improved agility.

It could take years before Windows can be as flexible as Microsoft needs it to be to pump out new features quickly. But the cultural shift is in swing. Microsoft's Office group is now using some of Mr. Srivastava's tools to improve its code. "It's amazing the invention those guys have brought forward," Mr. Gates says. "I wish we'd done it earlier."