The long march to Longhorn

Longhorn needs a new name.

It's not just because I'm tired of the bovine references and the "shorthorn" jokes. No, the problem is that the future version of Windows formerly known as Longhorn in many ways won't resemble the Longhorn that Microsoft described last fall at its Professional Developers Conference.

Back then, the centerpiece of Longhorn--and the source of much of the excitement and anticipation surrounding its release--was a revolutionary new storage and file system called WinFS. Microsoft said WinFS would finally allow us mere mortals to easily find answers to simple questions using plain old English (or Chinese, Italian, German or what have you) terms.

In a way, WinFS would finally unchain us from the goofy, outmoded constraints put in place years ago by DOS. No longer would it be up to you and me to remember whether we saved a restaurant review in e-mail, a Word document or a bookmarked Web page. The software would do the heavy lifting for us.

As strange as it sounds, we've been waiting for more than 12 years for this simple concept to manifest itself in a new version of Windows, and it looks like we'll be waiting for at least a few more. You've probably heard, or even remember, that Microsoft tried--and ultimately failed--to bring the "unified storage" concept to life with an aborted Windows NT update code-named Cairo.

You might not remember the details of the Cairo saga, or recognize the parallels with Microsoft's current dilemma with Longhorn. Here's a quick recap:

Back in 1992, Jim Allchin, the Microsoft executive in charge of Windows development, announces that the company is beginning work on Cairo, which will include a new "object file system" for storing document files, spreadsheets, multimedia files and other information in a unified way. The goal is to enable searching not only by file name, but by also file content. Cairo is due in 1994.

In 1994, after two years of relentless hype, Microsoft announces that Cairo's debut will "slip" into 1996. The company later moves Cairo's debut to 1997. Then catching up to Netscape and the Internet boom becomes Microsoft's top priority, and work on Cairo is "reassessed." In 1996, Bill Gates says Cairo's storage system is a vision, not a product, leading none other than Rick Sherlund--the Goldman Sachs analyst who helped Microsoft to go public--to say that Cairo has "lost its definition."

Fast-forward to 2002: Allchin tells CNET News.com editors that the old unified file system concept is alive and well and is planned as a major component of the next version of Windows, code-named Longhorn. Allchin doesn't set a date for Longhorn's debut, but analysts speculate it will arrive in 2004.

In February 2004, Allchin says that some of the programmers working on Longhorn have been reassigned to work on Windows XP Service Pack 2, a much-needed security revamp, and that as a result, some of Longhorn's features have been cut. In May, a Microsoft executive says WinFS will be included in Longhorn, but that "some of the functionality of WinFS and some of the scenarios may be limited in terms of what it can do."

Finally, last month, Microsoft said WinFS will not be included in Longhorn at all, but will instead ship as a test release sometime in 2006. Gates--recognized as the driving force behind the universal storage idea--then updates News.com on the technology. WinFs--"I'd be the first to say--is very ambitious," he says.

Why can't Microsoft--the world's largest software company, with thousands of talented programmers and billions of dollars in the bank--bring the unified storage concept to life?

There are lots of reasons. Legacy support, for starters. WinFS needs to work flawlessly with previous generations of Microsoft's applications, custom-built software and third-party tools or it will never get off the ground. That's hard, time-consuming work.

Then there's the problem of project dependencies. Microsoft is an enormous company with many simultaneous and interlocked projects. When one development team sneezes, another catches a cold. The team building the next version of SQL Server, code-named Yukon, ran into some development snags earlier this year. Some of the WinFS work was built on Yukon technologies.

Microsoft was also simultaneously building new development tools and redesigning the Windows programming interface. That's in addition to finishing Windows XP SP2 and an update to Windows Server 2003. Something had to give.

So what does that leave in Longhorn? Plenty. A new graphics system called Avalon, a new communications and Web services subsystem called Indigo, and plenty of other new features that we learn about almost daily.

But the grand vision of a unified and simplified search tool for Windows once again has been delayed.

There is one major difference this time around. In the 1990s, none of Microsoft's competitors could really take advantage of Cairo's downfall. Not so this time.

Next year, Apple plans to launch new search tools as part of the Tiger release of OS X. And the Linux camp isn't far behind: Novell says it's retooling its iFolder software to give its SuSE Linux unified search capabilities.

The capabilities of these various schemes vary, and details are sketchy. But clearly, Microsoft--which popularized the idea of unified search--will likely lose the race to market with actual product. Now the challenge for the company is to convince customers that Longhorn--or whatever you think we should call it--is still worth waiting for.