Columns

From where I sit: XML is more important than Linux

• By David Chappell
• May 30, 2001

Not too long ago, I was en route from Minneapolis to San Francisco when somewhere over Nevada the pilot made an unexpected announcement. "We seem to have a warning light on here in the cockpit," he said, "indicating a fire in the aft cargo compartment." The plane then made the quickest and most direct landing I've ever experienced. After a moment of sitting on the tarmac, the pilot made this announcement. "Easy Victor," he said. "Easy Victor."

The flight attendants immediately made a transformation. No longer were they their usual calm and gentle selves; they were now Storm Troopers From Hell. "Bend over! Stay down!," they shouted at passengers, and we all meekly hunkered down in our seats.

I had always harbored a desire to take part in an emergency airplane evacuation, and I was now thrilled to have my chance. To my great delight, the attendants popped open the plane's doors and blew the escape slides. Getting out of the plane was fun - the inflated rubber slide felt like an attraction at an amusement park. The aircraft was empty in no time, and a crew of firefighters in shiny, high-tech suits promptly arrived. The whole event was thrilling.

It was also somewhat anti-climactic. The entire exercise was caused by a broken light in the cockpit. There was no fire.

I think there's a lesson here for people who work in technology, one that we all need to keep in mind. It is that we must be able to distinguish between a broken light and a real fire. Broken lights are important, and we should take them seriously. But they never have the far-reaching impact of real fires.

The best example today of a broken light is Linux. Like the light in my airplane's cockpit, Linux is worth some attention, and a good number of people are using it. But ultimately, it's just another Unix. True, it's a very inexpensive Unix, and so it's bringing changes that no previous version of this OS accomplished. But despite what Linux cult members believe, this OS doesn't fundamentally shift the technology landscape - it just repackages technology that already exists.

Just as important, it's not a serious threat to Microsoft. The reason for this is simple. When choosing a server OS, enterprises typically begin by making the Unix-or-Windows NT decision. This choice is commonly driven by reliability and scalability requirements, current vendor relationships, availability of knowledgeable administrative staff and other concerns. If Unix is chosen at this stage, the next step is to select a specific version: Solaris, HP-UX, Linux or something else. Given this, Linux winds up being mostly a serious threat to other Unix vendors. While some firms do make a direct Linux-or-Windows NT decision, in my experience this is not that common. Ultimately, Linux is more of a threat to Sun than it is to Microsoft, no matter what Steve Balmer wants the Department of Justice to believe.

This is not to say that Linux is unimportant, any more than the broken light on my airplane was unimportant. But broken lights have far less impact than real fires. And, in today's world of enterprise technology, the real fire is the eXtensible Markup Language (XML). Like a fire, XML compels our attention - it will change the landscape.

Unlike Linux, which is essentially an updated version of an old technology, XML offers something that's truly new: a widely accepted scheme for describing information. XML was first applied to describe data transferred across the Web, but its use has now expanded. For example, XML is a required technology for DBMSs. Oracle 8i, IBM's DB2 and even Microsoft's next release of SQL Server all provide support for translating data into and out of XML-defined formats. XML is also likely to be used to define services available on the Web and elsewhere, probably pushing out the traditional IDLs defined by CORBA, COM and other technologies. Even new protocols, such as Microsoft's Simple Object Access Protocol (SOAP) are defined using XML. XML is a real fire - its impact will be huge.

Distinguishing between broken lights and real fires isn't easy, and in the short run, we may react the same way to both. I'm glad my pilot took that plane down so directly, even though it ultimately proved to be unnecessary. But if he'd figured out that it was just a broken light, he could have avoided a mass evacuation. In charting our technology future, it's important to watch for this distinction.