Change is in the air

The irony is that when it comes to the prediction racket, Sun Microsystems' CEO will never be confused with the Amazing Kreskin. For years, McNealy wrongly argued that the era of client-based computing was about to give way to a new approach where--borrowing upon the company's tag line--the network was the computer.

Still, you can only be wrong for so long. McNealy's prediction is coming true, though not because companies are suddenly mad about network computing architectures or software components produced at Sun.

The real game changer is broadband. More than any development since the late 1990s, the proliferation of inexpensive high-speed Internet connections represents a profound structural shift in the computer business. Among other things, it promises to usher in an era in which network access becomes the equivalent of a global utility.

Truth be told, the broadband rollout in this country has proceeded at a snail's pace. While nations like South Korea and Japan surged ahead, the U.S. dithered. The good news is that America's high-tech industry has hurdled most of the bureaucratic obstacles put in its way by a distracted federal government.

When this build-out reaches a tipping point, what happens then? That remains an open debate, but one point already is beyond contestation: Most of the important software development work being done these days is almost exclusively focused around the network and network services.

I'm not exaggerating. When it comes to the desktop, the only real innovation remains in the realm of so-called helper apps that support media or other communication applications. (For example, you've got things like iTunes, software media players, instant messaging and various desktop applets--Sidebar, Google Desktop. Big deal.)

McNealy's peripatetic No. 2, Jonathan Schwartz, believes he knows where all this is leading. Last week, he told an audience at the American India Foundation that the PC is on a collision course with irrelevance. If Schwartz is right--and I think he is--then the platform of choice inevitably will move to the network. Talk about a pregnant scenario!

In this gauzy future, some industry veterans talk about a vast Internet cloud in the sky acting as a virtual computer with limitless storage. For many of the mega-companies that have long dominated the computer industry, that's not necessarily great news.

For instance, a desktop computer doesn't figure to dominate a future where people can use a multiplicity of mobile devices to tap into that cloud. In other words, good-bye PCs, hello Dynabooks (or something quite similar). Is it any wonder that IBM sold off its PC business to China's Lenovo?

The people responsible for shepherding Windows also have to wonder how their cash cow fits into a world where operating systems are losing relevance. Maybe the same will also apply to middleware suppliers.

"It's the inevitable path of all software engineering," said Grady Booch, the chief scientist of IBM's Rational Software business who predicts the emergence of organizations banding around architectural standards that live above the middleware level. "It's operating systems today, middleware tomorrow," he said.

McNealy and Larry Ellison--he of the Network Computer--first put the issue on the agenda way back when. The computer industry's been debating the timetable ever since. Turns out they weren't wrong, they were just too early. Broadband availability was the missing piece of the puzzle.