So why does it often seem that plus ca change, plus c'est la meme chose? Outside the electronics world, seeing the Yankees once again dominate baseball reminds us why "deja vu all over again" has become a popular cliche. Electronics too has recurring issues, even though things have changed phenomenally since the pre-PC era.

One big bombshell of a couple of decades ago came when Hewlett-Packard announced that chip quality wasn't really up to snuff. That's starting to become an issue again, though the system-on-chip and huge ICs being built today are vastly different from their counterparts of the early 1980s.

It seems almost unreal that a chip could be the root of problems in new systems. After all, device designers have never had to build in something like control-alt-delete. It was quite a surprise to see just how much energy companies like Motorola, Sun, HP and Intel, and researchers at the University of Illinois, are putting into efforts to figure out how to avoid transients and other problems in the deep-submicron devices and customized SoC designs that are gaining acceptance throughout the industry.

For engineers, this could go beyond technology, and have profound psychological effects. For some of them, it may be crushing to learn they can't simply blame Microsoft when a PC or server goes down. The fault could lie in the silicon.

Chips are late; recalls occur. It's a bit like the late '70s-early '80s, the point when HP sounded its call for chip reliability. Everything old is new again.

But the problem is in figuring out which old trend will return, and how it will have mutated over time. A number of technologies could be the next ones to resurface. And a number of old trends that have interesting possibilities probably won't return, at least in a form that's close enough to bear comparison.

Will the parallel-processing techniques that flopped in supercomputers come back afresh? If so, it's time to get into the field of parallel software. Will home networks finally live up to the potential that's been promised for a decade? If so, the people who pick the right technology will make a lot money. The rest will have to keep searching for the next trend, be it old or new.

Sometimes, it's wise to look backward instead of forward when you're trying to figure out where things are headed. If nothing else, the root causes of previous successes or failures will help in today's evaluations.

The engineers and managers who embrace a trend have some control over whether their rejuvenation effort will bear fruit. I'm not as lucky. Back before the days of Ruth and Gehrig, another team dominated baseball. Here in the Windy City, we're still waiting for the Cubs to reprise that role.