Every chip, relay, servo and processor must prove beyond a glimmer of doubt t

Airplanes vs. Computers

The differences are large.

When it comes to reliability, there's a fundamental difference between the aviation field and just about every other walk of life. Nowhere is the chasm more visible than when you compare aviation's safety culture with what goes on in the computer industry. I know. I just spent 10 days in computer hell, and partly to retain my ebbing sanity I'd close my eyes and count the ways this could never happen with an airplane.

My troubles started when my year-and-a-half-old Dell laptop went into convulsions. The next week and a half involved making multiple visits to the diagnostic center; researching and buying a new computer; having my backup data transferred to the new machine; and then doing it all over again four days later when the new computer also went catatonic and I had to buy a third laptop.

Each step of the way, the technicians I dealt with assumed the usual computer-issue posture-they shrugged. "Yeah, those hard drives crash all the time," they said. "I see 'em defective like this out of the box every day."

Imagine if it was my life—not just my livelihood—that was on the line every time I booted up.

Now contrast my computer experience (and probably yours) with the level of reliability that aircraft and their systems need to attain to achieve certification. For example, business jet manufacturers routinely cold-soak test their models as part of the certification process. Recently, one "luxury" jet was submitted to far-from-luxurious conditions during certification trials: it was cold-soak tested at -33 degrees Celsius in the wastelands of northern Canada. After the ground crew shoveled the jet out from under a blizzard's worth of snowdrifts, not only did it have to start up and take off unscathed but everything had to work. Everything. Even the processors driving the cabin entertainment system.

Modern aircraft and their innards are torture-tested far beyond Marquis de Sade's wildest dreams. Every chip, relay, servo, and processor must prove beyond a glimmer of doubt that it will be as reliable as the bolts that hold the wing in place.

As reassuring as that might be, that culture clash is one reason it has taken decades to upgrade the air-traffic-control system. As each proposed computer and software upgrade was presented, it went into a circuitous trial and approval pipeline overseen by aviation-safety zealots. By the time all the testing and tweaking was done (and the computers rendered as fail-safe as those on board an airplane), the rest of the information-age industry had left these processors' specs in the dust. Some congressman's aide would look at the fine print in the final proposal and snort, "My kid's iPod has more RAM than this!" And it was back to square one.

The good news is that the upcoming Next Generation (NextGen for short) ATC system has finally built in some techno-expansion capability to accommodate advances in computer performance. And further, more of the technology burden for controlling traffic along the airways is being transferred from ground-based radar to navigation systems on board the aircraft. The system will be far simpler, more efficient, logical, and ultimately safer. But only if aviation continues to defy the culture of expected failure that dominates mainstream computer development.

So far it's doing that quite well. One more example: the Bendix/King KLN88 loran-C receiver/processor that was installed in my airplane in 1988. Not only has it operated flawlessly ever since but its accuracy has mirrored that of the most modern GPS receivers I've stacked up against it. Only the decommissioning of loran-C ground transmitters by the U.S. Coast Guard last month could render my unit obsolete.

Last week I tried to describe my loran-C to a technician who was attempting to resuscitate one of my laptops. It was as if I was speaking a different language.
And, in a way, I was.

THANK YOU TO OUR BJTONLINE SPONSORS