Complexity

Originally Posted November 5, 2008

My wife and I each have our own alarm clock. Her's is a traditional analog type with the little pointer that indicates what time the alarm will go off. Mine is a fancy digital clock radio that has two different alarm times and lets you choose either an alarm tone or the radio.

If she is in a hurry to set the alarm, it can be as much as 15 minutes early or late. If I am in a hurry, I am likely to get a very precise 12 hour error.

“Fools ignore complexity. Pragmatists suffer it. Some can avoid it. Geniuses remove it.” - Alan Perlis, Computer Scientist

There is no question that my alarm is the more complex of the two, so I suppose it is no surprise that it would be more prone to operator error. This relationship is found in our cockpits as well. Avionics have made a huge jump in capability, and complexity, over the last decade. I often instruct in aircraft equipped with the Garmin G1000 and Avidyne Entegra glass cockpit systems. While I feel I am very proficient in their use, I still learn new things about them from time to time. And my customers have to work hard to take advantage of everything these systems can do.

Researchers tell us that there are two different types of complexity. The obvious one is “detail complexity”, which relates to equipment or situations that involve many facts or elements. My digital alarm is a perfect example of this. It has lots of knobs, buttons, and indicators.

In aviation, detail complexity shows up in avionics systems (particularly GPS navigators), airspace, ATC procedures, and flight and security regulations. The trend is toward increasing complexity.

The other variety of complexity - called “dynamic complexity” - involves situations where cause and effect are separated by time or space. In the fast moving world of aviation, this kind of complexity is increasingly apparent. Just like a error setting my alarm (like missing the small AM/PM indicator) leads to a surprise 12 hours later, we see cases where the result of something we do in the cockpit may not show up until some time later.

Dynamic complexity is particularly troubling in aviation. We see this in IFR flying, where an error copying and comprehending a clearance may not manifest itself until later when we fail to follow the expected path. Autopilots can be incorrectly programmed and not behave as intended. Faulty entry of a flight plan into a GPS navigator can lead to inadvertent airspace incursions hours later.

The bottom line is in managing this complexity to ensure it doesn’t lead to an unsafe situation. The key to this is knowledge of the systems you are using, proficiency in their use, and keeping workload at a reasonable level so you can monitor the complex systems.

Fly Safe!

Neil