I'm a person, not an RTOS

Like many of you, I often multitask: check e-mail, talk on the phone, read a magazine, and eat, all more or less simultaneously. This ability to do so many things at once makes me feel so efficient and productive that I impress even myself. But I am deluding myself. A few days ago, I was trying to remember a transient thought I had about a project, had a fleeting idea that I was going to mention to someone in an e-mail, and was trying to grasp the noteworthy point of the article in front of me, all at the same time. If I were a processor, I would describe my symptoms as buffer overflow or interrupt overload.

Then I realized my misconception. Although some researchers and scientists are trying to make computers and software more like people, most of us are moving in the other direction, trying to emulate the capabilities of physical processors and their real-time operating systems. We assume that we can be like those processors and operating systems with which we work so easily, but we can't. They can do their context switching with low latency, and, therefore, so should we. They can serve prioritized interrupts; thus, we should do the same.

In reality, we cannot easily perform such emulation. In fact, we're lousy at it, despite what we tell ourselves. A recently published article reinforced this point (Reference 1). It explains how, in some carefully controlled performance and efficiency tests, the subjects' attention and coherence dropped rapidly as they tried to multitask. Contrary to what many people would expect, the subjects accomplished less when they did several things at once, especially when some of the tasks were at least moderately difficult.

Another basic engineering rule also doesn't apply to us, although we think it does. Nyquist's theorem states that if you sample quickly enough, you can capture all the information a signal conveys. But this fact does not mean that you can flit from topic to topic, meeting to meeting, or task to task and still have the concentration and insight you need. Unfortunately, some overworked engineers and their managers think that you can.

If you review what advanced researchers know about how the brain works, you'll learn that the short answer is "not much." Despite years of serious investigation, testing of individuals, scans of brain activity, and more, researchers know little about how the brain operates: how it stores and recalls basic data, still and moving images, and sounds; implements short- and long-term memory functions; searches for data in foreground and background modes; performs pattern-recognition; recognizes voices and pictures; understands and translates languages and idioms; makes judgements; deals with new situations and problems; and takes those creative leaps that even the best supercomputers, such as chess champion Deep Blue, cannot make, despite their incredible computational prowess. It's foolish to think that you can become like the processors and operating systems you know and use when the underlying systems differ so radically.

How does this concept apply to engineering and design? The lesson is to focus when you are doing something that counts in your design effort and concentrate on the priority. Don't let the sirens of multitasking, context-switching, or Nyquist sampling lure you away. When you are performing design-tolerance analysis, posing what-if scenarios, developing product documentation, writing a critical report, or performing any other project tasks, your quality and productivity will improve when you work on one important thing at a time—despite how easy it seems and how pressured you feel to do many things at once.

Contact me at bschweber@edn.com.