A recent issue of Time magazine analyzed the multi-tasking behavior of modern children. The upshot of the cover story was that our interpretation of multi-tasking was flawed for the simple reason that children do not explicitly do more than one thing at the same time, as they would were they simultaneously walking and talking. Instead, they concurrently engage in many things, like talking on the phone, typing instant messages, doing their email, watching TV, and working on homework. Such behavior is clearly distinct from true multi-tasking; it is more the continuous cycling between tasks. Taken purely as a matter of logic, it should make no difference were the child to take up one task, finish it, and then proceed to the next one. This sequential completion should not take more or less time than when the tasks are all worked on a little bit at a time, but never simultaneously.
However, it is still useful to distinguish pure sequential processing of tasks from the incessant cycling between tasks, which may be better denoted as time-sharing. As with a vacation time share between three people, it does not really make a time difference were the time per person to be used in one block of four months, or four separate blocks of one month each separated by a periodic hiatus.
Of course, there is one small fallacy here – in that time-sharing may be more efficient than sequential processing when the former allows downtime periods in sequential processing to be filled with another task. For example, we may be working on proving a theorem, having resolved not to undertake any other task unless the one at hand is completed. However, if one were to get stuck in the endeavor of proof, then it may be efficient to step away from the task and use the time getting some email out of the way. Time-sharing can be more efficient than sequential processing since it is the same as the latter without the constraint that any one given task be allocated to a single block of time. Time-sharing may be less efficient when there are switching costs that are incurred when moving from one task to the next, as we cycle through them all.
Whichever mode is efficient, there is one strikingly interesting parallel with modern computers. microprocessors are designed to time share, switching between processes on the computer at the speed of its internal clock. This has made multi-tasking on the computer very efficient. It is indeed ironic that so much of science fiction has been devoted to creating machines that behave like humans, when in fact, if anything, today we are closer to microprocessors in the way we handle tasks. Hence, a better view is to see that humans behave much more like microprocessors than vice versa. Given the complexity of human behavior, and the simplicity of the operation of the microprocessor, this is hardly surprising. Man to machine convergence is probably happening faster than we think.