This old post from Joel is just a fancy way of saying what psychologists studying task switching have found: it’s better to do things in batches. This is also something that GTD, Do It Tomorrow, and other productivity methods attest. The whole point of tagging things with contexts in GTD is to be able to do all related things at once (when you are in a certain context).
But wow much better is batching? Well, here’s where things gets blurry. Most task switching experiments are done with extremely simple tasks, like pressing a key when a triangle is red, and another when it’s blue vs. a similarly abstract task.
So we know surprisingly little about what tasks gain the most from being done in batches. But this reflects a more profound lack on our theory: what is a task? How similar are two tasks? For example, how similar is driving a car to driving a bike?
A dirty operational definition would be that two tasks are more similar the more they gain when done in the same batch. However, as far as I know nobody has tested this idea. And there’s a big ‘but’: sometimes similar tasks interfere with each other. Try singing a song and writing a different song on a pentagram. Or brushing your teeth and combing at the same time. So this operational definition doesn’t seem to work well.
Why is studying multitasking important? Well, if you have a browser open while you work, you know the answer already.
Do you know of any interesting multitasking studies that use realistic materials?