I've tried many ways to manage tasks, with varying degrees of success, at home and at work. I've never settled on any method for a long time—I don't know, maybe all methods tried have a flaw (overcomplication being the most common one), or there's another method over the next hill just waiting to be found so the current one can be abandoned.
I was drifting a little this morning and thought about what a task is, what it means—working backwards from the things I actually do, and breaking them apart into smaller and smaller bits. If you plan a project, it's some form of putting all of these bits together into larger and larger bits, and—hey presto—that collection of bits is the plan, more or less. So it seemed logical that if you don't have a good feel for what those bits are then your plan will always be wrong.
Break a task down and you get task molecules and task atoms that make up the larger task. Now let's overburden the analogy: what happens when you look inside the task atoms to prepare to break them down further? Subatomic tasks. Can you count those? Sort of—but not by observing them, maybe through statistical measurement of where the details of those subatomic tasks are expected to be.
It seems a reasonable, if overwrought, analogy—if you divide the work down into smaller and smaller bits, there is a threshold where the divisions cease to be useful. There is no thing anymore, but a cloud or probability of a thing. And what have you for all the trouble of that division? Time and effort spent—probably in unplanned tasks—yet no certainty, likely worse off than having found an appropriate level to understand the work and stopping there. A voice rumbles from the horizon: leave well enough alone.