What the metronome teaches

Everyone who has been made to practise with a metronome has had the same experience: the passage you thought was fine turns out to be fine only because you were quietly taking more time at the hard bit. The metronome does not make you play in time. It makes the borrowing visible.

I keep finding the software version of this. Estimation is the obvious one — timeboxing does not make work fit the box, it makes visible which work never fitted. But the better example is a test suite with a hard timeout. You do not learn that the suite is fast. You learn exactly which test has been quietly taking four seconds since March.

The instinct to turn it off

The immediate reaction to a metronome is that it is wrong. It feels rigid; the music obviously wants to breathe here. Sometimes that is true — rubato is real, and mechanical playing is its own failure. But the honest version of that argument is only available after you can play it in time. Before that, “it wants to breathe here” and “I cannot play this bit” produce identical audio.

The same test applies to a lint rule you want to disable, or a budget you want to raise. Can you meet it, and are choosing not to? Or can you not meet it? Those feel identical from the inside and they are not the same situation.

Then you turn it off

The point of the metronome is not to play with one forever. It is to know where you were borrowing, so that when you take time you are spending it rather than losing it.

That is the whole of it, really. Constraints are not there to be obeyed. They are there to tell you what you are doing.