The fallacies of todos

Searching a code base for “TODO:”s is a quick and easy way to measure technical debt. Reading through them often feels like going to a museum, with lots of names from people that have left the codebase a long time ago. Github currently lists more than 51 million TODOs.

I’ve come to realize that TODOs rarely get accomplished. The ones that are really critical usually get done while the rest gathers dust and becomes unnecessary technical debt.

I’m in favor of not using TODOs altogether. Leaving TODOs everywhere is like publishing incomplete code. It’s just as ennoying as someone starting a sentence and never

dark
sans