I use comments to document assumptions that are likely to be wrong, either now as I write it, or later when someone (probably myself) changes it. It is absolut

I use comments to document assumptions that are likely to be wrong, either now as I write it, or later when someone (probably myself) changes it. It is absolutely useful to do, and really not too difficult. Languages that allow for more formal assumption-checking (especially before runtime) are even better, but comments have an additional benefit of being understood by a human directly. I wish languages with static analysis could somehow allow authors to encode human-friendly/sematic errors that you often see as runtime exceptions into the static analysis itself...