[gknoy]

I agree. I use TODO indicators as a code smell for things I forgot to add before making my pull requests, but I also see them as valuable reminders for future needs.

Sometimes there's a quick-but-less-generalized way of solving something, and you know there's a more elegant way of doing it, but for now you need to get something working/fixed and don't want to make sweeping changes.

Your immediate change might be to add a way for users to see that X is finished. A useful TODO message would be something which explains what you're doing

  # Tell users when their Solve-My-Foo is done
  # (Currently done in an ad-hoc way)
  # TODO: Add a notification infrastructure so that 
  #    arbitrary components of the apps can present
  #    notifications of Important Things
  #    e.g.:
  #    - celery job is done
  #    - a thing you've subscribed to has changed
  #    - a report/download/etc is ready to be retrieved

Building such a system might be overkill right now, if this part of your system is the only one that uses it, but a TODO message helps me see (the next time I make changes to it, or try to extend it, or use it as a template for reimplementation elsewhere) that there's more than one place we need this feature, and so it might make sense to plan/do those more comprehensive changes.