Ask HN: How to decide when to do something?

You're writing up some instructions you are going to use later and the word 'positivity' is marked with a red line by your editor indicating it's not a word. Do you go off to google to learn more about the definition of the word and related words ?

You're adding a new feature that your business partner and a customer have asked to be completed for review/testing by tomorrow. As you're coding you notice that a library class that's used in various places is miss-named. Do you refactor it or do you add it to a list of stuff to do later ? I worry if I don't do it now it might never get done, yet that one little side street of improvement can lead to noticing another 2 things that need to be improved, which in turn ...

I find it demotivating to see code that I know should be improved but if I jump into refactoring working code I worry about breaking functionality and not progressing with features fast enough.

All feedback appreciated.