Sounds like they’re still very relevant and very important. Python isn’t a language I’ve used a lot but I’m still surprised I’ve never heard about docstrings till this tutorial. Thanks for the info.

Yeah, they’re definitely still the standard. They’re not really replaced by comments, comments are more for explaining why bits of code are the way they are. Docstrings are kind of like comments for functions/classes/etc that Python knows how to handle specially. The interpreter will parse the docstrings and make help text out of them available to the help builtin function

Yes, use them. One big advantage is if you hover something in an IDE it will show you the docstring.

If you’re writing Python you should be using Pylint (or Ruff) and it has a lint to ensure you write them.

The exception I usually make is for class member variables because it’s super weird that the docstring comes after the variable. I think that’s very confusing for people reading the code so I normally just use comments instead.

At my work we use linters that refuse code with missing or bad docstrings.

Yup, use docstrings and be descriptive. Learn by reading docstrings in other codebases. (Django’s codebase is always an excellent example of how to do Python, including how to do docstrings.) a) Anyone who sees your code will thank you and b) if you can’t explain it in human language, you probably don’t understand it well enough to implement it well.

Yes they are sometimes used, as is doctest, but Python’s built in help system isn’t as good as it could be, so the docstrings aren’t that useful. Most annoyingly, for built in library classes, help(whatever) spews machine generated prototypes at you before you get any actual documentation.

I generally like to write some kind of explanatory text along with any nontrivial function that I write, but I’m not very consistent about doing this as a doctring vs as a code comment. I do use type annotations heavily nowadays.

yes