Did they control for that people don't write tests and test assertions in notebooks?
Did they control for that people tend to not maintain expository code in notebooks in the same way that they maintain modules?
Are most .ipynb demo notebooks?
What percentage of users use a notebook-first workflow; such that code written in an .ipynb is auto-exported to .py modules?
ipytest is one way to run pytest in a notebook. assert statements in named test functions are another way to run tests in a notebook.
How much of the difference in code quality between notebooks and modules is due to the tool and how much is due to the type of code that people tend to write in notebooks?
Did they control for that people don't write tests and test assertions in notebooks?
Did they control for that people tend to not maintain expository code in notebooks in the same way that they maintain modules?
Are most .ipynb demo notebooks?
What percentage of users use a notebook-first workflow; such that code written in an .ipynb is auto-exported to .py modules?
ipytest is one way to run pytest in a notebook. assert statements in named test functions are another way to run tests in a notebook.
How much of the difference in code quality between notebooks and modules is due to the tool and how much is due to the type of code that people tend to write in notebooks?