Reason 1 (Your pipeline should be reproducible) for avoiding jupyter doesn't make any sense. That's the whole point of jupyter. Out of order execution can happe

Reason 1 (Your pipeline should be reproducible) for avoiding jupyter doesn't make any sense. That's the whole point of jupyter. Out of order execution can happen, but you can just as easily restart the kernel and run all... I can only imagine this is a problem for someone who doesn't understand the tool they're using. Reasons 2 and 3 for avoiding jupyter are more justified but easy enough to work around between jupytext and papermill.