Hi all, One issue we have in NumPy is that there is fairly common confusion about operations that mix NumPy arrays and scalars (Python scalars or NumPy scalars). This also adds trickier logic to NumPy it may help both users and the NumPy code complexity if we could move forward with a simplification. There are however, different ways to approach the problem and possibly different ideas altogether about what the preferred result is, so I made a poll here: https://discuss.scientific-python.org/t/poll-future-numpy-behavior-when-mixi... to get some input. There are no wrong answers :), mainly I would like to have a better idea about expectations are. As well as your opinion on the feasibility of moving things forward. The poll tries to go from "easier" to "trickier", so feel free to stop on the way. Please also feel free to use the this or the forum to discuss things further. Cheers, Sebastian