Stephan Hoyer

> * What is the design pattern to work around `np.asarray` in downstream libraries? I opened https://github.com/data-apis/array-api/issues/122 for this one before I saw this issue :)

6. How should libraries check array types, e.g., what is the equivalent of `np.ndarray` for use in `isinstance` or type annotations? Perhaps we should include the equivalent of an `Array`...

> @asmeurer Re: integer array indexing. As mentioned during the previous call (03/06/2021), similar to boolean array indexing, could support a limited form of integer array indexing, where the integer...

It's definitely possible (but not necessarily easy) to rewrite every call to `np.take` in terms of `__getitem__` with integer arrays. For a library like Xarray, support for all integer indexing...

What is the concern with supporting integer array indexing in `__setitem__`? Just the fact that it may not be implemented in otherwise compliant array libraries?

> That, and also that it's non-deterministic when indices are not unique, as noted in the PyTorch and TF docs on `scatter`/`scatter_nd`. I think we could probably safely leave this...

I wonder if the right way to handle this for now is to add a low level API for indicating costs, either in terms of a multiple of the original...

I noticed that `IPCompleter.limit_to__all__` is listed as deprecated in the docs now: http://ipython.readthedocs.io/en/stable/config/options/terminal.html Is there another supported way to do this? (For what it's worth, I have also found this...

TensorFlow goes to [some heroics](https://github.com/tensorflow/tensorflow/blob/e9de087fa7f59c39bbe12ac2c83c5547c83f746c/tensorflow/python/util/all_util.py) to delete undocumented variables from modules to work around this same issue.

> If ipython finds packages through dir(), Python 3.7 will allow overriding dir for a module (https://www.python.org/dev/peps/pep-0562/), thus giving back power to developers to control what is visible through tab-completion....