Tyler Reddy

> Sorry I've left this for so long: as I understand it, the goal is to mark docstrings as executed if doctest executes them? Yes, that would be cool, but...

I think it is pretty common for us to have to use a build from a specific git commit hash rather than right at a release point because of sensitivity...

This seems hard to judge, on the surface I think my initial view is similar to Oli's. Is progress on the CZI objectives impeded by this, or it is just...

Did I see on Discord that some folks at a conference wanted multi-precision? Could someone summarize that here? For the specific linked PR gh-3544, that subset of changes looks like...

So, I think I'm a bit more positive about changes like this in isolation if the tests pass and we're backwards compatible, at least initially. Not sure we should jump...

I *think* we're agreeing that flexibility is good and single-precision default is good (me for backward compatibility reasons, and you for performance), IIUC.

I'm also supposing that if we're mostly single-precision internally, but sometimes returning double, that it isn't too much of a "break" to start returning in the true precision used for...

> Is the best strategy perhaps to work on type flexibility around the edges Might be, at least for me the idea of starting with the more straightforward changes sounds...

I'd suggest only applying the CI linting to the `diff` if we do this, rather than driving the full code base through a large amount of churn. SciPy is working...

There may not be any major drawback, I just mean it could be sensible to wait to see what the rest of the ecosystem decides to do (if anything) about...