> Rather than being arbitrary, the motivation here is to limit amount
> of the import state that is specific to CPython.

What is gained by doing that?

> Finally, modules (the biggie) is accessible as sys.modules.
> importlib uses sys.modules.  The C import implementation used
> interp->modules directly. [1]  Most (all) of that usage is going
> away.  Again, _if_ that is the case, why keep it around?

I think each of them should be considered on a case-by-case basis.
Feel free to submit a patch that eliminates ->modules. People may
find reasons not to do so when they see an actual patch.

I suggest to close this issue, and encourage people who have the
desire to eliminate certain state as individual patches.

> Just to be clear, I do _not_ want to make changes willy-nilly.  (I've
> even grown more conservative in what discussion topics I bring up.)
> This issue has no urgency attached to it, in my mind.  It is the
> result of an actionable conversation that I didn't want to lose track
> of.

Then I think it doesn't belong in this bug tracker. I have five or
ten "grand plans" of things that should change in Python at some point;
putting them into a bug tracker is only confusing people, though, since
no implementation might be coming forward (for some of the things, I
have been pondering for the last eight years). For many of the things,
I ended up writing PEPs since they were significant changes.

So if this is one of your grand plans, feel free to mention it on
python-dev. Putting it on the bug tracker asks for specific action.
If I had to act on this issue, I'd outright reject it.