[RFC] How much of the standard library should be supported

Summary

Given the recent discussions about the size of CPython's standard library, and proposals to drop a few modules, splitting the core standard library from the extra batteries, and so on, perhaps would it be a good idea to define the extent of the standard library to be implemented/provided in rustpython.

Detailed Explanation

Issue #8 calls for the expansion of the standard library, but I feel that a blind call might not be the best idea. Instead, a more directed call for help would provide more cohesiveness of the work.

However, more importantly, discussions about the size of CPython have been surfacing here, a view of the future here, suggestion on how to approach a kernel-python PEP-594 to remove modules (https://www.python.org/dev/peps/pep-0594/).

The question is, where does rustpython want to be with this current context.

Drawbacks, Rationale, and Alternatives

The main rationale would be to allow rustpython to be an example of how a minimal modern Python interpreter could be when given the freedom from the ground up. This would also reduce the maintenance burden and the required work to produce a working interpreter by focusing effort.

This issue is somewhat linked in spirit with issue #1192 and #1125.

Unresolved Questions

How to define what would be the minimal standard library to be shipped with rustpython?

Start by reviewing issue #8 and the related issues #1192 and #1125, then read the linked discussions about CPython's standard library and PEP 594. Done means reaching and documenting a project decision about the minimal standard library RustPython should provide.