David Brochart

@minrk it would be great if you could be the shepherd for this JEP.

I started an [implementation of sub-shells in ipykernel](https://github.com/ipython/ipykernel/pull/1062), that can be useful to validate the kernel protocol changes required by this JEP.

Although tests fail in https://github.com/ipython/ipykernel/pull/1062, it is functional (you can try it e.g. with JupyterLab) and the [sub-shell test](https://github.com/ipython/ipykernel/blob/83e35021fda1693ecb7b4e0fb747a70c1c22263f/ipykernel/tests/test_subshell.py) shows that sub-shells run in parallel.

I opened a [PR in JupyterLab](https://github.com/jupyterlab/jupyterlab/pull/13729) that uses the [ipykernel PR](https://github.com/ipython/ipykernel/pull/1062) to implement "sub-consoles". Together, these PRs can serve as a proof-of-concept for this JEP.

> This might be a first step to enable a highly-requested feature: built-in variable explorer (e.g. [jupyterlab/jupyterlab#443](https://github.com/jupyterlab/jupyterlab/issues/443)) which would not depend on the debugger being active (or supported by the...

I made changes according to the comments in https://github.com/jupyter/enhancement-proposals/pull/91/commits/d389802464d170b6cbac26f17c8d5e31d3d8bc16. I removed the [points of discussion](https://github.com/jupyter/enhancement-proposals/blob/ac519a06cbe14e44c7aa974f0bf07e431f79d064/kernel-subshells/kernel-subshells.md#points-of-discussion) because I think they don't really make sense: the operations I mentioned (create sub-shell, delete...

I think we can discuss both approaches in this PR. To help us decide which one we will choose, it could be useful to list their differences. Let's start with...

Thanks @kevin-bates, let's call these two options "kernel sub-shells" and "dependent kernels" as you suggested, since they reflect well the fact that for the former we really expose sub-shells as...

OK so not very real-time friendly, but that could be a quick path.

Good to know, would you be open to a PR?