Continue supporting JupyterLab 3?

[timkpaine]

xref: #187

[dhirschfeld]

I'd vote for no because maintainer time is a scarce resource which could be put to more productive uses than maintaining backward compatibility.

In general, I think there has to be some onus on the consumers of free software to keep their stacks up to date to minimise the burden on maintainers.

That will no doubt impact some users unable to upgrade but they can usually continue to use the older version until they can update their dependencies to be compatible with the latest and greatest.

[vidartf]

If someone (possibly me) makes the effort to make a 3.x compatible version with latest features, we could consider releasing that as a release under the previous version number family (from a branch). But I don't think we can try to have dual support on the main branch.

tl;dr: if time permits

[timkpaine]

jupyterlab/frontends-team-compass#231