We didn't really discuss it, but it's certainly an option! (I'd argue that other things like VOTable belong in the core package, but no strong opinions on astropy.samp)

I also agree that votable should stay in the core since it is very much integrated with the unified I/O.

This would also require to de-couple the submodules from the top-module.
(It's not necessary to import the whole dal stack for just broadcasting a table.)

At Astropy Coordination Meeting 2019, this was decided to be a good idea. But io.votable would stay in astropy.

Is there a roadmap for this?

Renewing this effort...

The goal: We would like to deprecate then removeastropy.samp from astropy in favor of SAMP capabilities in pyvo.

Note that pyvo/samp depends on astropy.samp, so that reference will need to change. This work should ensure that existing pyvo SAMP capabilities are retained and that astropy SAMP capabilities are supported and documented.

For more details on this see the notes from unmerged PR #239.

Please cross link the new PR here when you have one. Thank you so much! 🙏

I'm removing the milestone from this one as this is not a thing that should block the 1.5 version.