Support for Deeper Nesting?

[dougbrn]

Currently we're focused on just a single nested layer underneath the base layer, as motivated by the time-domain astronomy use case. Do we want to enable further nesting? The main questions would be:

• What limit if any do we pose on the number of nested layers? 3 covers the cases we've considered in astronomy (time-domain spectra).
• How should the API be built to accommodate this?

[hombit]

We need more use-cases for that