In order to properly support lexicographical data we need the concept of a sub-entity for Lemma and Sense.
Things to consider:
* Structured IDs (we will use the hyphen as a separator, as in L768234-F7)
* Lookup and Store (EntityRevisionLookup, EntityStore)
* Assigning fresh IDs (even after sub-entities have been deleted).
* Just track it as part of the parent entity? Where should it (not) be exposed?
* How to represent sub-entities in wb_terms.
* See also {T159718}
* Kill wb_entity_per_page, PropertyInfoTableBuilder, SqlEntityInfoBuilder, …
* see also {T95685}
* Display and editing
* investigate work needed for static display of nested entitites, as well as for making the dynamic editing UI work with multiple (nested) entities.
* Define an RDF binding (Special:EntityData).
* Check how much we can rely on LEMON.
* Serialization + Diffing (should be trivial)