[D9 feature parity] Killer feature: Fieldable Fields in core

This is basically the same as https://www.drupal.org/project/ideas/issues/607396 (and related to the older https://www.drupal.org/project/ideas/issues/501366)

Problem/Motivation In most Drupal projects, the need comes up for a collection of fields. This is normally achieved with a light-weight entity such as Paragraphs or Field collection.

In Drupal 8 the go to solution for this is the Paragraphs module, which over time was also expanded to control content-type display, so that content editors could effectively control layouts with it.

Now Drupal core has a stable layout builder, the need to use it for layouts is no more, but there is still a need to use it for compound fields. This however doesn't integrate as well with layout builder, because there is a disconnect between content-editing and layout building.

In addition, in core we have a use case for such a compound field in the ingredients field on the Umami recipe content type. At present it is implemented as a text-field, which prevents us from using Drupal's modelling tools to build powerful features such as a view listing 'all recipes featuring chocolate'. This is because the ingredients field is used to list three distinct items. E.g.

1 cup of flour

When really this should be three fields

1 (quantity) cup (unit) of flour (ingredient)

If we had a way to build compound fields from the UI in core, we could construct a new field-type for the ingredients list comprising the following sub-fields

• Quantity - number
• Unit - text-list or taxonomy term
• Ingredient - taxonomy term

This would unlock Drupal's relationship modelling power.

Proposed resolution Add a compound field type to core with a proposed target of being an experimental module in Drupal 9.1

The field will provide a UI to allow site-builders to build new field-types by combining existing field types.

All other functionality should continue to work.

• Views support
• formatter support
• widget support
• data storage and retrieval
• serializer support (JSON API etc)
• translation support
• revision support

Remaining tasks Work has commenced in both the Field union (PreviousNext) and rich field (Lullabot) projects to investigate how feasible this is.

The two have agreed to join-forces and work towards a common goal

We have a slack channel #multifield