Establish our preferences for interaction/data flow between library components

See this slack thread: https://dhis2.slack.com/archives/CBM8LNEQM/p1592465129115600

There are different ways for our library components to interact.

• Mapping and cloning children
• Context
• Something else?

It would be good to establish what we prefer as a team, and in which situation (and why). That way we all operate on the same assumptions, as it seems currently we're not completely aligned in what we prefer here.

Plus it allows us to investigate the upsides/downsides and document them. And we could see what other libraries have settled on as well.