But this is very important I think if your database is the real world. If you want to reason about poverty and why there is poverty, then you can’t control the database. A database on that subject under someone’s control would be less useful. The open world assumption is very important because it is an open world. It’s a muddy messy shared world.
But there are many cases where you do need to deal with pieces of it as a closed world under your control and that’s where data shapes come in. They allow you to work in isolation from the muddy mess when you need to.