The assumption that data is something to be maintained internally, at best hidden behind an interface (a procedural one) and at worst "exposed" is so ingrained that it's hard to think of it any other way.
However, why can't we have "data" as the interface? The WWW and REST show that we can, and that it works quite swimmingly. With a REST interface you expose a "data" interface (resources that you can GET, PUT, DELETE, POST), but these resources may be implemented procedurally internally, the client has no way of knowing.
I've also done my bit, with In-Process REST[1] and Polymorphic Identifiers[2] for the "REST" half, and Constraint Connectors[3] for the "reacting" half.
[1] https://link.springer.com/chapter/10.1007/978-1-4614-9299-3_...
[2] https://www.hpi.uni-potsdam.de/hirschfeld/publications/media...
[3] https://www.hpi.uni-potsdam.de/hirschfeld/publications/media...