There are a few problems with this approach, however
First, how the model works in _practice_ is not how the model works in theory. Indeed, how the model is _required to work_ in practice in order to behave even somewhat efficiently is nothing like how it works in theory
This is a lot like my impression of how people talk about programming and it doesn't work _there_ either. "I prefer programming languages that don't rely on magic" is mostly a lack of knowledge about modern systems
2/