Yes, getting a domain model right up front can be very hard. That's why behavioural interfaces that hide as much as possible were found to be more forgiving w

Yes, getting a domain model right up front can be very hard. That's why behavioural interfaces that hide as much as possible were found to be more forgiving when we realise that we failed to predict the future correctly when dreaming up V1 of the product. They make it easier to evolve.