When joining a project, you have to work with the code that you’re given. It’s only when you start to understand the project that you can start to change the way the problems are being solved. Most projects that I’ve joined use an object oriented language, but not all projects use object oriented principles. In order to be able to discuss this issue, I’ve started to call it “Rich Domain vs Structs & Services”. Both are ways of solving problems, or of implementing the functionality that is needed. Both have their advantages, but I think that the Rich Domain is usually preferable.
Structs & Services
Structs refer to (complex) data structures, without any logic. The most common use is in C, which is a procedural language. It was a step towards an object oriented language, but it’s not a class yet since it doesn’t contain any logic. Operations on this data are handled in procedures. These procedures are replicated in object oriented services with (usually static) functions.
A Rich Domain uses classes with fields and methods that operate on those fields. All domain logic is captured in those methods. Services are only used to interact with other systems or dependencies. These services are not part of the core domain, since the core domain should not have any dependencies.
In my opinion, using Structs & Services in an object oriented language is a mistake. Not a big mistake, because you can still create clear, working programs using that pattern. It is more intuitive to separate the data and logic, and therefore it’s easier to build. But when the subject matter gets more complex, it’s easier to have a robust and flexible domain. According to Domain Driven Design, the insight that is encoded into the domain will help to evolve the domain to be more valuable. Another argument against Structs & Services is that you’ll end up having to implement a single change in multiple locations. If a struct needs another field, the service needs to make use of that field. This makes things more complicated than needed.