I looking at the source code for Code Camp Server, and noticed that in the domain layer there are interfaces for each repository. However, the actual implementations live in the DataAccess layer. This is a little different from the standard layered architecture approach where each layer only references the layer directly below it. In this case, the DataAccess layer references the domain, not the other way around. I understand why this is done (keeping domain PI), but was wondering, what if (rare, I know) the domain objects need to access the repositories? This is mentioned in Fowler's Poea. How can we do this without creating a circular dependency: domain -> data -> domain etc... ...
Go to the complete details ...