How Repository, IdentityMap, UnitOfWork, DataMapper, QueryObject and VirtualProxy work together?

Question

I know how all those patterns work separately, I have read some Martin Fowler articles and books. The problem is that I can find only separate different implementations of each of these patterns but I still haven't found some reference architecture diagram (UML or other) which I could use as a starting point to develop a prototype for an object-relational mapping tool, which provides persistence ignorance for domain model.

Could anyone direct me - where I could see how all those patterns work together? How all these things are connected and where I can use concrete implementations and where abstract interfaces should be used?

I know that there may be debates when and how each pattern should and should not be used, but for now let's pretend that I want to use them all together and see how they fit with each other the best way.

Thanks.

Answer 1

You might want to have a look at Davy Brion's 'Build your own DAL' series. I've only skimmed it myself so I'm not sure if it addresses your question.

Be sure to also have a look at Ayende's comments to the series.

• A guide into OR/M implementation challenges: Reasoning
• A guide into OR/M implementation challenges: Mapping
• A guide into OR/M implementation challenges: CRUD
• A guide into OR/M implementation challenges: Hydrating Entities
• A guide into OR/M implementation challenges: The Session Level Cache
• A guide into OR/M implementation challenges: Lazy loading
• A guide into OR/M implementation challenges: Custom Queries