We want to create a simple database based on CSV-Files. But in order to be able to write proper unit tests, we need to get rid of the direct file access on the hard disk.
We want to introduce a new abstraction layer for the persistence. Then we could swap the file-based implementation with an in-memory approach for our tests. Also sometime in the near future it is panned to add support for cloud storage, our boss just learned about the "new clouds".
Currently the Database is limited to store only Addresses. Extend the implementation in a way, that arbitrary new object types can easily be added in the future. Verify your concept by implementing a new type "Task".
