Patterns of software development: “Coincidence pattern”
Software source code is not a big truck a static thing. It evolves over time. Some piece of code being repeated in several places doesn’t always mean that it should be DRYed right away. There’s always such thing as coincidence. When you are putting some functionality into reusable module it should have a meaning applied to the business logic, not just “repeated lines of code”.
What’s the big problem with the “premature drying”? When compiler looks at the code, it just executes it. It doesn’t really matter how deep in the classes and functions some expression is located. It is just a matter of performance.
But when developer looks at the code, he constructs abstractions and loads them into his brain. Developer speaks abstractions language, not expressions. When I see some method/module being used in several places, I understand that as a kind of relationship existing between those places of the system. I see something in common. This idea influences further understanding and I’d better not to be wrong at some point.
Rule of thumb: don’t try to save less than 273 bytes of code if it may lead to confusion and create more abstractions or rules to be remembered.