I remember seeing a debate about this somewhere, and am currently considering removing a base object that every business object, in a system I’m working on, inherits from. It contains a few properties, some database logic, and some constructor logic.
Is this an anti pattern, or is the jury still out? Would it be better to have a base contract to inherit from, which would require a certain amount of boilerplate coding to be done in each object?
EDIT: I do like dsimcha and feel it reflects very well on the issue, I am still happy to hear any further answers
The standard rule of thumb is to use inheritance only to provide flexibility for users of a class through polymorphism, and use composition if you want to reuse code from other classes. However, as long as you’re not violating the Liskov Substitution Principle it’s probably not too bad. Writing tons of boilerplate is inherently a bad thing, too, since it obscures the parts of your code where the real action is happening and is anti-DRY. If you are violating the Liskov Substitution Principle, though, then absolutely this is a bad idea.