在控制反转中，回调和依赖注入之间有什么区别？

Question

I understand what's going on enough to know I'm missing something. I get how in DI I'm using and Interface or Abstract Class to define what my function/method can receive. As long as my object implements or extends the given type, I can pass it to, good decoupling.

In a callback, I'm passing a function to a method that calls my other function. LIke in C, I can use function pointers. OOP language same thing but with Interfaces/Abstract Classes.

I see the decoupling, why that's good, the software can be extended, and so on, but not the difference between the two.

What am I missing?

EDIT: Is DI the pattern and a Callback the implementation?

Answer 1

Dependency injection is a software design pattern which as you say lets you decouple your code. Making for example the ability to swap your SQL backend with an Oracle backend or swapping email hosts in dev environment or live.

This naturally descourages adding say database code to your front end project which is obviously not a great idea.

A callback or delegate is where you pass in some code as a parameter to a method. You see this a lot in things like collections where you may want to use different sort functionality in different consumer code.

So if you where to write your own DI container, you would be binding a reference to an interface, say IRepository to a concrete implemention, say SQLRepository or WebRepository. Your DI code what allow consumers to define which concrete implementation to use at compile time either using a configuration file (eg unity) or in code (Ninject).