I'm curious about the two main approaches to resource management between various services or managers in large-scale projects. The first is Dependency Injection (DI) where you simply pass the needed resources into a manager's constructor. The other is the Service Locator pattern, which registers managers in a centralized registry so they can access each other's resources by requesting pointers.
In large, multithreaded projects with extensive collaboration between managers and a focus on unit testing, which method do you think is superior? Are there other options I should consider as well? Any advice would be greatly appreciated!
1 Answer
I've found that the effectiveness of DI versus Service Locator often depends on the programming language and the project's specifics. For example, in Java, DI frameworks help avoid many wiring issues and enforce explicit dependencies, which is great for static analysis and compile-time safety. On smaller projects, especially in JavaScript, a simple constructor injection usually suffices. Service Locator could introduce hidden dependencies since it makes the injection less explicit, complicating the dependency graph and making it harder to track down issues when they arise.
Thanks for the insights! I'm working in C++, and I see how what you shared applies here too. But how would you handle a situation where a class requires too many dependencies? Is that a common problem?