Inversion of Control, by any other name...

Table of Contents

We often hear (at every bloody interview) about the SOLID Principles. It’s regarded by some as the major driving force of good software. But why do so many engineers get (at the very least) 1/5 of the principles wrong at a fairly consistent rate?

The D in SOLID

SOLID stands for:

  • Single Responsibility
  • Open/Closed Principle
  • Lisckov Substitution Principle
  • Interface Segregation
  • Dependency Inversion

If I had a coin for every instance an engineer has told me that the last “D” in SOLID stands for Dependency Injection, I’d have 20 coins (I’m young, sue me).

By any other name…

Enter the other name for Dependency Inversion…Inversion Of Control.

I lie here a bit. Rather than a synonymic term, Inversion of Control is to Dependency Inversion, what the ocean is to a sea. A bit of history, Martin Fowler popularized the term in his blog around 2004, along with outlining the Dependency Injection PATTERN. He ‘concretised’ the IoC principle by making it tangible to a real use case - the management of dependencies.

Re-enter Dependency Inversion. This became a semi-concrete term for inverting control of dependencies. Since Dependency Management is where the IoC principle is (arguably) the most useful and tangible, the name has become popularized into a synonym for IoC (I suspect to complete the acronym ‘SOLID’, which I actually appreciate tbh. The word ‘SOLII’ sounds a bit flimsy imo).

Yet Another Explanation of IoC

The net is ripe with much much much better explanations and definitions of IoC. Here is yet another one:

Inversion of Control outlines where ownership of execution should live within the design of a system. It advises that implementation details be isolated to the implementor and hidden from the driver, with the implementor accepting a command to do something from a driver, rather than exact instructions detailing how to do it.

Basically, to quote Jedi Master Fowler, its the “hollywood principle”. “Don’t call us, we’ll call you”.

Where the term “Dependency Inversion” shines at a software implementation level, the term “Inversion of Control” shines at the “integration between software” level. I’ve personally found the most use (not that other implementations are not useful) in integrations between multiple systems, particularly where these systems are owned by different teams. The principle can help direct questions where there might be 2 good answers.