The word toggle is used to describe a switch that moves between two positions, or states. A toggle can be a pin passed through a loop on a rope to fasten it, or a part of a computer that switches between two functions. The earliest use of the word dates back to the 18th century, when it was used to describe a type of fastener that could be pulled and pushed through loops or holes. Today, it’s most often found in computers as a way to move between two functions, such as opening and closing an application or turning on and off a screensaver.

Toggles are a useful tool for implementing feature flags, but they’re not the only way. Savvy teams look at their toggle inventory as a kind of inventory that comes with a carrying cost, and aim to keep it low by being proactive about removing toggles that are no longer needed. Some teams have rules in place where they always add a task to the team’s backlog whenever a new toggle is introduced, with the goal of eventually removing it. Other teams put expiration dates on their toggles and create test cases that fail (or refuse to run the application) if a toggle has expired.

Depending on your architecture you can hardcode your toggle configuration in a variety of ways, from simple but less dynamic through to complex and highly sophisticated. In many cases, especially with Experiment Toggles, the most effective approach is to use a runtime configuration system that allows for dynamic in-memory re-configuration of a toggle. This will prevent you from having to restart your testing environment, or even worse, redeploy your app in order to flip a toggle.