I’ve been exploring the topic of schema entropy, the idea that we can understand the complexity of a schema by exploring all possible variations of the schema. The more variations, the higher the schema. The higher the schema, the more complex it is. When evolving an API, it’s helpful to look at how the entropy of schemas change. For example, with request bodies, it’s necessary to increase entropy for each schema change.

Entropy is concerned with the amount of thermal energy that is available for work. The higher the entropy, the less the available energy for work. And vice versa. The second law of thermodynamics states entropy will increase over time in a closed system until it reaches equilibrium. All the available energy will level out eventually as entropy increases. I recently wrote about the idea of understanding the complexity of a schema by measuring its entropy.

I’m working on a personal project with the goal of giving insights into the complexity of an API design. We don’t have many ways or tools for looking at complexity, and the ones we do require experience and knowledge of API design. I’m hoping to find ways to look at complexity through a different lens. One way I’m working on now is what I call schema entropy. An intro to the idea of schema entropy Schema entropy is a way to look at the complexity of an API schema by determining how many variations of the schema are possible.