The strategy on our release train is to write all new APIs in Kotlin since it has many advantages over Java.

The microservices at Auspost follow the RESTful API style, use JSON, and are mostly written in Java and Kotlin using a variety of Java frameworks such as Spring Boot and Spark Java. My previous article mainly focused on the client so let’s look at the changes required in detail on the backend.

I always consider strategy in Software Engineering as a marathon, not a 100-meter sprint, changes take time to implement. We still have quite a few more APIs to convert, more teams to talk to and we may never hit 100% but it’s definitely a good start. Twelve months later we now have 7 APIs using Kotlin Serialization for JSON parsing, these models are published to our Enterprise Nexus Repository and used by our Android App - this represents about 20% of all the APIs used in our App. We were never really sure about our strategy to share our API models with our clients would really work not because it was a bad idea but because we knew it would be difficult to get buy in from our API developers. Continuing on our journey from the article “ The first steps on your Kotlin Multiplatform journey - changing your microservices” that I wrote last year, I wanted to give an update on our progress and share some learnings.

Android kotlin serialization