API evolvability is key in a system of systems
— Greg L. Turnquist (@gregturn) March 28, 2017
Heh. That tweet was a follow up to Oliver’s latest presentation on the topic “API evolvability is key in the system of systems”. So what does this stuff mean, and why is it important?
API evolvability is one of the core underpinnings we get when we use the constraints laid out in Roy Fielding’s dissertation. And what is that exactly? In essence, the web grew FAST and is a major success. Many of the reasons are that is constrains us in certain ways, thus forcing certain things forward. Web sites are easy to evolve WITHOUT evolving the web browser. The value of this cannot be stated enough.
For decades, developers have faced the issue of evolving APIs. The concept of REST is take some of the underpinnings of the web and take them into API-land. Evolving APIs with similar constraints should make it possible to scale such changes and avoid some of the brittle, backwards-compatible breaks we have suffered for years, which can lead to slow downs in evolution.
How does this tie in with Spring?
I recently moved back to the Spring Data team. I had been on that team before, working on Spring Data REST. A key library that supports it is Spring HATEOAS. In Spring HATEOAS, we have an API meant to build hypermedia-powered RESTful APIs. And due to the incredible popularity of ALL Spring Data projects, Spring HATEOAS hasn’t received enough love.
Since jumping on this fab team, I have started putting together a list of proposed updates, enhancements, and polishings. Next week, I’ll be in Germany with the rest of the Spring Data team having a coordinated meeting. Among the many topics, Oliver and I will be hammering out this plan to propel Spring HATEOAS forward to be benefit of the community as well as the Spring Data REST community.
And one thing that Oliver has invested much time in speaking at various conferences about DDD, REST, and API evolvability. I have a strong desire to create some sort of code repository, a playground if you will, where people can checkout things and see how to evolve an RESTful API using Spring HATEOAS and the concepts of REST.
— Mark Paluch (@mp911de) March 28, 2017
That way, we aren’t just preaching a high falutin’ concept, but showing how it can be done.
And I also imagine that in putting together such a repository, we can also find the gaps and holes in Spring HATEOAS. I suppose it’s risky putting this all out in a blog post, but I just wanted to share some of the ideas floating around in my brain as we ramp up support for Spring HATEOAS.