It's this time of the year again, when our industry tries to predict the future, to decode what we are seeing and how it translates for the market and users of technology.
My predictions for 2016 are focused on APIs, probably the hottest concept and technology in both IT and business today. We just looked at how APIs will march toward ubiquity, and here is the second prediction:
2016 will see crystallization around an API meta-language
OK, first, what is an API meta-language? In layman's terms, it's a language that is used to define and describe an API. And why is it important? Web APIs are open, and even though they are supposed to adhere to specific architectural constructs, they don't carry dead weight. Older, legacy technologies are more formalized, and part of this formalism is a way to describe the service or structure. In databases, it's catalog tables: Query these system tables, and you get a description of all the database tables and their structure. In SOAP, it's WSDL: a SOAP Web Service carries along a descriptor file in WSDL, that explains what can be done with that Web Services, which methods exist, which parameters can be passed and retrieved, etc. But the REST style allows for no such formalism. As a result, RESTful APIs need to be properly documented, outside of the API itself. This documentation can simply be human-readable (HTML page, PDF, printed on paper) -- or it can be machine-readable. This is where API description languages come into play.
Today there are several API description languages: Swagger, RAML, and API Blueprint. I won't get into the debate of which one is better and why, suffice to say, they each have their merits (and fierce proponents).
In 2015, Swagger, one of the three languages, got acquired by a vendor (SmartBear). Several months later, a group of API technology stakeholders (including SmartBear and others such as Restlet), launched the Open API Initiative (OAI) under the umbrella of the Linux Foundation, with the goal to standardize how APIs are described (disclosure: I work for Restlet). While the OAI is initially focused on advancing the Swagger specification, some members, including Restlet's Chief Geek Jerome Louvel, intend to drive OAI to look beyond Swagger and to cooperate with other areas of the API community, looking at involving for example the teams behind RAML or API Blueprint.
This foundation sets the stage for a much broader understanding of the need for and the importance of an open API meta-language. Its success however will be dependent on its ability to cooperate with these other open source and industry specifications. Although it's easier to focus on Swagger and more difficult to incorporate other communities, that is the real potential here: define and promote a new community specification that represents the requirements of a larger community. Only then can we truly speak of an API meta-language, one that crystallizes the needs of everyone and covers all use cases.
The stage is set for 2016 to see this crystallization happen.