In 2004, pretty much nobody was worrying about RESTful APIs. But we’ve seen an exponential increase in interest over the past 13 years. If that interest continues for the next 13 years, 2030 web resources might be very different. We might hit the API Event Horizon!
It’s possible that the RESTful paradigm might go the way of SOAP. But HTTP has been around for over a quarter-century now. It’s very probable we’ll still be using it in 2030. And if we’re not, something else will take its place that we’ll have to worry about.
So it’s a good bet that API, in one form or another, will be used for data and information exchange well into the future.
2030: Specialized Search for APIs?
Web crawlers have been crawling the web for more than two decades. But we need to have intelligent agents that can recognize what they find.
I can find the API of a company or its developer program – but I just get a link to a site. I still have to suspect that the API exists in the first place. And I’m on my own when it comes to understanding what the API does, and how to get it to work.
It would be great to have some kind of specialized search agent that spends its time looking out for APIs. And when it found one, it wouldn’t just flag it up to a human. It would analyze the documentation, work out what the endpoints do and how the authentication scheme works. It would even be able to write the documentation just from knowing what the API does.
Intelligent Knowledge Bases in 2030
How could it do this? Through its intelligent knowledge base.
The knowledge base could consist of a series of rules, much like what might have been in a 1980s-style expert system. For example, “If you find something that suggests the API is using OAuth2, then do this.”
But by 2030, we will long be past the age of even IBM Watson and AlphaGo. We’ll finally have systems that will combine depth and breadth of knowledge with an agent that is robust and flexible. These agents will learn by watching what API engineers do, talking to them about it, and reading up on the subject.
They won’t be true Artificial General Intelligences. They’ll know enough to know what they don’t know, but they won’t be afraid to ask. And they will never forget once they’re told.
Of course, we shouldn’t just rely on agents stumbling onto APIs. Discovering any kind of web resources has been a problem since the beginning of the Web. By 2030, an intelligent web ecosystem will finally have emerged in which Web resources describe themselves and then publish their details on appropriate registers.
These registers might be something like Programmable Web or Google. Or it might be something more like a special kind of agent that listens for announcements from resources and then itself passes on the information it receives to other agents.
The agents will be able to handle the most complex authentication scheme. Indeed, by 2030, the schemes we use today will long since have passed into history, replaced by ones that are designed around the capabilities of the agents.
2030 Web Resources: The Big Question
A big question with APIs or access to any other web resource is, who do you give it to? This is where blockchain-like technologies and trust exchanges will come into play.
I might let you use my API if I trust you. But if I don’t know you?
Reputation comes from third parties. It’s not just what an agent has done. What do other people think of the agent has done? But there’s the possibility of bad faith here. As on TripAdvisor and Amazon, people can pay to post low ratings and poor reviews about rivals. Fake news!
Hardly an issue restricted to APIs, but trust and reputation exchanges and brokerages will help reassure you that the agents that want to use your API won’t try any shenanigans.
And from APIContext’s own particular corner of the API universe, the agents will constantly monitor the APIs they are interested in. If something goes wrong with an API, it will be identified immediately. Sophisticated tools will detect potential problems with APIs before they affect users and provide deep understanding of overall performance and quality – and whether the APIs are meeting their Service Level Agreements.
This won’t be free, of course. You will be able to trade information for trust and reputation, and convert trust and reputation into cold, hard cash, probably via some crypto-currency.
Smart agents and their associated knowledge bases will have a pretty revolutionary effect on every aspect of web-related activity. I’ve just sketched out a few ways in which these technologies will impinge on a niche, but vital part of the noosphere.
We are certainly going to be living in interesting times over the next decade – and it will be a blast to observe exactly how things evolve.
One thing we can be sure of: there’s going to be plenty of scope for the development APIs and their ecosystem in the years to come.
