Web 2.0 Expo in San Francisco and are getting inundated with press releases about new APIs and developer platforms, many from companies we've never even heard of in the first place. How long ago was it that the forward-looking thinkers argued that APIs and platforms would soon be available everywhere?We're here at the
That time is clearly fast approaching and it makes us wonder: now that this matter is settled, what comes next? We asked a variety of people here and around the web what they thought will define the next frontier, what will build on the emerging foundation of ubiquitous APIs. We got some interesting answers.
Readers who appreciate this discussion may also enjoy our previous post titled "APIs and Developer Platforms: A Discussion of the Pros and Cons."
For now though, let's acknowledge that there are scores and scores of companies that have considered the pros and cons and decided to launch APIs. Answers to our question about the future ranged from the technical to the social, political and economic. They often fell into a small number of big-picture trains of thought. We hope you'll share your thoughts on the future in comments.
The most interesting replies to the question - "what comes next after APIs become ubiquitous?" can be summarized as follows:
- Business models
- Filtering for information overload
- Standards and interoperability
- Outsourcing API services
Backlash is included in our list because there is definitely some push back. Some folks think APIs are for suckers. We don't agree but those responses were interesting as well.
Perhaps inspired by a touch of cynicism, many of the people we talked to said that finding business models was the next frontier for an API enabled web.
Nick Gonzales of ad network Social Media said that the early rush to build apps on the Facebook platform should be considered the exception more than the rule. He says it was remarkably easy to build apps on that platform but that hasn't helped developers make money outside of Facebook. It hasn't been the kind of opportunity that many big companies have taken advantage of yet, either.
What he believes may be necessary in order for big money to be made is long term agreements between platforms and developers, assuring continuity of availability and opportunities to monetize.
Gonzales addresses monetization from an ad network perspective, but other business model options include premium access to APIs or - more likely in this author's opinion - APIs leveraged for lead generation.
Filtering for information overload
It's one thing to smash together different streams of data, but making sure the results are user friendly is another matter. Many people we talked to said they wanted APIs and platforms to increase their capacity for determining relevance.
Student entrepreneur Abhishek Nayak put it well when he called for a future characterized by "better platforms like FriendFeed, to make sense of all the information and noise from your social networks." Blogger Eric Eldon of VentureBeat felt similarly when asked what comes next. "FriendFeed will rule," was his three word answer to the question. How incredible is it that such a young startup has gained Twitter-like metaphor power already?
Ian Kennedy of Yahoo's MyBlogLog predicted that filtering and duplication removal will be big. Software consultant Lokkju Brennr got even more specific: "It will have to be actual natural language parsing," he told us, "combine that with all those APIs, and you have knowledge, instead of just data."
All of those are fun things to think about and bring to mind the semantic web technologies we write about here often as well.
Standards and interoperability
The most obvious answer to this question as far as we're concerned is that the next step is to make ubiquitous APIs standards-based and able to work together.
Tech consultant and co-founder of the Yahoo! Developer Network Jeffrey McManus disagreed with the assessment of APIs being truly widespread but said "the next step is to make them not suck and support them well."
Probably the easiest way to do that is to build them on existing standards. Though many of those are still half-baked pipe dreams today, that's not true in every case. See Anil Dash of SixApart, for example, who says that "AtomPub has become the standard for accessing cloud-based data storage."
Florida venture capitalist Dan Rua puts it this way: "after ubiquitous APIs comes category subsystems/adapters, allowing for write once, run with any similar service type abstraction."
One of the most interesting replies we got to our question was from Aaron Fulkerson of MindTouch. Aaron says that the next step is for developers to engineer for concurrent processing; to make APIs not just interoperable but intelligently orchestrated to be called in concert. The idea here is to create multi-step or functional mashups. He didn't just come up with that off the top of his head on the floor of the Web 2.0 Expo - it turns out that Mindtouch offers an Open Source framework to accomplish exactly what he's describing as the flavor of the future.
Fulkerson wasn't alone in his hopeful prediction, either, though. Ruby geek Audrey Eschright called for something similar: "Real innovation in the services built on top?" she asked rhetorically. "Not just content mashups, but new kinds of tools."
That sounds great to me.
Outsourcing API services
One of the companies we write about often here is Mashery, a startup that manages APIs for companies who want to offer them but prefer to outsource their management to knowledgeable experts. Mashery reports rapid growth and could represent a key part of the future.
We hear whispers about a number of beyond-stealth startups, too, that are aimed at solving the scalability problems faced by some of the most popular APIs on the web. That's not at all a dry matter - commoditization of solutions to the biggest technical bottlenecks in making APIs work would open up a whole new world of possibilities.
While this may be the most hard-business vision of the next frontier, it's also one of the visions I get most giddy about. Any time I get giddy it's probably a good idea to talk about...
Many people we talked to said that the next step was likely to be one going backwards, away from the frothy wave of "Me Too" APIs and platform announcements.
We heard this from some people we really didn't expect it from. David Janes, creator of a sophisticated lifestreaming app for groups called Onaswarm summarized his feelings thusly: "How about a return to using well-known protocols (as opposed to APIs) for doing well-understood tasks, i.e. publishing and posting data. E.g. RSS/MetaweblogAPI or Atom/APP...It's insane...I've had more than my fill of working with these APIs." When I pinged him to confirm those lines - he said that it would more accurately explain how he felt about the APIs he's been working with if there were some obscenities sprinkled into his quote. That from a man who has put his hand into the dragon's mouth. If you will.
Similar sentiments were expressed by Len Kirby, engineering director at Flock. Flock is a social browser that brings together a large number of social data streams and functionality from all around the web. None the less, Kirby is no fanboy of the latest wave of APIs and platforms.
He told us that he thinks the next step in fact may be going back to predictability based on finished standards; as opposed to the half-baked protocols of industry luminaries that didn't finish developing proposed standards. Tempering his vitriol just a touch, this manager at a company built on Mozilla technology also shared some sympathy. "Just like any visionaries there's only so much time to make things real and [for example, coming out of] Mozila and RDF there's been a lot of very good things - but time to market rules." Kirby made sure to affirm as well that Flock really does love Mozilla.
Everyone's got a soft spot for inventors, but a substantial number of people are pushing back on the deafening roar of announcements about new frameworks for invention.
That said, you won't likely hear any of those voices blogging here at ReadWriteWeb! We think that today's crush of APIs and platforms is just the beginning, that we're at a turning point of innovation. We love it and intend to chronicle the next steps as best we can.
Here at the Web 2.0 Expo there's plenty of opportunity to discuss what that future might look like. Internet time traveler Dion Hinchcliffe counted nine major web mashup announcements at the event before lunch today and asked if critical mass had been reached. If that's not true yet, all indications are that it certainly will be soon.