OpenSocket: a Thought Experiment
Posted by Dan Lester on November 1st, 2007 8:00 AMIf there’s anyone who gets the concept of an API, it’s Oren Michels of Mashery. At app-camp the other weekend, I heard him extol the virtues of opening up your web site to the world. “Facebook Platform,” he explained as part of his warm-up act, “has a reverse API“. You open up your API, to a particular specification, and Facebook agrees to host the data and the experience that you’ve opened up to them. “Yeah, yeah, that’s just semantics,” I thought - because I didn’t think.
Later on, I felt conscious that everyone was wondering what I was doing sitting in this small but crowded gathering, so I thought I’d get involved. Relieved that I didn’t need to apply for an API key to access his interaction points (ears), I asked him if any customers whose websites he has helped open up have come to discover that their API undermined their business model or revenue streams.
Not to Oren’s knowledge, no. If you have a web site in the first place, it’s because you think your presentation layer is your secret sauce. And in fact you should be looking to source other APIs through your web site. [Now here it comes...] “Well, that’s what Facebook did, right?!”
It’s true - they had a beautiful website, an all-knowing algorithm that tells you only what you need to hear… So they create the platform; now we can build the applications that are too numerous (or useless) for Facebook to bother developing themselves.
This invaluable exchange with Oren was back in my mind following the recent announcements about OpenSocial. It occurred to me that we don’t even need to wait for Zuckerberg to surrender: as a pre-release partner in OpenSocial, I used my API key to develop a Facebook app code-named OpenSocket that implements an OpenSocial container within Facebook. It’s a bit like running a DOS command prompt within Windows, but as a means for me to tinker around with both APIs, the concept was interesting enough to see what I could get up and running.
And, at three o’clock in the morning, there it was: Visual Bookshelf for OpenSocial running inside Facebook. Everything’s gone full-circle! Life is complete.
Facebook - the guys who showed us the future - reduced to being just a container in the words of Google. Back in your box! Doing what you always did best: presentation. The shopping mall where we hang out with our friends even when the shops are closed.
Of course, my app doesn’t really work properly. And I wasn’t really trusted as a pre-release partner - hell, I’d have broken the internet or something.
But listen, Facebook: there’s always a place for you at this party. If I could choose where to access OpenSocial apps, and choose where to pick up an intelligent digest of my friends’ activity, I know where I’d rather stay.
[Dan Lester is a backseat driver in the Facebook revolution. He advises the social media strategies of non-profits and multinationals alike. His consistent number one piece of advice is to seek advice from someone better than him. Stay away.]
(4.64 out of 5)
(4.17 out of 5)
Add New Comment
Viewing 2 Comments
Thanks. Your comment is awaiting approval by a moderator.
Do you already have an account? Log in and claim this comment.
Do you already have an account? Log in and claim this comment.
Very interesting post. You and I are on the same wavelength. As somone who built a 15+ year-old business around mapping API's (i.e., Mac2Win -- the Mac Toolbox API implemented as a Windows DLL so companies like Fractal Design, Macromedia, Claris, etc. could port their Mac apps to Windows with just a recompile and link).
In the porting technology business, all you have to do is map one Native API to another Target API and this insulates the developer from having to learn all the nuances necessary to write apps for these other platforms. The idea you described would be as trivial as what Marc Andreesen did when he spent the weekend mapping the OpenSocial API's into Ning's existing API's.
So, I agree that anyone (not just facebook) could build an API mapping layer and voila! you would be able to write an OpenSocial-compliant app that when linked together with your OpenSocial2Facebook layer would work just fine as a facebook app. The problem with this approach is that there are no OpenSocial apps that need to be ported over to facebook.
Therefore, I suggest this thought experiment. What if Facebook (or someone else) wrote a porting technology that would take the facebook API's and build a mapping layer from them over to the OpenSocial API's. This way you would be moving in the direction from LOTS OF APPS needing porting to LOTS OF WANNA-BE SOCIAL OPERATING SYSTEMS needing apps. Whoever writes this mapping layer could also target MySpace and Yahoo.
In this case, Facebook developers focus on facebook first (which is where all the innovation is) and the Facebook Porting Technology company keeps up with Facebook by writing updated porting layers for each other Social Network that manages to get a critical mass of users.
Thanks,
Lee Lorenzen
CEO, Altura Ventures -- the first facebook-only VC
(c) 2007 Altura Ventures LLC.
Do you already have an account? Log in and claim this comment.
To really tear down these walls, lets get an OpenSocial app running in my OpenSocket for Facebook, but appearing in Orkut courtesy of your Facebook Porting Technology...
Add New Comment
Trackbacks