AOL Surprises Developers: the Future Opens?

Last week was a memorable week for me and, I think, for many developers, architects, and product managers at AOL. The news about AOL and OpenID wasn't published in a formal announcement, but rather in a brief blog post that was written in response to rumors and curiosity circulating around the Web about what AOL was up to. The site openid.aol.com had been discovered, and questions were being asked. Microsoft had publicly announced its OpenID initiative at the RSA Conference the week before.

The news that 63 Million AIM users now had an OpenID URL was greeted very positively by most people who know what OpenID is. My Friday post "AOL OpenID News Excites the Blogosphere" even received 4 diggs and its Digg page became today's top Google search result for "aol openid".

But there was also a significant amount of surprise: "Huh? AOL did this?"

Huh? AOL? You Mean, Like, the AOL that Sends Me All Those CDs?

Over at the ClaimID.com blog, Fred Stutzman's post AOL Embraces OpenID brought these reactions (among others):

  • "I'm surprised AOL of all companies would do this, but that's a great step forward"
  • "AOL's doing their best to get people to love them again"
  • "Very cool news. Now I just have to get my head around this!"

OpenID and "AOL is Open"

Having attended the Web 2.0 Summit in November and listened to Jonathan Miller and other AOL people tell us about "AOL Is Open" and then investigating the APIs, speaking with AOL developers at conferences, by email and AIM, having visited the AOL headquarters in Dulles, Virginia -- and writing blog posts, finding authors who were willing to investigate the APIs and write articles and blog on dev.aol.com/blog... and feeling like it was all being done in a vacuum, void of notice by anyone -- I was very much ready to see some appreciation from the technology community for what's happening at AOL!

And I don't say that as just a partisan statement either. I'm quite realistic, almost 4 months into my tenure working on this site:

  • Yes, it's true: AOL used to send us all too many CDs we didn't need offering us N-thousand free hours. Most people still don't know AOL is no longer in the paid-subscriber business. They're not. They're trying to reinvent the entire company on a new business model. That's not an easy thing to do.
  • "The APIs are buggy!" True, the new APIs are not perfect. 18 months ago, there were no APIs at all. Some of the APIs are still listed as "Betas" which means all the bugs haven't yet been worked out. But, when they are all worked out, they'll provide access to a fully-scaled infrastructure that has been supporting tens of millions of users on a daily basis. Which is what the non-beta APIs already provide. Not a bad infrastructure to bring into your own application or startup site.
  • "But they called it REST and it's not REST!" Yes, in the huge transition AOL is attempting, there is some chaos. In what company where a large percentage of the people are laid off is there not chaos? The company is trying to survive. Buzzwords may seem to some people (non-developers) like they will help the company survive. Hence, if you look, you'll find some inconsistencies between what is stated on a Web page and what's really in the software. But that doesn't mean that what's really there in the software is flawed. The software itself can still be highly robust and reliable -- just mislabeled by someone who was asked to construct a Web page to attract visitors, and most likely the misrepresentation was unintentional.
  • "The documentation is poor (or nearly nonexistent)." In some cases, yes, in others I see nothing to complain about. The documentation varies, which I'd expect since the APIs interface with software that was entirely proprietary not long ago. I can also tell you, the API documentation issue is being addressed in multiple ways. First, we are beginning to publish articles that demonstrate how to get started with the APIs. The flow has just started, and our first articles are introductory in nature. The articles will go into more detail as time proceeds, providing developers with illustrated tutorials on the APIs and working, reusable code. In addition, I've seen a plan AOL is putting into place to fill in the gaps in their own API documentation, and document the APIs using a standardized format that, to me, looks quite sensible.
  • "But... it's AOL!" Yup, it's AOL. The company that catered to non-tech people and got them online through dial-up. Also the company whose developers built an infrastructure that was capable of doing that in a more efficient and reliable manner than anyone else in the Internet's early years as a public resource (hence, their survival, while others like Compuserve and Prodigy fell away). I'm an infrastructure guy, myself. I'll take AOL's infrastructure and trust the developers who created it and who today are creating the APIs that make that infrastructure available to all developers.
  • "But... it's AOL!" OK, I'll admit, I have an unfair advantage. I've talked to the AOL developers, I've watched and listened to their presentations, I've tested out their Beta codes, I've seen just how much they love what they're doing and enjoy the freedom AOL is giving them to experiment with new technologies... So, I know first hand that these are fine, geeky people, just like you and me. I suggest that you give their latest creations a test drive, then decide what you think.

Conclusion

In sum, the AOL APIs -- and the new "open" AOL -- are works in progress. Now, if Google or Microsoft or Yahoo! or enterprising start-ups have "works in progress" that's considered a good thing. So, why do so many people, developers especially, seem ready to make an adverse exception in the case of AOL?

I'm glad that the AOL OpenID news is very surprising to many people. Maybe now, some of them will be curious enough to take a look at what other "surprises" (though they've been doing their best to publicize them) AOL's developers and architects have been working on lately...

-- Kevin Farnham
O'Reilly Media