Where Einstein Meets Edison

Semantic Technologies Series: Interview with David Recordon of Facebook

Semantic Technologies Series: Interview with David Recordon of Facebook

Jul 21, 2010

This interview continues our 6-part series on semantic technologies bringing you interviews with industry thought leaders.  To follow on Thursday and Friday this week are interviews with Will Hunsinger (Evri) and Jamie Taylor (Metaweb).

I recently interviewed David Recordon, Senior Open Programs Manager at Facebook, where he leads open source and open standards initiatives. He joined Facebook from Six Apart where he focused on platform strategies, and previously worked at VeriSign in the emerging business group. David has played a pivotal role in the development and popularization of key social media technologies, such as OpenID and OAuth.

David, what you do in your current role?

The past few years, I have worked a lot on technologies like OpenID and OAuth and how we create technologies that help to connect the web together and a lot of this has been focused around APIs, so I joined Facebook last August and I focus on Open Source and standards to build out Facebook’s Open Source and Standards team, and I have three people working with me on that. I spend most of my time on the Open Source side.

Basically my role is to help engineering and product teams throughout the company using, creating, and leasing open source software and standards. I got really involved more recently in OAuth2.0 which shipped at f8.

How many developers are currently working with the Facebook platform?

There are about a million developers working with the Facebook platform.

What does Facebook do to support them?

There are a lot of things. The most basic thing is to make technologies that are really easy for developers, like an easy API. Additionally, having really good documentation; we offer some forum for developers to interact with each other and with us. We run events – Facebook developer garages – around the world; we just had one in London last week which actually Mark Zuckerberg went to as well so we paired it with a hackathon. We had teams of developers hacking on projects over night and presenting them to Mark and Mike Vernal from our platform team and a few other people and then having a developer garage in the afternoon, talking about the Facebook Graph API.

Also, we actually went out and spent a day with TechStars in Boulder, presenting our Graph API, social plugins and in the afternoon met 45 minutes with each of the TechStars companies.

So, the startup community and entrepreneurs are really important to us and I think a lot of that comes from how Facebook as a company works: lot’s of small teams, focus on impact, empower individual engineers to make a decision. We also realized as we are going to develop new technologies, startups are the ones which are able to move really quickly.

Chris Messina to whom I was talking early today had the following question for you: When will Facebook become an Open ID provider?

I think it’s probably about 2 years ago that Facebook started allowing people to sign in using an Open ID. I think to date, Facebook is certainly the largest website in the world that lets you sign in using OpenID. It’s functionality that isn’t all that apparent to users; we actually use it for bits of our sign-up process; so if you are coming from certain providers like Google we are using OpenID and OAuth to interact with your contacts help you go and find people you already know on the site when you are joining

In terms of becoming an OpenID provider, I think that’s where we are focused on right now, how to build identity on top of oAuth2.0. The past six to nine months we have been thinking: ok, if we have all these providers which are offering oAuth2 APIs, how do we go and build or re-build OpenID on top of that. That was the OpenID Connect proposal which I floated a month or so ago, looking at so now that we have OAuth 2.0 which works on web browsers, desktops apps, and mobile devices, how are we going to make OpenID something which sits on top of it, getting us to that stack of http, SSL, OAuth, OpenID.

From the perspective of a partner site, how do you decide which service to go with: Facebook OAuth, OpenID-

Twitter OAuth or Windows Live… Yes, I think that’s one of the challenges we have today, it’s not a simple decision of going to implement an OpenID 2.0 provider, even if you are looking at all the different companies, you are still going and supporting OpenID 2.0 sign-in for Google and Yahoo, Windows Live ID for Microsoft, Facebook’s platform on top of OAuth 2.0, Twitter’s platform on top of OAuth 1.0… That’s more the question I am interested in: how do we go and having learned a lot, now how are we going to try to solve the problem of multiple standards; not just how are we making OpenID better, but I think one of the ways we measure success is having 5 different companies using the same technology and people are going and interacting with it on a daily basis which you have in some areas but it’s not quite yet where it needs to be.

One of the heavily discussed topics here at the conference is the notion of the Personal Data Locker. Will Facebook become the Personal Data Locker in the future?

I don’t know! I mean obviously today, people trust Facebook with a lot of different information.

In some ways Facebook has already integrated parts of the Personal Data Locker vision. 

Yes, I think [David Siegel’s] vision, in terms of having data from all these different types of places be interoperable and having applications be really acting on your behalf, makes a lot of sense. I think there are some questions in terms of how we get there though. I think you can go and make some comparisons to the Facebook platform in terms of allowing developers, any developers, to build on top of information which people have trusted Facebook with, whether that’s status updates or things they like, people they are connected to and that you are going to get some connections.

He [David Siegel] didn’t focus much on social interactions but we are getting closer and closer to a web where social is built into its core and that very much takes social becoming a piece of the web using platforms like Facebook to do it.

Would you go as far as saying that keyword search will be replaced by social search e.g. if I want to find a restaurant you will source the information, a recommendation entirely through your social network?

I think the example that you gave of going and finding restaurants is really powerful; I think that’s what we see for news feeds to begin with – giving you a filter into the world based on your friends – that’s a really powerful concept.

The Facebook Open Graph protocol allows us to make a big step forward in that direction. However, critics like Will Hunsinger, CEO of Evri, say that the Facebook Open Graph is not truly open and leaves room for people to game the system.

I think that’s just a little bit different. I think that using the Open Graph to be really authentic will give you the best results. Because if you are looking at the web through a social context:  If I create a page for the Hilton on the web, say it’s the Hilton, add in all the right keywords, so that it would rank well – unless that one of my friends is connected to it, I am not going to see it if I think about it from a graph perspective.

But it leaves room for publishers to choose their site as a canonical representation to drive traffic rather than linking to the real Hilton site.

It all comes back to what are your friends connected to, and what you are connected to. For me, it might be this restaurant in San Francisco, one of my favorite Sushi restaurants. I want the representation for their own website, but if one of my friends had written a review about it, I might want my friend’s review instead. Hopefully this gets me to the content I am looking for.