Tapio by Alex de Carvalho

« Funny first Seesmic video from class | Main | OpenSocial makes the web better for all »

OpenSocial conversation with Brian

Google's announcement about their OpenSocial platform got a huge amount of play among the Twitterati today. I had a brief but interesting conversation with Brian Breslin of Infinimedia and Twitbin about OpenSocial.

Brian: http://blog.pmarca.com/2007/10/open-social-scr.html

Me: yeah I saw that.

Brian: btw I have a theory on OpenSocial: the most valuable piece of info is the graph in this case its the evolution of hyperlinks.
v1 was links between documents.
v2 was letting users create those documents.
V3 is the link between people.
So all these sites will be running these widgets in mini silos of sorts with plenty of windows for people to look in but their data is likely to be contained. OpenSocial provides a common markup language and structure sort of like xhtml or microformats. Eventually people will start communicating between silos and google will be there to facilitate and index and monetize that link so they will be building their social graph data. A meta social graph. Since they are likely to be looking in the silos (individual sites), and the conveyor belts transferring data between them.

Me: Right. A hyperlink is a hyperlink, be it between documents or between people. For example, when one blogger links to another, is this a link between documents or between people? What is see is that people are "dressing themselves up" online, just like in real life. Just like you would put a watch on, here you can put a widget that shows your iLike music preferences and what you're telling others is: music is important to me, here is what I like, and if you like it too, then let's talk. This is the premise of social objects.

Brian: right but you're talking about a granular level, and I agree w/you on what social object is, etc. But I'm talking about the platform provider level.

Me: so when you say the link between people is important, you imply Google will target ads based on the information gleaned from these relationships and user profiles.

Brian: yeah, since they will surely get into the graph somehow, on what you like and what your friends like and your activities and the context of the page you are reading. Imagine you're reading a CNN article about auto safety and the ads show "get new tires for your Toyota" instead of something more contextual like "auto safety specialists" since they will know you've got a Toyota and your browsing history.

Me: ok, but they can already do that without OpenSocial. But I see what you mean, they can get demographic data from your friends' user profiles. Location too.

Brian: ultimately it depends on how deep the socnets let google look.

Me: I was thinking how on the one hand you have members on your site adding their OpenSocial applications and at the same time, your site is an OpenSocial application somewhere else. And the right strategy now is to make yourself an FB app and an OpenSocial app and enable users to add OpenSocial apps on your own site.

Brian: Well the way we're building our apps is so that they are easy to republish to different platforms, so once the specs are out for OpenSocial, we'll have a set of modules that let us recompile for different standards since myspace is likely to have their own. But facebook could also open up fbml and bam, OpenSocial is dead in the water.

Me: why would OpenSocial be dead in the water? I think it would coexist.

Brian: well what would you build for, something with one way to publish to Facebook's 50Million + the members of whomeever adopt it or OpenSocial which would support Orkut, Linkedin, and hi5 [and MySpace]. If you are targetting US users, facebook's trounces them by far, but they are likely to coexist. I was kind of exaggerating on dead in the water. It’s going to be tough technology-wise for socnets to decide which to adopt, and how to fuel their own growth. Since they won't be able to compete on features since any socialnetwork could instantly add 1000 features via the developer community. But ultimately by being late to the party, hi5, etc couldn't make their own platform.

Me: you're right, it heavily depends on who you're trying to reach and these will probably coexist. I've been talking about "why my social media is not your social media" and OpenSocial takes it a step further.

Technorati Tags: open social, opensocial, google, facebook, orkut, hi5, social network portability, snp