Facebook is the Most Accurate Social Graph
Posted by Nick O'Neill on February 4th, 2008 9:30 AMFollowing Google’s announcement Friday of the release of the Social Graph API, a number of experts have had time to voice their opinion. Tim O’Reilly, an apparent Google evangelist, believes that this new API brings us one step closer to “the Internet Operating System.” Others such as Jeff Jarvis, Dare Obasanjo and myself believe that this is good in theory but not very practical. Some could argue that eventually there will be tools that make it much more automated to define our connections between others on the web via hyperlinks.
Unfortunately (for Google), the majority of internet users will not switch to this new system when they have already become used to defining their connections via social networks by simply clicking “Add Friend.” Perhaps I am being a little short-sighted and developers will rush to build tools that make it easy for general web users to embrace the new Social Graph API. I doubt that is going to happen anytime soon though because currently there are very few incentives for developers to build applications using the Social Graph API.
When the Facebook platform launched, many developers rushed to build applications because of the opportunity it provided. Suddenly developers could build an application an instantaneously and have hundreds of thousands of users check out what they had built. This is in contrast to build a traditional website which requires an immediate advertising budget or continuous content production to obtain any sort of traction.
So for the time being, Facebook has the most accurate representation of the social graph. Until Google can provide significant incentives for developers, the Social Graph API is going to face significant challenges in attempting to gain any traction. It is simply too easy to add a friend on Facebook versus adding a contact via the Social Graph API. Do you think Google can make the new Social Graph work?
(4.64 out of 5)
(4.22 out of 5)
February 4th, 2008 at 8:52 am
When it comes to discovery of a user's social graph, the Social Graph API is essentially retroactive — if a social network implements it, then all of its users' “Add Friends” (past, present and future) will be discoverable via the API. If significant adoption occurs, the real question won't have to do with how easy it is to extend your social graph, it'll be whether Facebook's social graph is more valuable than the sum of all the social graphs available on every other network that implements the OpenSocial API.
In that sense, the Social Graph API is taking on Facebook in the same manner as OpenSocial: Facebook vs. everyone else. And as with OpenSocial, its success will be measured by how many major players get on board.
February 4th, 2008 at 9:27 am
The use of the rel attribute is cursed with the same fundamental problem with meta-data in general.
People can't be bothered to complete meta-data.. and even if they could be bothered, the user generated annotation will be disjointed and rarely follow any standards.
So - I totally agree that with Tim O'Reilly - while this is one step closer, there's a long way to go before its social graph data is as useful as Facebook's social graph.
Paul Reilly
Founder of the Church of Facebookology
http://www.facebook.com/group.php?gid=5489961708
February 4th, 2008 at 9:52 am
When it comes to discovery of a user’s social graph, the Social Graph API is essentially retroactive — if a social network implements it, then all of its users’ “Add Friends” (past, present and future) will be discoverable via the API. If significant adoption occurs, the real question won’t have to do with how easy it is to extend your social graph, it’ll be whether Facebook’s social graph is more valuable than the sum of all the social graphs available on every other network that implements the OpenSocial API.
In that sense, the Social Graph API is taking on Facebook in the same manner as OpenSocial: Facebook vs. everyone else. And as with OpenSocial, its success will be measured by how many major players get on board.
February 4th, 2008 at 10:27 am
The use of the rel attribute is cursed with the same fundamental problem with meta-data in general.
People can’t be bothered to complete meta-data.. and even if they could be bothered, the user generated annotation will be disjointed and rarely follow any standards.
So - I totally agree that with Tim O’Reilly - while this is one step closer, there’s a long way to go before its social graph data is as useful as Facebook’s social graph.
Paul Reilly
Founder of the Church of Facebookology
http://www.facebook.com/group.php?gid=5489961708
February 4th, 2008 at 10:44 am
@Paul, the idea behind the Social Graph API is for the social network to populate the metadata straight from it's own social graph, so in theory, as long as the social network fully populates the metadata, there should be no difference between that data and the data that is inherent to Facebook's social graph.
February 4th, 2008 at 11:44 am
@Paul, the idea behind the Social Graph API is for the social network to populate the metadata straight from it’s own social graph, so in theory, as long as the social network fully populates the metadata, there should be no difference between that data and the data that is inherent to Facebook’s social graph.
February 4th, 2008 at 12:27 pm
Nick, I'm afraid that I don't see any evidence in your post supporting your assertion that “Facebook has the most accurate representation of the social graph.” Did I miss it?
February 4th, 2008 at 1:08 pm
I have no interest in having my social graph exposed to the world. I would deliberately falsify my link info. Theoretically, Facebook has more security.
I also do not like the idea of Data Portability where my info could end up on a Social Network that I did not sign up for. I don't mind my Facebook or Linkedin friends having it but I don't want them spreading it over the Internet.
February 4th, 2008 at 1:27 pm
Nick, I’m afraid that I don’t see any evidence in your post supporting your assertion that “Facebook has the most accurate representation of the social graph.” Did I miss it?
February 4th, 2008 at 1:31 pm
Google Social Graph API sucks as much as Facebook. Hopefully these two evil companies will merge and then disappear like Netscape did.
February 4th, 2008 at 2:08 pm
I have no interest in having my social graph exposed to the world. I would deliberately falsify my link info. Theoretically, Facebook has more security.
I also do not like the idea of Data Portability where my info could end up on a Social Network that I did not sign up for. I don’t mind my Facebook or Linkedin friends having it but I don’t want them spreading it over the Internet.
February 4th, 2008 at 2:31 pm
Google Social Graph API sucks as much as Facebook. Hopefully these two evil companies will merge and then disappear like Netscape did.
February 4th, 2008 at 2:35 pm
@Kevin, i didn't really place any evidence. That's a great point. At this point all I have is anecdotal evidence. Are you on any other social networks that have a more accurate representation of your social graph?
February 4th, 2008 at 3:02 pm
Nick, I don't think any SNS has a full and accurate representation of my social graph. But anecdotal information won't help us establish which SNS currently has the most accurate representation of more users' social graphs. In fact, I would think that actually measuring that might be quite difficult. *If* all of the data (XFN and FOAF) are out there then it seems that Google's approach might actually be a really good one. I guess we'll find out shortly…
February 4th, 2008 at 3:35 pm
@Kevin, i didn’t really place any evidence. That’s a great point. At this point all I have is anecdotal evidence. Are you on any other social networks that have a more accurate representation of your social graph?
February 4th, 2008 at 4:02 pm
Nick, I don’t think any SNS has a full and accurate representation of my social graph. But anecdotal information won’t help us establish which SNS currently has the most accurate representation of more users’ social graphs. In fact, I would think that actually measuring that might be quite difficult. *If* all of the data (XFN and FOAF) are out there then it seems that Google’s approach might actually be a really good one. I guess we’ll find out shortly…