I've been thinking about a solution to the open social graph problem, and think I've come up with an answer which could work nicely, and has aspects in common with what other people have been suggesting here.
Essentially I propose an extension of OpenID (I call this 'xOpenID') which also stores data on relationships. So as part of every OpenID there is a database of other OpenIDs which the user somehow has a relationship to, alongside some information on the nature of that relationship, most likely by means of XFN. I have written a more general slideshow about open social networking, and this solution is explained in more detail there: http://www.slideshare.net/huw/open-social-networking
Another aspect of 'xOpenID' is the exchange of information between OpenIDs. If there is a mutual relationship established between xOpenIDs, then it would be possible for user A to see which social networks user B is signed up on. It would then be possible to automatically (depending of course on the nature of the relationship between the users) create a secondary relationship between the user A and user B on a given network if both were already members.
The social networks themselves therefore sit upon this 'master social graph', allowing users to keep control of their data and niche social applications to tap into the wider social graph (as is possible with Facebook applications at the moment), but without the caveats and restrictions present on the Facebook platform.
The system decentralised from one perspective (no single company owns all the data), but also centralised in that each user's relationships are stored in one place. I think this achieves a good balance, as it largely removes the need for syncing whilst being truly open.
The only problem with the system is, as I raise in the slideshow, the need for mass adoption by social networks. I believe that this could be overcome if a large network such as Facebook could be persuaded to do so, which is not inconceivable. Ultimately I believe that any solution which is to work effectively will have to create incentives for social networks to adopt it, and these incentives (in the forms of ease of use and access to a wider social graph) will be in place if a major network adopts it.
Yes, if you had a clean sheet of paper to start with, this is clearly the way to go. Rather than duplicating information, and putting the onus on the user to keep their information synched, it would be much cleaner to have the user maintain their info, make it available ala OpenId, and have the soc-net sites subscribe to it (with RSS?).
The issue, as you point out, is getting the soc-net sites to adopt it. There is a school of thought that says that that is impossible, practically speaking. But, I for one would like to see an implementation like this attempted, perhaps with some demo soc-net sites as subscribers. I think if OpenId succeeds, then xOpenId for soc- net portability would be an easy enhancement. There seems to be some momentum in that direction.
Really, a system like this would be a great benefit to soc-net sites because they would gain access to a much, much wider group of potential users through the web of individual social graphs. Ironically, all these "Web 2.0" sites are resistant to giving up what they perceive to be control and ownership of their users. Ass- backwards, IMHO.
I would be very interested in participating in any effort to this end.
> Ironically, all these "Web 2.0" sites are resistant to giving up what > they perceive to be control and ownership of their users.
It's the same old story. Think AIM, ICQ, MSN, and Yahoo for IM. Think Skype for VoIP. The folks who finance these things want captive eyeballs. The only wonder is that email is decentralized, but I suppose that happened before big money started paying attention to the 'net.
Based on my experience fighting for open, decentralized IM in the Jabber world (8 years and counting to break down the silos of the big consumer IM services), I'd say that the only hope here is to starting building out an open infrastructure for decentralized social networking -- open protocols, lots of open-source code to help folks get started, etc. Eventually some of the big boys might join, but don't expect them to do so out of the goodness of their hearts -- expect them to do so once the open SN network gets so big that they'd be foolish not to open their silos up to that network.
> The issue, as you point out, is getting the soc-net sites to adopt it.
It struck me that rather than trying to write something that Facebook would adopt (for example), perhaps a bootstrap strategy would be to write something to which Facebook application developers could "port" their apps. I have only browsed the FB API docs, but I think in many cases, the amount of services a FB app requires from FB boils down to simply getting your list of "friends".
Certainly, FB app developers are by nature early adopters! And even if a lot of the apps are kind of silly, they are real examples of how people want to leverage their social graphs.
I agree about the social networks not joining in out of the goodness of their hearts. Once the open social graph reaches a critical mass, individual social networks would be stupid not to join in. The problem is getting it there, and I think the best hope would be to put a big social network (probably Facebook) under as much public pressure as possible. In the coming months and years, there could be a growing public concern about data silos. The problems of a closed social network are not hard to understand, and the young demographic which use them are likely to be more liberal, and therefore more bothered, than other groups.
Of course this might not happen, in which case I must accept that there is little hope of this, or a similar architecture, ever taking off.
There's also the small point that 'xOpenID' doesn't exist yet, and I'm not a developer...
Huw Leslie <huwles...@gmail.com> Fri, 31 Aug 2007 15:16:37
>I agree about the social networks not joining in out of the goodness >of their hearts. Once the open social graph reaches a critical mass, >individual social networks would be stupid not to join in.
A good test of this is to watch for when they support OpenID. And note that Plaxo (which is pretty big) now does OpenID.
-- Julian Bond E&MSN: julian_bond at voidstar.com M: +44 (0)77 5907 2173 Webmaster: http://www.ecademy.com/ T: +44 (0)192 0412 433 Personal WebLog: http://www.voidstar.com/ skype:julian.bond?chat *** Just Say No To DRM ***
|
