That is the question I had to ask myself when I first saw OpenSocial. Being the owner and developer of a social site to be, TallyHoh, I am very interested in the possibility of taking advantage of a central friend repository.
Tyler and I have been putting off the “friend” aspect of TallyHoh until last, in favor of developing features, and hoping that someone will come along and develop the Open Friend network. We have come across numerous solutions for this problem like foaf that attempt to solve the friend problem, similiar to how hCard solves the contact information problem, by using microformats. Then there are other people, like OpenFriend who have attempted to build upon foaf and add an aggregation system. Neither of these solutions have really seemed to have taken root yet.
Then Google announces OpenSocial. I have watched the video, read over the spec, and even read other people’s interpretation of what OpenSocial is or could be, but I am still not sure it is open friend.
I am confident it is open widget. So maybe a more accurate name for the project would be OpenWidget, the cross social network widget system. Because at this point I see nothing social about the API other than the ability to tap into each networks user data for the widgets. This seems nice, and a good answer to Facebook Applications, but I find it wanting as a social network API.
I am hoping that the People Data API will be what I want, but it still seems a little to decentralized to me, but maybe I am missing he point. I hope that something will emerge soon, because we are finally ready to start letting people use TallyHoh, but we really need to add a friend system!
Tags: tallyhoh, www
3 responses so far ↓
1 factoryjoe // Nov 13, 2007 at 2:25 am
What’s wrong with XFN?
And yes, OpenSocial should have been called OpenWidget.
2 Daniel Roop // Nov 13, 2007 at 9:42 am
@factoryjoe
I hadn’t looked to closely at XFN, but it does seem like it is almost what I am looking for. Looking at ClaimID, my openid provider of choice, I see they provide all the XFN data I would need on my main claim id page (http://claimid.com/danielroop). However that isn’t tied to my openID (http://danielroop.com) at all. I could be the uber geek and update danielroop.com to include my friend relationships.
Maybe that is the right option, we plan on allowing multiple openid urls per user (since this seems to be the standard), and if we just parse XFN data from each openid, that might work….I will look into it, thanks!
3 factoryjoe // Nov 16, 2007 at 10:50 pm
Actually, all you need to do is a have a link on danielroop.com that uses rel-me and a link on your ClaimID that points to danielroop.com that also uses rel-me and you’re all set!
Rel-me means “this is another link of mine” and since you’d have links on both sites confirming that, yes, indeed, those are both owned by the same person, a spider could start on danielroop.com and know to look at your ClaimID page for your friends!
Leave a Comment