This is a delayed entry about a small session I held on Reboot about social networks. It ties in nicely to a recent series on Four Starters about trust and how friends are a solution to this.
In this article I will lay out why social networks are too important too leave in other people’s walled gardens and I will lay out a tentative way to connect the gardens and cultivate your own using microformats and other open standards.
Feedback is greatly appreciated.
The buddy list is key
Social networks and your online identity were prominent in various Reboot talks this year. I lifted Stowe Boyd’s quote: “The buddy list is the center of the universe.” (see slide). This has been true always but is ringing more so as the web matures and we are seeing the breakdown of centralized application models.
In a recent article, Dare Obasanjo wrote about the same subject and how he thinks that Facebook is going to be a big driver in this space. Maybe for America, but I don’t see this happening for Europe in the near future.
And does it seem like a good idea to have all the social information of the world under the control of one company?
But as Dare says, knowing who I know and who I trust, whether that information is in your address book or in your IM application, is usable in other contexts and can greatly improve the trust and interaction in those contexts. You apply the wisdom of crowds to a subset of people —the people you know— to circumvent the trust breakdown Reinier wrote about in current sites.
Yet another social network
My session was after that of Willem Velthoven (write up) who talked about their anyMeta social networking application. Their angle is to reduce the duplication of effort and enable sharing of data. Willem spoke about how he got quite sick of filling in his profile on every social networking site he wanted to participate in.
I have the same feeling and that is mainly what stopped me from filling in my Facebook profile, that and the fact that nobody I know is on Facebook. I could not bring myself to fill in another profile and try to get all my friends onto a new social network just because it looks like the next big thing.
Putting your profile information into these closed social networks gives them a lot of value but you rarely get the option of retrieving that information or using it elsewhere. The facebook API is an exception and is one way of getting access to data while it stays firmly in the silo. At what terms you can get at the data and what you can use it for is firmly in the hands of Facebook.
Microformats to the rescue
What I propose and what we talked about during my session is the concept of Portable Social Networks enabled by microformats. A social network that is open and readable and can be with you anywhere you want. During the Mediamatic session we discussed the use cases and the issues that would crop up and I invited people to attend my session for a discussion on the technical aspects.
We had a brief chat afterwards with interested people and after a break I was joined by some people among whom Willem Velthoven and Jeremy Keith (his post) to talk further about the technical stuff. I got the impression that some of the people attending my session were happy —maybe relieved even— that there was also a technical session to be found on Reboot.
I will go into deeper technical detail in a following post but the concept is to use microformats to markup most of the information found on a typical Myspace, Facebook or Hyves (the Netherlands’ most prominent social network) profile page. This way you can either get the data out from supporting social applications or hook into the network by hosting your own identity web page with correctly formatted data.
So what can we do with technology we already have (POSH+microformats)? As it would seem, a lot:
hCard
Your hCard can contain most of your personal information including your personal details, your address and your picture. This is equivalent to the personal details and picture which are usually listed on any given social network.
XFN
Most social networks have a prominently visible list of friends and a count of your total number of friends. This is very easy to implement using XFN. You can markup the links to the people in your network with the rel=”" metadata which XFN defines. XFN allows you to hook into the network from anywhere. So an XFN link from suppose Hyves could also lead to a profile page on another social network or a self-hosted one. Or at least, that is the vision.
This way you can also differentiate between trust levels withouth using numeric values which Reinier also talked about would be necessary. I am bound to trust a friend more than a contact and you could derive more relations like that.
You could also link to other sites where you have a profile or store data such as your Flickr account, your del.icio.us bookmarks or any other site. The rel="me" value could be used for this, but it is required to be symmetric, so those other sites would have to link back using the same rel="me".
hResume
A lot of social networks also allow you to markup your current job, sometimes your previous places of employment as well and in many cases also your school history so you can get in touch with former school friends.
This information is very similar to the information you can markup using hResume. You could only present the information you want to share in a casual social network but you might want to enter a full hResume to provide all the functionality of professional social networks such as LinkedIn or Xing.
hReview
Most social networking sites and even Flickr have you keep lists of your hobbies, favorite music, movies and books. You could easily mark these up as hReviews and have the fn be an URL to a generally known catalog for that item. For books I would say something such as Librarything or Amazon (with an associate ID!), for music maybe Last.fm and for movies probably IMDb.
This also solves the problem Willem suggested could arise when we use different (language) titles for the same object. Just link them all to the same uniquely identifying resource.
OpenID
The microformatted information listed above can be on any page you want, in fact it probably already is if you have a Flickr or a Twitter account. I do think that there is a case to be made for linking this to your OpenID.
OpenID solidifies the notion of identity online and creates a place for everything to come together. It gives you a URL with which you can refer to a person and you can be reasonably sure that the person and the URL belong together.
Applications that will want to use this kind of information will probably already ask for your login credentials. Those credentials could very well be an OpenID from where on you could automatically retrieve a load of information as described above.
Consolidating or delegating
The question is: do you host this information yourself or do you have someone else do it for you? Since I already host my own OpenID at http://alper.nl, I have already started by embedding an hCard there and I am building the MySpace-esque portal page with all my information on there (preview).
Not everybody will want to host this themselves, but it is analogous to OpenID. Anybody can self-host their OpenID but they can also use a hosted version and the same for the providers. If at any time you want to switch OpenID providers, just change the reference.
The various internet sites which host your profile information such as the social networks and the profile sites such as MyBlogLog or 30boxes need only to mark up their essential data with these standards for it to become instantly accesible and portable.
Even for those hosting this information themselves, it would be nice to have some sort of interface beyond editing the HTML yourself. For instance it would be nice to have an ‘Add as friend’ button on your own site which would ask for your permission to add somebody to your XFN list.
Completeness and Clients
You can make the markup as rich as you want or you can leave stuff you don’t want to share out. Adoption and convention is completely up to you. There are advantages in adhering to a certain standard, but smart clients should be able to deal with holes in this picture.
Below are some use cases which can be realized today already and will also work if not all of the information is present.
Use case: Get an Avatar for somebody
I already talked about this in a previous article (“OpenAvatar - Combining OpenID and hCard”). This concept is just an extension which loads more data onto the page. If a page —such as an OpenID page— contains an hCard with an associated picture, you can retrieve it.
I already wrote a parser as a webservice which takes a URL and returns the associated picture. This parser can take either my OpenID or my Flickr profile (which contains an hCard). This way you can get an avatar for someone that they can manage and update to their own liking.
This concept had already been brainstormed on the microformats wiki.
Use case: Get registration information
A lot of information you need to fill in once during registration such as your full name, date of birth and some other stuff can be gleaned from the hCard. The site getsatisfaction.com already offers to scrape this information from an hCard supporting profile when signing up, saving you the trouble to fill it in.
Flickr lets users list their preferences in music, literature and cinema on their profile page. This listing could be marked up as an hReview with a rating of 1.0 on a worst/best scale of -1.0 to 1.0 (like is 1.0, dislike is -1.0). Then I could reuse it on all the social networking sites that want to know my favourite movies.
Use case: Find out who I trust
The stuff Dare and Reinier talked about with building trust networks and using that information can be realized by walking the XFN web.
Any site imaginable can be improved by adding the knowledge of my network. Imagine IMDb which shows you which movies your friends have recently watched. Or anything really, and all this without having to add your friends on every such network.
Wouldn’t that be a dream?
