October 29th, 2008
Identity Principle: Honor Existing Connections
by Joshua Porter | 
3 Comments
One of the problems with existing social networks is that if you’re not in the network, you are invisible. For example, if I’m in Flickr, Facebook, or LinkedIn and I want to get in touch with my friend Tony, I have to first figure out if I’m “connected” with him in that network. It turns out that I’m connected with Tony in Flickr and Facebook, but not LinkedIn. LinkedIn therefore don’t know that we know each other.
So when I’m in LinkedIn, I can’t connect with Tony. I have to get out of the application and contact him from elsewhere. This makes each site’s social graph only as valuable as it fully describes your real social graph.
At first glance this seems like a real drag…why can’t we communicate with everyone from every site? Then we wouldn’t have to care about which site we happen to be on…we would simply communicate from anywhere.
Then we would have to ask: which social network is going to be that fully-realized social graph? By their own admission, this is what Facebook is trying to do. They are trying to map the entire social graph so that they become the network of choice, the one place that people connect through. If Facebook can accomplish that, then they will have some serious lock-in to their walled garden. People would simply stay on Facebook because that’s where their friends are, not necessarily because it offers a better user experience otherwise. This is the network effect we hear so much about.
But let’s ask this question: is there a offline complement to a complete social graph? Is there an existing place that maps the entire social graph in the way I’ve described above…in the way that Facebook is trying to do?
No, there really isn’t. I suppose you could argue that census information or family tree services might get close, but this information isn’t relied upon in a day to day way in order to make connections. It’s there for reference, but not frequent use.
We tend to keep all this information in our brains. We keep track of all of our connection information there, and we don’t offload it to a third party. But why?
Well, our brains are good at it, for one. Evolution has equipped us with socially-adapted brains that allow us to easily keep track of our human connections (although some of us would probably appreciate being able to remember names and faces a bit better).
But I also think it has to do with how human groups work. People group together for many different reasons: common interests, close location, shared religious views, etc. I presume that new reasons are created all the time…people group in countless different ways.
The mere fact that most individuals have always belonged to multiple groups at the same time suggests that it might be beneficial to have connections in different places that don’t overlap. So, my local web geek friends will probably never meet my family, and my school buddies don’t hang out with my indoor rock climbing buddies. While there are some overlaps, for the most part these different facets of my life exist in harmony as separated concerns. I daresay that if they weren’t separate, then they wouldn’t work as well.
So one of the principles we’ve thought about at chi.mp is to go with the flow, so to speak, and honor the existing diversity of people’s connections. So in the online world this means that your Flickr connections can stay separate and different from your LinkedIn connections, and that’s OK. Those two parts of your life can co-exist peacefully over time and possibly never overlap. This is what we have always done…why should we change things to suit the needs of software?
To this end, we’re building a contact-management system that honors existing connections. When you use your chi.mp contact tools, you’ll notice that you can “import” contacts from various services across the web. But this import doesn’t mean that you forget the site afterward and manage everything from chi.mp, we’ve created it to periodically check on those connections so that when you add or change them in the remote site, then chi.mp recognizes that and makes the adjustment. We’re also hoping that data sharing efforts make this sort of thing common…we believe that honoring existing connections and allowing people to share what they want is the right thing to do.
Here’s an example of it in action. Over time I’ve connected to several services from my chi.mp domain, importing my contacts from each service. In some cases, where there is overlap, friends have been imported more than once. In the following screen I’m using the “manage multiple contacts” screen to consolidate 5 entries I have for my friend Christine. (you’ll note that as the result of a funky Gmail API issue I have two Gmail entries for her):
After I merge all of these separate contact entries, the contact entry looks like this:
Notice that the new contact entry for Christine recognizes that I’m connected with her on several different sites (and we’ve also gotten rid of the redundant Gmail entry). This is important because it allows me to connect with her easily, but also keeps existing connections intact. While Christine does have a chi.mp account, it’s not necessary for me to connect with her, since I’m already connected with her on Flickr, Facebook, and Gmail. A single existing connection is all that is necessary, and all that should be necessary for simple communication tasks.
There are also virtues of having tailored to those specific types of relationships, something you yourself have noted (http://bokardo.com/archives/the-power-of-niche-social-network-sites/). Stopping at non-overlapping groups of contacts implies that Facebook could simply outfit their existing “groups” feature with some additional privacy controls and call it a day. To me, though, that misses the whole point.
In the real world, we don’t have to worry about what “features” are available because, as humans, we can easily coordinate with others and plan activities that are tailored to our various groups of contacts. We can plan knitting clubs, scavenger hunts, church meetings and whatever else, without needing much infrastructure. Heck, we could even get together with our bowling buddies in our driveway and have fun with a ball and some 2-liter bottles; no bowling alley required. Of course it’s still useful to have infrastructure, and on the web it’s basically required.
Having a single site that simply tries to segment contacts from each other is useful, but it’d be like trying to plan meetings with your friends in a stark white room. Only, instead of all your friends being in the room together at once, you can pick and choose which friend to meet. But it’s still a featureless room with nothing useful to aid the interaction.
I think there’s a lot of value in activity-centered design (http://bokardo.com/archives/activity-centered-design/) even in the realm of contact management you’re talking about here. Ideally (to me), there would be more than just a “contact” button. If I had a friend on LinkedIn, I could have a “write a recommend” button, a friend on Netflix could have a “suggest a movie” button, you get the idea. I’m not talking about full integration of these features into a generic site like chi.mp, but just some useful shortcuts to some of that site’s more common social features.
In theory, an ideal world would have some common way of announcing which actions a site provides and what URL to point the button to. In the real world, though, every site needs a fair amount of custom code to integrate with anyway, so there’s little extra effort involved in adding a few useful links. I just think it’d be great to be able to view a contact and get some actions that reflect my relationship with that person. Email’s great because it *always* works for communication, but these niche social networks are so popular because higher-level interactions are incredibly useful. Couldn’t something like chi.mp integrate with those as well?
I gotta stop writing so much in these comments.
Of many things that need to happen in terms of (better) managing our online identities — or should we say, managing our online identity ATTRIBUTES, to differentiate from the identity PERCEPTION cf. Danah Boyd? — this certainly is one. And although it seems pretty straightforward, it’s not the one that I was expecting…
The feature that I am waiting for — somehow the complement of this — is a central place for managing your information, thus avoiding that one has to provide (and eventually update) the same attributes in LinkedIn, Plaxo, Blogger, Dopplr, ecademy, Google Groups, MyBlogLog, Socialmedian, Twine, Xing, etc.
Talking of which - have you guys seen power.com? Identity theft car crash nightmare, or groovy network aggregation idea? Anyone who thinks it’s the latter (as Mashable apparently do) needs their head examining.