In the process of reading what others were saying about Twitter, I read this blog post by Danah Boyd, a well-known social technology researcher. She has a lot to say about the pros and cons of Twitter, how it relates to another social service called Dodgeball, and her experience using it with the other attendees at SXSW last week. She makes one point that really stood out to me. She identifies something that has always bothered me about social technologies. Something that is begging for a solution. Friendlists are too simple.



Danah makes the point that the messages she sends to her friends on Twitter go to ALL her friends, with no discernment of which friends might be interested in the message. This raises the question: aren't friendlists out of date?

My friends are not all the same yet very few social services provide a mechanism for describing friends in any detail. You can often block friends. Xbox Live has the concept of a "reputation" and you can rate fellow players based on how well-behaved they are. On Facebook you can describe how you know a person with phrases like worked together, took a course together, and went to school together. But none of these describes my relationships in terms that can help direct the flow information intelligently.

Susie wrote a post recently that is related to this point. She tells a story of how she spotted Phil McKinney's photo on a gadget blog site. Phil happened to be in the same meeting so she forwarded the link to him and they laughed about it. Susie considers how this process of finding interesting content can be automated and she poses the question: Which is better: user-generated tagging or machine-based content analysis? In her example, I don't think a machine would have been able to identify Phil. For many things, people are still the best analyzers. However, I wouldn't mind having software help me filter my contacts for people who might be interested in what I'm looking at.

Could the solution be as simple as just being able to add tags to our friends? I was recently reading an interesting blog post about streaming content to the Xbox360. I wanted to be able to type "Xbox" into a search field and immediately have the option of forwarding that link to people I've tagged with "Xbox". This is also true for Twitter. Friends could use tags to filter what messages they receive or senders can use tick off the tags of people who would be interested. Tagging is probably the simplest approach to more intelligent modeling of networks.

Let me take the idea a step further, just for the sake of brainstorming. Could a system also characterize our friendships based on the strength or intimacy in the relationship? What if you are considering forwarding some sensitive information that you would only share with your closest friends and/or family? Is it possible to compose a set of questions that can be asked about everyone in our friendlist to help characterize the closeness of the friendship? There are some metrics that help determine intimicy such as

• how often we contact the friend
  
• how many modes of communication we use (phone, IM, email, SMS, etc) to contact the person.
  
• how long our communication sessions last
  

There are also subjective questions that the user could answer. For example, could we ask, on a scale from 1 to 10, how much do you trust this person? If you had a embarassing secret, would you tell this person? Or, if this person were to move to the moon, how sad would you be, on a scale from 1 to 10?

This is probably a job for a psychologist, but there must a series of questions, fewer than 10 if possbile, that could describe the strength or itimacy of the relationship and help model the relationship more accurately in software. This, combined with tagging, could go a long way toward a more intelligent and useful friendlist.

I'm just thinking aloud about it because I think it's interesting and because I think it will need to be addressed. Simple, one-dimensional friendlists are outdated and whoever addresses this challenge in the coming generation of social technology will be a big winner.