By Dennis D. McDonald

Successful system operation frequently depends on the quality of the data it contains. Social networking systems rely on the ability their members have to manage and keep up to date information about their identities. They also rely on the ability to describe and act upon data about relationships with other network members. If identity or relationship data are faulty, unstable, or inconsistent, the operation of the social network, and the performance of network based transactions related to it, will suffer.

I thought about this when I read the announcement of the Facebook Connect service. Facebook Connect allows Facebook members to log into other participating web sites and share selected information about transactions with that site with their Facebook “friends.” This announcement follows closely on MySpace’s own announcement that it will start using OpenID to support network logins across a variety of participating web sites. (Brian Solis says that Facebook Connect is “practically a competitor” to OpenID.)

Jeremiah Owyang described a typical Facebook Connect “use case” in response to a comment I made on his blog. First, here is my comment to him about his description of Facebook Connect’s significance:

Does this mean, now, that when you, Jeremiah, purchase a product at a participating Facebook Connect website, that when I log into that web site using my Facebook ID, that I will get a popup that says, “Hey Dennis, did you know that your friend Jeremiah — or at least someone using Jeremiah’s Facebook ID — just bought a Widget here? Wouldn’t you like to buy one too?”

This is how Jeremiah responded:

Dennis

that’s one possibility, but it sounds like most will be using it in this use case:

Login to Facebook on a third party site. Do some actions on that third party site (buy something, favorite something, vote for something)

Then, it will ask you “do you want to share this with your FB network?” If so, should it be a big, small, or text update?

Users will have the choice before first submitting it to your network.

As long as you have the choice, I can see how this type of information sharing could be useful. I see something similar in services like FriendFeed where one can watch, for example, what others are posting to their blogs, to del.icio.us, or uploading to Flickr.

Facebook Connect extends Facebook’s identity management system to other participating sites and allows for transactions to be shared with “friends.” It’s the issue of how to define “friends” that concerns me. Compared with the nuances of personal and business relationships that exist in the real world, the ability we have to express — and keep up to date — the nature of real relationships online is still pretty primitive. For example, different networks have different definitions and types of relationships including friends, fans, connections, etc. I and many people, for example, have designated people as “friends” in Facebook that are, in fact, friends; we’ve also designated as “friends” people we’ve never even met. Will a system such as Facebook Connect allow people to make finer distinctions about different types of friend relationships and what that entails for information sharing? Or would this make the system’s operation entirely too cumbersome in its support for website-to-website portability of certain types of data? (I may feel comfortable letting my “real” friends know about some type of web based transaction, while I would prefer that my “friends in name only” be kept in the dark.)

A lot of this has to do with the type of information people really want to be portable from one network to another. Another issue is how detailed and/or cumbersome the system needs to be in order to manage that information. For example, it’s one thing for a network to include in its signup agreement statement, “All your information belongs to us and we can share it as we see fit.” It’s another to say, “We will allow you to make data sharing decisions on an individual case-by-case basis.” Such variability is one of the reasons for the complexity surrounding the DataPortability Project as it continues to wrestle with its vision, its governance, and its eventual deliverables.

While I am optimistic that our ability to model and describe person to person relationships online will continue to evolve, I also think we need more open debate about what information we want to be able to share online. Another important question is how much control it is reasonable to give individuals over what is shared. On the one hand,. we need to protect individual privacy. On the other, we don’t want systems to be so cumbersome that they aren’t workable.

Copyright (c) 2008 by Dennis D. McDonald