Why the social data sharing problem can’t be solved
Sunday, September 9th, 2007The funniest part of the Data Sharing Summit organized by Marc Canter was the way everyone kept using the word “universal.” As in a universal API based on a universal database schema that will allow social network data to move freely from site to site. While this is an admirable goal, it ain’t gonna happen. Most of the attendees were engineers, so they believe in a mechanistic model of the world. They see social data portability as a machine to be fixed. That is what engineers do, after all. The problem is that social networks are biological systems, and living things resist mechanistic solutions. There is this tricky issue of emergence, where the right combination of chemistry (some might say the hand of God) allows a lump of flesh to beat as a heart, or think as a brain. Trying to come up with rules to explain or control this, or in the jargon of this meeting, a schema, is reductionist thinking. It’s as futile as predicting how a new drug will react in the body using physics.
Saying you are going to “solve” social data sharing with software is as naive as saying you are going to “cure” cancer. You can treat cancer. You can find a way of eliminating a single form of cancer in a single type of tissue. But you can’t find a universal cure, because biological systems have this annoying habit of mutating on their own. Sometimes they even do so in response to attempts to control them. The best you can hope for is a series of small victories in an ongoing struggle. This is what the software world should be trying to do with the problem of data sharing between social networks.
When I tried pointing out the incredible complexity of what people were attempting, they said things like, “Sure, it’s a moving target, but we’ll do our best to keep up.” Wrong metaphor. It isn’t a ball flying through space, obeying Newtonian laws. It is a living, thinking organism. I don’t mean the people writing social network software. or the code itself. I’m talking about the users, who construct new social behaviors all the time. Let me use one simple example from the summit. During a session called “What is the problem?” (the best session, in my opinion), someone brought up the issue of lying, as in agreeing to be someone’s “friend” when you really don’t want to see any of their updates. How can you communicate that type of behavior in XML or RDF code?
I’m not saying solutions are impossible. They just have to be much less ambitious than were being discussed here. Maybe Marc should have invited some politicians. Politics is the art of the possible, and that is the only thing we can hope for when it comes to social data sharing. Marc is a total pragmatist, and a consumate politician, so I’m sure he agrees with me, but as a politician he keeps rolling the ball forward towards his goal. So he ended up humoring the engineers as they spun out their acronyms, while trying to get as many of them in the same room as possible to win some small victories. All he cares about is getting something that works now, and there I agree with him totally.