There's been a debate raging on Techmeme, the website that turns the tech blogosphere into a mail list, about the value of Facebook.

I have two answers to the question: 1. A lot. 2. Or very little. Depending on what they do.

The next evolution of the web is to deconstruct social networks into their components. I'm tired of building networks of friends, over and over. Next time I do it, it'll be for keeps. It'll be the "real" social network, the one all future social networks build on, just as the format and protocol designed by TBL was the one we all built on for basic machine-level networking.

The "arcs" -- the lines connecting people -- will need to have better labels. And like the Internet, be subject to innovation by anyone, without anyone else's permission. Small pieces, loosely joined.

And the arcs will connect groups of people too. Big pieces that act just like the small pieces.

And there will be an easy way for an app to authenticate someone, and access data private to the app, and data that the user has let the app have access to. That way when I register to be part of a new community I don't have to re-enter all my data again.

If Facebook has the guts and vision to become this network, then it's worth everything, even $10 billion isn't enough. Instead of Yahoo or Microsoft acquiring them, they will be in the buyer's seat. And Facebook has a clear shot at doing it. But there's no evidence that they get this, and no evidence that if they do get it, that they're going to move aggressively to fill the need.

There are plenty of other identity systems out there that could be easily extended to become social networks. Yahoo, imho, made a serious mistake when they made 360 independent of their main identity system. But they did a smart thing (although painful for existing users) when they folded Flickr's identity system into their main system. Same with Google and Orkut. Don't start new namespaces, build off the one you have.

None of these mistakes are new, they've all been made many times over. IBM had the PCjr, when all the market wanted was a cheaper PC. And Hypercard should have been a scripting language integrated with the Mac OS, why have two graphic environments when your strength is the unity of your UI? What value were the rules of UI design if Apple immediately broke them? Unification leads to simplicty. Coalesce instead of fork. This is Postel's Law, still a very good one to follow. Fewer formats please. And fewer logons, and fewer networks of friends.

And Twitter still looks good. Let's hope they keep it simple.