(1) The average Internet user probably isn’t an active member of dozens of Web 2.0 services. While this may be difficult for some to believe, the truth is that most people don’t feel compelled to sign up for every new Web 2.0 service that launches. And quite frequently, users sign up for services that they eventually end up using very little. Data portability seems a lot less compelling when one recognizes that many, if not most, mainstream Internet users aren’t actively investing their time equally across a wide range of Web 2.0 services.
Actually you're wrong. Data Portability is not about 'Web 2.0' - it's about any web-based service. A typical user might use CNN, Yahoo Mail, Facebook, AIM, their cell phone and their PC or Laptop. That's a lot of apps. Imagine the possibilities of having them sync some aspects of your data.
(2) The average Internet user probably doesn’t need or want to take his friends along to every Web 2.0 service he or she signs up for. These services can be fun and entertaining, but the notion that every user wants to be able to import his data when signing up for a new one is asinine.
Really? I remember the same argument against Telephones, PCs and Cell phones. It's only asinine if you have a failure of imagination.
The point is not what users do today, but rather what new applications and innovation are possible in a standards based data ecosystem.
(3) Privacy is just as important as openness. Where does my data end and yours begin? If you believe that users of Web 2.0 services have some inherent “right” to control their own data but that this data is in inexorably linked to the “social graph,” what “rights” do users have to control where “shared” data goes?
Openness is the wrong word. The DataPortability project does not refer to the 'Open Web' for a reason.
Privacy is also the wrong word. Privacy is too broad a term that has no actionable attributes. We need to focus on words that represent features for implementation. Features that allow Access controls and permissioning for example.
As for shared or derived data, the lines are being drawn and the issues are being debated. Just because it's hard to work out doesn't mean it's not worth trying.