This is an idea I've been thinking about for some time. A one sentence summary would be that it's a generalization of the Gravatar service.
Motivation (aka what problems would it solve):
- When you something in an eshop for the first time, you must enter your name, adddress, etc. Whenever your address changes, you have to update it in various different services.
- When Google Reader got discontinued, people were trying out alternative RSS readers. But they had to export their data from Google Reader and then import it to the new reader, which is kind of annoying.
- The same applies for calendar events, contacts, bookmarks, avatar, etc.
- It's not just individual users. Reastaurants have to enter their location, phone number, opening hours to services such as Google Maps, Bing Maps or Yelp. When their opening hours chang, they have to update it in all these services. This is not only an additional overhead for the restaurants – another problem is that if someone creates a Yelp alternative with better UI and more features, they won't be successful, because they don't have the data – and restaurants won't bother giving them the data, because nobody uses that service.
- Companies have to send the same job offer to various agencies.
The proposed solution is something like Gravatar for everything. Users and organizations have something akin to a Dropbox folder, identified by a URL. There are subfolfers such as /address, /avatar, /bookmarks or /opening_hours. These subfolders, including their name and structure, are standardized. You can access anyone's folder via an open API.
So for example, if you want to purchase something in an eshop, you enter your folder's URL and then allow one-time read access to /name and /address and that's it.
Third parties can independently create their own subfolder standards. A bunch of Linux distributions may agree on a /desktop-background subfolder. When you install some Linux distro, you just enter your folder's URL and allow read and write access to /desktop-background.
What do you think? It's intended more as starting point for brainstorming then a final solution.
I think this could work if the information was somehow only handled by the browser or the api encrypted the information before users submitted their form. Otherwise any website could be phishing your information when you are browsing.