Given my use case (finding what you wrote), there's a route here. Standards and services that have taken off really fast in the last few years have some common characteristics. Start with self interest. Whatever it is has to give immediate value back to the person implementing it. Generate additional value from social activity. Feed that back into the self interest. I think this is why many of the solutions that require ubiquity to work, fail. They don't appeal to self interest in the first instance.

So. Let's say our identity metasystems have one component which is the automated equivalent of the AboutMe page at a known URL. We then need to have an HTML representation of some or all of the AboutMe data also at a probably different known URL. We'd then put in a protocol for auto-discovering the Identity URL from the HTML URL. Now we've got something that is of immediate use but with social, viral potential. So the next step is to have a lightweight protocol which is to encourage individuals to put their AboutMe URL into their signature everywhere they post. They get an immediate payback that they have a marker they can search for. All their posts can now be collected using one of the search aggregators (Google, Technorati et al) by searching for the marker. But this also encourages services that make use of the identity metasystems to automatically embed these URLs and to make further use of them. And it provides a route for new types of spider to crawl the emerging data cloud.

I think both LID and OpenID have this type of thing built. I'm not sure about others. We tried to do things like this with FOAF in FOAFnet but got hung up on the problems of working with very large quantities of RDF. One of the bigger issues here is having a common format for typical AboutMe data. The big adoption challenge is to get systems with large numbers of AboutMe pages to automatically embed the Identity URL and support it.

Notice the way this turns Identity systems on their head. Instead of focusing on small numbers of big systems verifying the identity of their users we're building a tool for large numbers of users to provide their identity to lots of systems. Everyone should run their own Personal Identity Provider (PIP). And solving the problem of "what did I write where" is the leverage we can use to make it happen.

This is why I'm passionate about simplicity and support for the bottom end and rant at the high end. I want to get a PIP into every Drupal, phpBB, WordPress, MT, etc installation. And I can't do that in the short term unless the technology requirements are very low.

[ << Approaching a definition of Web 2.0 - The Social Software Weblog - _ ] [ Dave Winer's latest OPML browser >> ]
[ 29-Sep-05 12:16pm ] [ ]