THE NEW XANADU® MODEL FOR THE WEB

We see the World Wide Web as a diabolical dumbdown of our original hypertext idea.  HTML caught on because of its shallow simplicity: it built on the existing computer convention of hierarchical files (deeply ill-suited to document structures-- as anyone can now see), and ignored the fundamental issues of version management, rights management and the origins of quotations.

The Xanadu method has always been very simple: make content available with certain permissions; then distribute and maintain documents simply as lists of these contents.  This means everyone can re-use all content virtually, simply by listing the desired content.  Since links are between the addresses of these contents, links are intrinsically bidirectional, may be made by everyone, and may overlap in vast numbers.  (This of course squares the search problem, but Hey.)

Our last several years have been concerned with figuring out how to move these concepts to the twisted environment (from our point of view) of the World Wide Web.  The new design (unrelated to the previous Xanadu code*) is now being written up.  It consists of:

• A new file type for virtual content (tentatively called .XVF)
• Browser plug-ins to fill in and present the .XVF file
• Editors for the .XVF file
• Servers delivering portions on request (using existing protocols)
• Overlay/Gateway micropayment system for content
• A deep cache for keeping content, maintaining multiple names

For reference, this design and these formats will be called OSMIC2 (Open Structure for Media InterConnection 2).

These formats will not be proprietary, in that they will be fully documented for anyone to write to.  However, we reserve the right to build them forward in successive versions according to our beliefs and the emerging politics of the Net.

---