Older blog entries for average (starting at number 1)

How to build the next Internet: make it p2p.

The current Internet is mired by its history of monolithic, server-centric computing centers. Most content now is generated by the users and shared in a peer-to-peer fashion. The DNS system isn't equipped to handle this.

Let's upturn the top three layers of the legacy OSI model to put the Session layer at the top where users are switching gears and locations often, make a new, 3d Presentation layer utilizing a unified data model, and rename the Application layer the content-centric layer which will replace DNS and ultimately the OS. Hellyeah.

All this is being done at pangaia.sourceforge.net. It advances the state of the art in computer science and data visualization (as well as being totally rad) and is looking for people to help code a prototype.

A working paper version has been made as proof-of-concept. It's shows how with a few simple rules it will create a very complex, self-organizing system of social information.

Advogato's trust metric is severely underutilized given the amount of collaborative platforms popping up in webland. What gives?

For all the interest, both academic and pragmatic, it's remarkable how little implementation there is (worth noting is UCSC's wikitrust). The problems do not seem intractible and the problem domain itself is quite fascinating, yet so little to show. Are people just using web site selection to do ranking ("bookmark trust") and leaving it at that? I'm surprised there so little desire for a unified field of discussion and knowledge, it could create a new economy of ideas as well as a marketplace for allocating decision-making authority and governance.