I'm hacking on this regularly enough I'd like to officially announce it here:
https://github.com/tarcieri/cryptosphere I'm shooting for an easy-to-use crypto capability system for a globally decentralized datastore, which more or less means Tahoe + Kademlia. Unlike Tribler, there is no search, because access to all data is restricted by a crypto capability system. This makes the Cryptosphere targeted more as a system for content curators to securely publish content to interested parties. This is the second P2P system I have worked on. The first was a P2P streaming system called DistribuStream: https://github.com/tarcieri/distribustream If you're looking to compare it to something else, it falls within the realm of MojoNation, FreeNet, or GNUnet. As far as what I'm doing different: I consider all of these systems overcomplicated, and have tried to simplify everything but the trust model. My goal is to derive trust as a function of self-similarity computed via a collaborative filtering algorithm. Technology aside: I have a lot of ideas about how to nail the user experience of a system like this and make it accessible to the unwashed masses. I also think I can make it simple to install with minimal requirements which is what I think is the big limitation to people adopting new technologies like this. -- Tony Arcieri
_______________________________________________ p2p-hackers mailing list [email protected] http://lists.zooko.com/mailman/listinfo/p2p-hackers
