James A. Donald wrote: >> The ideas used in Tahoe are useful tools that can be >> used to solve important problems.
Ivan Krstic' wrote: > Yes, and I'd be happy to opine on that as soon as > someone told me what those important problems are. <http://www.waterken.com/dev/YURL/Name/> More generally, we need an egocentric authentication system. The vast majority of users believe that if they "successfully" login, this is proof of relationship. Obviously we should make that true. When you receive a message that purports to be from someone that you have a relationship with, the relevant question is not "is the globally recognized true name of this entity "First National Bank of Omaha", but "is this an entity you have a relationship with, if so, how is this entity known to you, and how are you known to this entity?" In pre computer days, a business communication: 1. Could only be read by the sender and the recipient. 2. Identified the sender as the sender is known to the recipient, and the recipient as the recipient known to the sender. 3. Identified the primary key of the recipient's file at the sender, and sender's file or issue file at the recipient. (And if no valid recipient key, the communication is low priority) In the computer age, we need to have cryptographic enforcement of all this. Cryptographic enforcement of true names is not particularly relevant or useful. In the days before banks became part of the government, ordinary middle class people used a variety of financial instruments, called "notes" or "bills", wherein a communication from Ann to Bob could contain a capability that Bob could use with Carol, thereby granting Ann a corresponding capability with Bob. The "cheque" or "check" is the degenerate remnant of these once diverse and multifaceted capabilities. We need to recreate this ability to transfer capabilities in organization to human communications and organization to organization communications, starting with, and most importantly, capabilities representing money or promises to deliver services on demand. _______________________________________________ tahoe-dev mailing list [email protected] http://allmydata.org/cgi-bin/mailman/listinfo/tahoe-dev
