Quote from: benjyz on April 17, 2014, 10:20:21 AM

very good insights. I would note that hashing power accumulates in several distinct locations (data centers). The more geographically distributed mining would be, the more it would be owned by different people. unfortunately we can't just require that. imagine a reward that declines the more you mine from one location. miners would make the effort to open up new data centers, but this activity would be unprofitable, so that new miners could make more.

thinking about these things I find very fascinating, also because historically currencies were tied to geography / population, and there are good reasons for that. my idea of proof of location was a replacement for proof-of-work, or perhaps augment proof-of-work. essentially each of the describe segments (what you called cubes) would be tied to the money supply. the more segments get involved the better. one could imagine software for trading between those cubes as well. for example imagine a good being shipped from one location to another tracked by the Internet. each segment has a cost function. so instead of shipping companies you would have a public agent-transport-network. this can applied to just about any productive process.

To curb contest of space which is unreasonable, just tie transaction rewards to where they've been initiated. So if you have users somewhere, providing resilient internet access + mining will be profitable (and make sense).

On the spatial partitioning scheme: one could refer to recursive schemes like BSP or octrees. Then, in places of high use, high density of mining and high competition, the space would be partitioned. Much like the self adjusting difficulty algorithm, only tied to actual meaningful supply/demand.