On Mon, Nov 25 2013, coderman wrote:
On Mon, Nov 25, 2013 at 6:12 PM, David Vorick <david.vorick@gmail.com> wrote:
... Nobody cares how many petaflops the network is pulling, because the petaflops can't be put to use somewhere else... But if the mining was based on cloud storage, a dramatic drop in the price of the currency would result in a dramatic drop in the cost of storing data on the network.
i like the idea of "proof of _useful_ work" applied here to storage. if only mining had been applied to BOINC, GIMPS, or *@home efforts...
surely there is prior art?
I just remembered, there is in fact prior art, though it's state of the art, as in "nearly practical." They're called SNARKs: Succinct Non-interactive ARguments of Knowledge. You can take any computation and annotate it sufficiently that whoever runs the computation can generate a (constant size) proof that they did it correctly in (nearly) constant time. Andrew Miller posted about it on the Tahoe-LAFS mailing list. See http://eprint.iacr.org/2013/507.pdf for one implementation. I'm not sure if you could use this for a Bitcoin-style problem, though, since I have no idea if the difficulty can be adjusted smoothly. It's interesting that we can now take ALL computations problems and turn them into the "easy to verify" variety, however. -- Sean Richard Lynch <seanl@literati.org> http://www.literati.org/~seanl/