[Sovereign Keys] NameCoin

[Zooko Wilcox-O'Hearn]

On Tue, Jan 10, 2012 at 2:48 PM, Peter Eckersley <pde at eff.org> wrote:
>
> It seems to me that in practice BitCoin has come to depend on a bunch of important servers even though the protocol itself doesn't require it.

Do you mean the servers operating the big mining pools? There exists
an alternative (actually implemented and deployed) which would prevent
such servers from performing rollback attacks:

https://en.bitcoin.it/wiki/P2Pool

It is plausible that this will eventually become larger than the
traditional pools, since it also prevents the servers from stealing
from their clients.

> So I'm not sure how big an advantage the lack of Timeline servers is for BC/NC.

If BC/NC has an advantage here, it is that it doesn't *require*
Timeline servers, rather than that it doesn't *have* them.

This is a useful distinction to make because maybe BC/NC could add
Timeline servers in addition to its current longest-blockchain
technique. (Precisely how to merge the two criteria, I don't know. For
example, the rule could be "longest-blockchain unless a quorum of the
timeline servers agree to switch to a shorter blockchain". There are
probably better rules than that.)

Conversely, of course, a future version of SK could add a
proof-of-work system in addition to the quorum of Timeline servers.

Regards,

Zooko