[Cryptography] Is it mathematically provably impossible to construct a mechanism to test for back doors in programs?

[Judson Lester]

On Tue, Jun 3, 2014 at 10:03 AM, Phillip Hallam-Baker
<phill at hallambaker.com> wrote:
> A chained notary is only trusted as far as the last outputs that have
> not been recorded by other notaries. So even in the case of total
> collusion, default is very limited.
>
...
>
> Now obviously there is a theoretical possibility that they all might
> collude and default but it is pretty unlikely that they would and it
> would certainly be noticed. I think that is far better in practice
> than the BitCoin block chain with its known vulnerability to unwinding
> transactions.

There was a really interesting point you don't address though: if the
notarization system became untrustworthy, what do you do then? Actors
in this currency system have a significant fraction of their material
wealth pinned to that trust - how do you recover if 16 notaries report
one ledger and the others report a different one? (Or 12/20 for that
matter.) Worse, what if the split ledger is revealed as a falling out
of a colluders, so even the last agreed ledger is in question?

Judson