> The basic problem is that chaumian credentials are transferable.
> People who have no use for them will be able to sell them for a few
> £s, and domain name speculators will be happy to buy them. Someone
> who is willing to speculate $70 each on hundreds of domains can easily
> afford to buy a few transferable credentials.
That's a good thing about Anna Lysyanskaya's credential system. Imagine
that credentials were widely used in society, with one centralized
identity registry and each person having dozens of unlinkable pseudonyms
for different situations. In her system, to let someone else use one
of your pseudonyms and credentials, you have to give them your master
private key. This gives them access to _all_ your nyms and credentials.
They could drain your bank account, commit you to harmful contracts,
and do similar mischief.
One of the attractions of the DNS idea is that it might be a "foot in
the door" for a credential system like this. You could take your DNS
registry information and use it in blinded form for other situations.
You could use it for online voting and polls, for example, where it is
costly now to eliminate multiple voting. It would facilitate ratings
systems for Usenet and other online discussion groups, where presently
the difficulty of eliminating vote-stacking keeps these ideas from
getting off the ground.
You could even hope to move into ecommerce situations, using credentials
to access for-pay web sites. Currently these sites rely on ad hoc systems
to detect sharing of access codes, like noticing when the same code comes
from multiple IP addresses. You need something like a credential system
to allow unlimited use access codes which can't be shared with other
people.
Then, you could try to solve the really hard problems: using credentials
in place of the identification information collected by governments and
businesses. Drivers' licenses and other government documents could be
replaced by cryptographic credentials which don't encode your identity.
Economic commitments could be backed up by positive credentials for
successful completion of contracts in the past.
This approach offers some hope of getting off the path to the ultra
surveillance society that we otherwise seem doomed to follow. Obviously
the grander steps are a dream, with long odds against them. This is
why it will work better to start small, with a well defined situation,
and expand out from there. DNS would be a good place to start, since
it is worldwide, fundamental to the operation of the net, and about to
be redone.
