Re: DigiCash

Darren New:
> If I give you a dollar for an online newspaper and the 
> newspaper never gets to me, or I give you $50 for a textbook and the 
> textbook never shows up in the mail weeks later, who do I go to to get my 
> money back?

This is a bad-case scenario, since it involves long delay and
delivery of goods that, unlike information, have a nontrivial
marginal cost to the merchant.  These both give the merchant
incentive to skip the delivery step.  For this kind of business, 
there are several possibilities, none of which  is mature enough
to have actually been deployed yet (which is why ecash is currently 
aimed at small, immediately delivered goods and services).

* Escrow services, which take responsibility for delivery each
way (and develop a public reputation for reliably doing this).
This has the value that it still allows the parties to be mutually
anonymous (although additional tools are needed to actually
have strongly anonymous connections, the payment/delivery protocol 
doesn't preclude it).
* Ripped bill protocols might be able to substitute for escrows.
(these are quite esoteric, but can again be hidden behind a nice
metaphor).
* Deal only with a jurisdictionally reachable merchant and hope the
law is on your side (often not an available solution on the global 
Internet, but usually assumed by identified payment schemes, so
it's a reasonable fallback from that point of view).

For smaller transactions with rapid delivery times, complaining 
loudly and publicly on Usenet will soon put the business out of it.
Receipts will give such complaints more credibility, and will also
allow the clearing agent to credibly get involved.  (Obviously
in these cases anonymity is lost for a particular transaction,
but unlinkability with one's unchallenged transactions remains).
Finally, in a few markets the value add of confidentiality is enough 
to outweigh the risks for even large transactions.

Nick Szabo				szabo@netcom.com

> Darren New:
> > If I give you a dollar for an online newspaper and the 
> > newspaper never gets to me, or I give you $50 for a textbook and the 
> > textbook never shows up in the mail weeks later, who do I go to to get my 
> > money back?
> 
> This is a bad-case scenario, since it involves long delay and
> delivery of goods that, unlike information, have a nontrivial
> marginal cost to the merchant. 

Well, the online newspaper is certainly information. 

> * Escrow services, which take responsibility for delivery each
> way (and develop a public reputation for reliably doing this).
> This has the value that it still allows the parties to be mutually
> anonymous (although additional tools are needed to actually
> have strongly anonymous connections, the payment/delivery protocol 
> doesn't preclude it).

If you can do this, then you don't need an anonymous payment system, do
you? If you want to go through a trusted third-party escrow agent and
that agent doesn't know what you're buying, why do you need to hide what
you're spending it on? How is this different from the bank being the 
trusted third party holding your money in escrow anonymously?

If the escrow agent knows who you're buying from and who you're selling
to, you've lost the anonymity. If the escrow agent doesn't know what the 
goods are and doesn't know who is exchanging the goods, then you don't 
need to pay that agent anonymously.

> * Ripped bill protocols might be able to substitute for escrows.
> (these are quite esoteric, but can again be hidden behind a nice
> metaphor).

I still think that at some point, either you have the money before the 
goods or the goods before the money. Basic Byzantine General problem, yes?
In addition, if it's online cash, I have to clear it with the bank before 
I know it's any good. I read the Certified Electronic Mail paper, and 
(while I didn't follow it all) I didn't see how that solved the problem.

> * Deal only with a jurisdictionally reachable merchant and hope the
> law is on your side

And... how do you know? How do you know that the LLBean website that's 
taking your money *is* LLBean?  (Probably TM's in there somewhere.)
I can set up a web site, run it for three days, and walk off with a bunch 
of money, can't I? How can someone distinguish this from LLBean having 
shipping delays?

> (often not an available solution on the global 
> Internet, but usually assumed by identified payment schemes, so
> it's a reasonable fallback from that point of view).

First Virtual assumes that the seller loses in the case of a dispute. 
Bearer instruments assume that the buyer loses in the case of a dispute. 

> For smaller transactions with rapid delivery times, complaining 
> loudly and publicly on Usenet will soon put the business out of it.

That's assuming the business meant to stay in business in the first place.
Ever hear the one where you wheel the fake ATM machine into the mall for
one day?