From: nobody@soda.berkeley.edu
Date: Tue, 24 Aug 93 18:55:32 PDT
To: cypherpunks@toad.com
Subject: Re: Digital Gold
Message-ID: <9308250154.AA01671@soda.berkeley.edu>
MIME-Version: 1.0
Content-Type: text/plain


At  4:23 PM 8/24/93 -0400, hnash@mason1.gmu.edu wrote:
>
>I tried to imagine a digital currency which is not backed by 
>any bank, but just exists by mathematics and convention, like 
>gold.  The result is the following currency system which 
>could be called digital gold.  It involves three conventions, 
>(1) a convention for valuing coins, (2) a convention for 
>claiming coins, (3) a convention for transfering coins.  
>
>I believe the resulting currency is unforgeable, 
>uninflatable, and untraceable.  Let me know where I've gone 
>wrong (gently).

I love this scheme, but I have two questions about it.

1) How will disputes be settled?
>In order to claim a digital gold coin, the claimer must 
>publicize a "claim certificate", containing the signed 
>integer and the public key required to recognize the 
>signature.  The first person to publicize a claim certificate 
>will be recognized as the owner.  
Some variant of the keyserver could be the public registry of gold. Can
this be done in a reasonable way, where a large part of the transaction
cost is not involved in verifying that the buyer really owns her cash? Is
there some way to make this more like a web of trust? Can it be blinded in
a way that maintains the trust but allows anonymous transactions?

2) Why should anyone value a digital gold coin in the first place? Money
has value because of people's expectation (based on experience) that if
they take it to a store they can buy stuff with it. Where will digital gold
obtain its bootstrap value? Somebody has to start using it.

My own feeling about how digicash will develop is that it will be added
into an initially non-anonymous digital based debt clearinghouse. With PGP
(etc.) we already have the means to create verifiable IOU's, contracts,
loans, etc. (Is anyone using them?) If such IOU's were used widely but
non-anonymously, it would require only a small innovation to move to
anonymity (either through blinding, or use of agents with reputations, or
?). The problem is getting any kind of online action to begin with. 
(AMiX?)