From: nobody@shell.portal.com
Date: Thu, 19 Aug 93 14:01:05 PDT
To: cypherpunks@toad.com
Subject: Re: Physical to digital cash, and back again
Message-ID: <9308191211.AA12617@jobe.shell.portal.com>
MIME-Version: 1.0
Content-Type: text/plain


As usual, Hal Finney writes an excellent post on digital cash.
However, some assumptions need to be addressed.

> [triangle model divides the world into banks, shops, and customers]
> Customers would thus primarily turn real cash
> into digital cash which they send to vendors; the vendors then turn the
> digital cash back into real cash.

In its first few years of operation (at least), a digicash currency
will be of highly uncertain value and reputation:

Currency will:
* Be of questionable legality in some jurisdictions.  Once
they catch on, governments will actively suppress it.
* Lack a track record.
* Be able to purchase only a limited number of specialized
information services -- the customer won't be able to use it to 
buy groceries, make house or car payments, etc.

Thus, as a customer, the *first* thing I will be concerned about is
the ability to convert my digital cash into currency that can
be used to purchase physical goods & services in my local
jurisdiction.  Specific application #1 -- an information worker
or contractor being payed in digital cash will want to convert that 
digital cash into local currency, early and often.  Specific 
application #2: customer deposits money into an account to pay 
for online services.  In most cases the customer will be concerned about
refunds and return of unspent funds.  Specific application #3:
on-line casino, with digital cash chips.  Casino customers'
first priority will be the ability to "cash out" with as little
hassle as possible.  

Now we could stretch things and call the person who wants to
cash out the "shop" (especially for case #1), but then the
necessity of a user-freindly system would apply just as much
to shops as it would to customers.  Better to have a symmetric 
relationship between agents rather than dividing the world into 
customers and shops.  Any of these agents may wish to make 
payments in either direction.  In many cases the amount of payments 
will be quite assymetric, but in most cases the ability to make secure 
payments both ways should exist.

This gives us the following model:  (using FRNs as the local
legal cash currency):


					-------
					|     |
			              	v     |
FRN bearer acct <--> d-cash acct <-> agents   |
	^				^     |
	|				|     |
	v				-------
  physical cash


> [customer] sends another check
> to the bank and gets another batch of digicash.

Not a good idea, checks are not private!  Try cash or money
order.