From: hughes@ah.com (Eric Hughes)
Date: Fri, 22 Jul 94 20:16:40 PDT
To: cypherpunks@toad.com
Subject: Small transaction amounts
In-Reply-To: <9407221954.AA14906@smds.com>
Message-ID: <9407230254.AA00519@ah.com>
MIME-Version: 1.0
Content-Type: text/plain


   Right...I think.  What has to scale is the "semantics of money."
   Within a small area ("box"), security is guaranteed by how the enclosing
   system works, and over a larger area it's done by crypto 

There are several ways to make the boundary porous.

1.  Differing rates of clearing a smaller system to a larger.  I can
clear to a larger system once an hour, once a day, once a month, etc.
One can keep a risk bound steady in a system with increasing
transaction flux simply by increasing the rate of clearing.

2.  Probabilistic verification.  In a system where verification is
used, the transactions at the low end might be certified in real time
at some rate.  This decreases the cost of provision while keeping an
eye out for the upper bound on risk.

3.  Net settlement.  A system where one can both add and subtract
value can clear periodically only the net difference in funds.  Net
settlement works really well for small scale systems, but systemic
risk increases proportional to system size.  

4.  Exposure caps.  In a net settlement system, there might be a
maximum positive or negative balance that would be permitted before
clearing to another system was required.  Futures markets have rules
similar to this.

5.  Intraperiod overdraft loans.  A "daylight overdraft" is a running
net negative balance in between clearing times.  By charging for this
money as a short term loan, there is an incentive to minimize its use.

There are more, certainly, and any student of financial markets could
name another five without too much thought.  There are some
interesting and significant issues involved in verification of some of
these policies.

Eric