From: qzmorgan@hushmail.com
Date: Sat Mar 30 2002 - 02:58:19 MST
Non-member submission from [qzmorgan@hushmail.com]
Sounds like an exercise in smart contract design
(http://szabo.best.vwh.net/idea.html)
The current arrangement in many U.S. cities right now is pretty
simple -- certain emergency vehicles have the right to give themselves
green lights, via a wireless transmission from the vehicle akin to a
garage-door opener. Presumably this is recorded and audited by local
governments to deter abuse. Normal vehicles have no such ability (and
correspondingly no such right). I understand that the security is not too
good and a good cracker can spoof the lights in some cities. Strongly
secure protocols exist to solve this problem.
The ownership of the lights, i.e. who has the final decision of when
lights are red are green, need to be specified. Ronald Coase originated
"transaction cost" analysis -- now a thriving school of
economics which is used when determining what kind of contractual and
ownership structure (ranging from single ownership of all lights and cars
to each light contracting with each other light and each car) is most
likely. There are many possibilities for the ownership/contract structure
most natural for this problem, and it's very related to the issue of whether
streets in general are a "natural monopoly".
For simplicity let's assume for a moment the lights are owned by a single
party
(say, the local government) and each driver contracts independently. Then
how do we design the contract between each driver and the stoplight utility?
Hal's auction suggestion is a great starting point. However, we need to
define
specifically what is being bought. Spike Jones suggests one is buying
"time saved" -- one bids, say, $10 to save one minute, and thus pays $10
if one minute is saved or is refunded $10 if so delayed. This is a great
idea if it's auditable -- if your car sensors and computer can compute how
much
time you saved, and compare it to how much the streetlight utility says
you saved. Drivers and the utility (or the computers under their control)
would, I think, have to agree on a fair baseline schedule for each car, in
order to determine whether they are gaining or losing time versus this
fair schedule. Assuming such a fair algorithm, and a friendly user
interface
that allows the user to change their bid to suit their urgency, and then
displays in a straightforward way the advances or delays and the
corresponding
payments -- if this is all practical, then idea of minutes of advance or
delay is easy for the driver to understand. The resulting system would then
have low enough mental transaction costs to be practical. (Certainly any
product definition and auditing scheme that requires unaugmented
human drivers to understand stoplight scheduling is out of the question --
the mental transaction costs then are far too high).
A remaining smart contract issue is designing a protocol that preserves the
privacy of drivers, so that the stoplight utility and fellow drivers can't
use the bids to keep a dossier on who is travelling where. Crypto
protocols
for this are possible, but may be too time-consuming for a real-time
auction.
Presumably drivers, not car owners, are the persons contracting with the
stoplight utility, so drivers need to log in and out of the cars so that
the driver, not the owner, pays or is payed.
Let's now go back and look at the contractual and ownership structure.
Start with a scenario where each light is owned by a different person.
Would there be a profit in one light owner buying out others instead of
contracting with them? The two most relevant factors here are
"relationship-specific investments", such as the need to connect one
intersection with the next by one street with no alternative routes, and
"contractual incompleteness" -- the inability to predict future demand or
other contract-related contingencies. If investments are specific and
contracts incomplete, a commodity market style of interaction breaks down
and it makes sense to integrate, perhaps all the way to a single monopoly. A
single monopoly seems likely here, but OTOH it might be possible to design a
minimal "night-watchman monopoly" with smart contracts that somehow
eliminate specific investments, perhaps with "virtual" routes, and either
predict future contingincies or render them irrelevant.
For further research on transaction costs and contract/ownership
structure, net search words are the above quoted terms and the
economists Oliver Hart and Oliver Williamson.
Hush provide the worlds most secure, easy to use online applications - which
solution is right for you?
HushMail Secure Email http://www.hushmail.com/
HushDrive Secure Online Storage http://www.hushmail.com/hushdrive/
Hush Business - security for your Business http://www.hush.com/
Hush Enterprise - Secure Solutions for your Enterprise http://www.hush.com/
Looking for a good deal on a domain name?
http://www.hush.com/partners/offers.cgi?id=domainpeople
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 09:13:10 MST