From: ghio@temp0119.myriad.ml.org (Matthew Ghio)
Date: Mon, 29 Sep 1997 03:02:54 +0800
To: cypherpunks@cyberpass.net
Subject: Re: Remailers and ecash
In-Reply-To: <d9ad723431a59140b039a71c9cecc923@anon.efga.org>
Message-ID: <199709281854.OAA24329@myriad>
MIME-Version: 1.0
Content-Type: text/plain



Adam Back wrote:
> I think someone did write the code.  I think it was Sameer, and I
> think he ran a remailer with this feature for a while.  I expect the
> code is on the berkeley ftp site somewhere.

Yes, it was a nymserver (omega.c2.org) where the user paid up front per
month per nym, and then for traffic over a certain amount.

Lucky Green wrote:
> You certainly could use DigiCash's software with Type 1 remailers. Though
> why anybody would want to pay for a service as insecure as Type 1
> remailers is beyond me. Type 1 remailers should be removed from service.
> 
> Furthermore, the barrier to entry is too high for the consumer. All these
> problems are about to be solved by third party software. Perhaps then we
> will see for-pay remailers.


What people often forget is that these structures tend to evolve thru
incremental changes rather than the immediate adoption of entirely new
paradigms.  The 'barrier to entry', is often the deciding factor.

omega.c2.org was a failure because, despite having features which
(nearly) everyone agreed were desirable, the up-front cost was too
high.  Shortly thereafter, I started a nymserver which, despite having
lower security and fewer features, featured a relatively low barrier
to entry.  This nymserver was so popular that it became the basis for
all the current nymservers.

Type-1 remailers were (and are) useful because they provided a system
which could be operated with low initial investment.  After providing
a proof-of-concept, they became the basis for PGP-enhanced remailers,
latency and reordering experiments, reply blocks, and finally mixmaster.

Various anonymous message pools were tried, and failed. 
alt.anonymous.messages worked because most people already have a
newsreader.

Same goes for blind ecash, DC-nets, secret-shared message pools, etc.
Despite our conceptual ability to describe the technology that we want,
we too often forget the cost of building the infrastructure to support
it.

So, assuming that for-pay remailers are the goal, how do we get there?
If attaching chaum-cash to remailer messages were the answer, everyone
would be doing it.  I think we need to look at the situation at a more
basic level.

Money can be defined as the lowest common denominator thing of value
shared by the greatest number of people.  (Does this describe Mark Twain
accounts?  No.)  So, how else might one pay for remailer usage?  Setting
up your own remailer is the obvious answer, but there are others, such
as selling cpu-cycles.

One potential scenario is "I'll let you use my remailer if you let me use
yours," where people earn remailer credits by remailing other people's
messages.  These remailer credits allow one to send anonymous messages
via other remailers.  Once these remailer-credits become sufficiently
valuable, they can be sold using whatever monetary system is popular at
the time.
 
Another potential scenario is to assume that a.a.m is going to continue
to grow until it becomes unwieldy.  Then people will have to pay someone
to store-and-forward the messages, or split it into smaller pools for
which people can provide their own support by running small news servers
on their machines.

Both of the above scenarios provide the desired end, which is that people
pay for their remailer usage, and neither involves the (unlikely) model
where people attach digicash to their remailer messages using the current
system.