From: phelix@vallnet.com
Date: Mon, 29 Sep 1997 20:54:59 +0800
To: cypherpunks@Algebra.COM
Subject: Re: Remailers and ecash
Message-ID: <342f598c.15227348@128.2.84.191>
MIME-Version: 1.0
Content-Type: text/plain



On 29 Sep 1997 02:10:39 -0500, "Robert A. Costner" <pooh@efga.org> wrote:

>
>At first I thought some of the stuff Monty Cantsin was discussing was
>interesting, but it has gotten out of hand.  I've asked the question
>before, Why would remailer operators want to accept Ecash?  After seeing
>the conflicting messages coming from Mr. Catsin, I to rephrase it, why does
>*Mr Catsin* want remailer operators to use Ecash?
>
>The first answer I got was so that remailer reliability could increase.
>The theory as I understood it was that remailers were run like a hobby, not
>a business, so the money would be an incentive to bring in professionalism.
> So I proposed, and documented, that the minimum level to achieve this
>would be a $50K investment over a year's time.  This is in fact, more or
>less what the Cracker remailer takes to run.  Most of the resources are
>donated in some way, but this is their equivalent retail value.  Even so,
>with Monty's pricing structure and Cracker's current level of traffic it
>would be enormously profitable.
>
>Then Mr. Cantsin seemed to go back to the all you need are some spare parts
>theory of remailer operation.  Enormous profits of $200 per month, or even
>$5,000 per year.  Well, Cracker handles close to 25% of worldwide remailer
>traffic of it's kind[1].  And it's not much.  I would say this is due
>mainly to the user interface.  Making remailers more difficult to use by
>adding Ecash is not going to increase traffic significantly.
>
>  >I recommend that people start pricing at a quarter per message per hop
>
>In my opinion, and I've been known to be wrong, this is a seriously messed
>up comment.  A quarter per message is too much, much less a quarter per
>hop. A price of 1/100 of a penny per message is closer to a proper
>valuation. But the problem here is in the pricing model.  It should not be
>transactional unless to encourage the very casual user.  A pricing model
>should be flat rate.  One price for a month, or even a year's service.  The
>net is based on a peering price structure, not an inter-lata structure.
>Trying to compute or add charges at each hop is against the nature of
>information flow for the net.
>

Why?  Each hop adds value to the message (in this case, increased
anonymity/security).  Shouldn't a person have to pay more for increased
value? 

Also, how would a user pay per-month to each remailer and maintain total
anonymity and ease of use.  Flat rate pricing requires one to open an
"account" with each remailer he would ever use before sending the first
message.  Then at each hop, the message is linked to some "account" on the
server which is someway linked to the user (specifically, a set of messages
going through a server can be linked to the same "account", which to me is
totally unacceptable).  This seems overly complicated and could compromise
the users anonymity.

Finally, with flat rate, what would prevent a person from spamming?

>  >We are talking about how to get a working payment system up for
>  >remailers which gets us great service and provides privacy and
>  >security.
>
>The point I have never gotten past is how you expect a payment system to
>change the level of service?  The next point I'm still shaking my head over
>is what about remailer services is not up to your standards?  The only
>thing I have heard you mention is latency, which is a feature programmed
>into the remailers.  If anything, people would pay to add latency, not to
>take latency away.
>
For me, the main problems I see are the number of remailers available and
spam prevention.  Going commercial solves these problems.  Also, there will
still be free remailers around.

>  >Remailers are used by a small highly
>  >specialized market of perhaps a few hundred people.
>
>There is some truth in this statement.  But there are also remailers run by
>a variety of companies such as hotmail, juno, and the like.  They encompass
>millions of users.  Millions of users who want a remailer, but will not
>tolerate the level of entry required for a Type-I or Mixmaster remailer.
>Until client software can be improved and made as easy to use as an
>integrated spelling checker, the "advanced" remailers will have no true
>market share.  (Oh, I forgot.  Most of the world uses email clients without
>integrated spell checkers.)

Agreed.  I've just finished going through the Eudora API docs and I think
sending a message through remailers can be as easy as clicking one button
on the toolbar (just like with signing a messages with the pgp plugin).
For everyday use, everything else (like choosing the remailer chain) can be
automated.

-- Phelix