[e2e] We do need anti-spam mechanism for e2e

[Jon Crowcroft]

ah, i was wondering if someone would notice that - we have a solution
(which is in an inceitves scheme for sharing wireless nets, but could
work here too) 

 Modelling incentives for collaboration in Mobile Ad Hoc Networks
Jon Crowcroft, Richard Gibbens, Frank Kelly and Sven Ostring
Proceedings of WiOpt'03: Modeling and Optimization in Mobile, Ad Hoc and
Wireless Networks, March 2003, 

see
http://www.cl.cam.ac.uk/Research/SRG/netos/papers/
for link

idea is a "use it or lose it" control law for credit, but also a "slow
decay up to a unit" of credit for those in debt - this allows sort of
workahead and workbehind (think of it like a slack term in a scheduler)

the goal is to attract sponsorship from people who don't enter the
spirit of the thing, so messages not accompanied by ecash haev to be
accompanied by REAL cash...

an offline conversation has conveinced me that we could SELL this idea
to Visa and other credit card agencies

effectively they become adversiting agencies, and mnanage the escrow
system for us, and manage collecting fees from people who want to send
unsolicited mail to us...that way we all benefit

they have the databases all ready
In missive <200401071411.i07EBGvL008120 at codex.cis.upenn.edu>, Michael B Greenwald
 typed:

 >>   Wed, 07 Jan 2004 10:02:26 +0000
 >>   Jon Crowcroft <Jon.Crowcroft at cl.cam.ac.uk>
 >>
 >>   um, no - it doesnt cost you 22.57$ 
 >>   
 >>   what happens with a payment scheme is that we all put some real money in
 >>   escrow  and we get some e-cash
 >>   
 >>   we use the e-cash to pay for sending to e2e, BUT remember who gets the
 >>   e-cash? the members of the list do. so all members who participate get
 >>   tokens (ecash) but unsolicted mail from outside gets a REAL bill. as we
 >>   accrue real money, we can use it to put in escrow instead of our own
 >>   money. after a while we start to profit from the spam, and we stop
 >>   complaining all the time.
 >>   
 >>   end of problem.
 >>
 >>It seems to me that the incentive mechanism you describe can't work.  
 >>
 >>Won't anyone who sends more than 1 out of 2257 messages to the list
 >>lose money, and not make a profit?  And if one actually *can* make a
 >>profit by listening, then there is an incentive for people to
 >>subscribe to the list simply to lurk (and have a filter drop all
 >>messages without reading them).  As the number of lurkers rise, the
 >>cost to send legitimate messages (as well as spam) rises too, and
 >>makes it increasingly likely that sending even *one* message will cost
 >>you money that will take receiving a *lot* of spam to recover from.
 >>In other words, if legitimate but prolific senders can afford to send
 >>to the list, then spammers can simply join the list and can also
 >>afford to spam the list.  If legitimate senders can't afford to send,
 >>then the list is dead (and I don't see how you can make it hard for
 >>spammers to send if they are willing to join the list in some guise).
 >>
 >>Maybe some more complicated incentives *could* work, (sending costs
 >>more per recipient than the recipients receive? But then who collects
 >>the excess money?  Or, maybe recipients can vote on the spam-level of
 >>a message and senders that exceed some measure of off-topicality will
 >>be kicked off the list before they recover their costs.  But that can
 >>be defeated by joining with two accounts.  Maybe we delay the
 >>repayment of real money to listeners by a sufficiently long time (or
 >>bar withdrawl for some long time), so that the lost interest on the
 >>real money is a disincentive, but immediately remunerate people with
 >>e-cash.  That way, people who want to send must be willing to identify
 >>themselves so that spammers are identified with a real address.  But
 >>...).  Maybe something would work, but now we'd no longer have quite
 >>so simple a scheme as you originally proposed...
 >>   

 cheers

   jon