From: "Vladimir Z. Nuri" <vznuri@netcom.com>
Date: Sun, 19 May 1996 05:25:53 +0800
To: tcmay@got.net (Timothy C. May)
Subject: Re: The Crisis with Remailers
In-Reply-To: <adbfebe7040210046482@[205.199.118.202]>
Message-ID: <199605171738.KAA05800@netcom9.netcom.com>
MIME-Version: 1.0
Content-Type: text/plain


so list attention turns once again to a looming remailer "crisis"...


where are there so few remailers? the reasons are pretty obvious.
these problems have been transparently apparent from the very 
beginning.

1. there is no economic incentive.

as soon as there is a good economic incentive to run remailers, you
will see them proliferate. but currently they have no virtually
no value to the creator. it's like building a house for other people
to live in out of humanitarianism. note that with web pages, you
are buying free publicity for your company. but in fact you are
typically buying yourself *negative* publicity by running a remailer.

what is the current incentive to run remailers? answer: adulation
by other cypherpunks. hmmm, not necessarily all that motivating
to very many.

2. there is no good way to deal with spams or other so-called "abuse"

I commend the remailer operators for starting a mailing list to
deal with spam. but the solution remains essentially "stop
spam by hand". spammers still have the ability to be a serious
threat to the network. this has been a threat from the beginning
and has never been resolved. note that "spam avoidance" is a
very, very difficult problem that plagues far more than remailers,
such as mailing lists and usenet. but it is particularly acute
with remailers.

3. liability

there is a lot of liability to the operator of a remailer, and
again, this risk is totally unsupportable from their current
returns (nil). Hal Finney recently suggested restricting posts
from remailers to avoid copyright liability. this will limit
the liability and risk but does not totally remove it.

4. no need for a network

in fact there is not really a need for a remailer network on one
level. there is only a need if the service is not available. why
is there only one anon.penet.fi? well, because of the above reasons,
and also by the fact that only one is sufficient to serve all of
cyberspace, virtually. what I mean is that there is easily enough
traffic to justify another anon.penet.fi type remailer, but it's
not totally critical (i.e. to the point that someone puts their
resources where their mouth is) as long as anon.penet.fi is running.

5. etc.

==

if people want to know why remailers haven't proliferated  in
the same way that other cyberspace infrastructure has in the
past, such as news servers and web sites, you have to focus on
the above issues. remailers are NOT like other cyberspace services.
they are a tremendous burden to run, instead of being of high
use to the maintainer (even though they don't generate cash)
in the way a web page or usenet server is.

the main problem, getting cash for the service, is slowly dissolving
to the point that it will not be an obstacle. I predict that 
remailers (and many other unusual services) may begin to proliferate
at that point-- but not as much as other areas of cyberspace such
as the web. remailers are always going to be plagued by the other
problems I mentioned above unless some really brilliant genius
comes along to solve what seems to be the unsolvable.

another tact the cypherpunks might take to get anonymity into 
the cyberspace infrastructure is to target forum architecture.
instead of trying to create remailers that "feed into" other
networks, why not build in remailers into those networks themselves?
I am thinking of the way NNTP could be a massive anonymous 
remailer network, and that in fact it was once but that this
was purposely designed against in the protocol (preventing people
from anonymously submitting articles to NNTP hosts).

I propose that as long as there are serious elements involved
in building up cyberspace that are hostile to anonymity, you
are not going to see it flourish in the way other services have.
it seems to me the major obstacles to widespread anonymity
are perceptual, not technological. if people can find a way
to handle the above issues and still provide anonymity, it will
spread. otherwise, I doubt it will ever become very "mainstream".
perhaps the above problems are intrinsic to anonymity, which would
be a pity in my view.

BTW, TCM laments that he hasn't seen master's thesis on remailers.
I consider Lance Cottrell's mixmaster work to be really on that
level, and highly commendable. LC has really advanced remailer
technology by tremendous leaps and bounds since putting his mind
to it. also Levien's remailer page is another very outstanding
service. it is possible that all the real research into remailers
is being done at the NSA <g>

seriously, though, I think cpunks have an opportunity to do some
introspection here. it seems a pretty good rule in cyberspace that
"cool and useful services flourish and grow". witness Usenet
and the web. why haven't the cpunks been able to tap into that
kind of exponential force with remailers? the problems are not
merely technological. I would say the technological problems
associated with the remailers are the most straightforward to
solve. its the complex social issues that are seemingly insurmountable.

I really believe that if anyone wants to get more anonymity in
cyberspace, they must deal head on with the sociological 
"anonymity taboo" in society. why is there a taboo in society
against anonymity? could it be there are some good reasons for it?
is it possible to create a "socially acceptable" anonymity? of
course this line of thinking is going to be utterly repulsive
to some on this list, but I contend it is essential to remailer
growth strategy.

of course if people don't want remailers to ever go "mainstream"
anyway, well then there is no problem. the remailer network still
has an "underground" feeling to it and perhaps that will always
be part of its draw, and its actual structure.