[ExI] <nettime> Forced Exposure: Groklaw closes down

[Eugen Leitl]

tl;dr we must be careful with language and concepts

On Mon, Aug 26, 2013 at 08:05:15AM +0200, Anders Sandberg wrote:

> >The function of a darknet is uncensorable transport and tamper-proof
> >publishing across of untrusted network. It is not about limiting access.
> 
> I use the term in a wider sense than just anonymizing P2P
> communication. Many organisations want to maintain VPNs that are

Unfortunately the term is ill-defined, it warrants a specific itemization
of features in oder to speak meaningfully about the resulting properties.

> uncensorable, tamper-proof and ideally invisible, even if they might
> not want to have anonymous participants (indeed, I assume NSA really

Darknets are not necessarily anonymous. Anonymity in general can
be very weak to rather strong. We do not have provably anonymous
systems, yet. If they're at all feasible, they're probably high-latency,
and not transport efficient.  

> wants to be able to log who is doing what in a tamper-proof way).
> The limited access is an optional thing, but a natural effect of
> making the network invisible (and even more so if identities are
> strictly controlled).

Again, we can make anything we want, provided the set of features
is well-defined (and is self-consistent).
 
> I guess the key question is what the inherent consequences of
> different capability choices are. Uncensorable seems to imply a
> broadcast model, even if there is just a single recipient. Lots of

Well, of course. Nothing is ever free. "Lots" does mean different
things to different people.

> bandwidth usage. Tamper-proof likely requires cryptographic
> protocols, and hence stable nyms. Invisibility requires using

Darknets are impossible without cryptography. Tamper-proof doesn't
mean nyms belonging to single people or users.

Again, we can make whatever we want. It's important to not treat
the whole space of possibilities as something with unified features.

> "normal" protocols or encryption good enough to hide inside other
> communications, or an entirely separate system. Being able to look
> "out" into the wider Internet requires member computers to face it,
> which means intrusion risks and more subtle information leakage.

A darknet co-exists with the wider Internet, or anonymizing systems
*can* (but not must) allow egress into the wider Internet. The
Internet is not friendly, neither are darknets.