<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
</head>
<body text="#000000" bgcolor="#ffffff">
On 05/01/13 16:12, Graham Burnside wrote:
<blockquote cite="mid:50E850FC.50402@gmail.com" type="cite">
<meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
<title></title>
On 05/01/13 15:20, Jonas Smedegaard wrote:
<blockquote cite="mid:20130105152049.30178.98668@auryn.jones.dk"
type="cite">
<pre wrap="">Quoting Graham Burnside (2013-01-05 15:15:05)
</pre>
<blockquote type="cite">
<pre wrap="">I've had a couple of hours recently to catch up on how freedombox is
developing. I was reading through the debian wiki and noticed that the
method being used to resolve the IP address of nodes in the network is
via Tor hidden services.
</pre>
</blockquote>
<pre wrap="">It is true that some are discussing how to penetrate masquerading
firewalls and how to discover nodes. Also true that some find Tor
suitable for handling those issues, and actively explore that option.
Some of us, however, consider FreedomBox a project to mainstream
privacy-related functionality that is already mainstream among geeks.
With "mainstream among geeks" I people who install packages with
aptitude (or apt-get) and generally rely on Debian to provide sensible
defaults for those packages.
I believe Tor is only relevant to consider including into FreedomBox
when someone can provide sensible defaults for it which can be
integrated with the Debian packaging of Tor.
Until then, Tor is outside the scope of FreedomBox, in my opinion.
Don't get me wrong: Tor is a _very_ relevant tool, my point is that it
is not yet mature enough to mainstream among non-geeks, when not even
mainstream among geeks yet.
Same goes for any other exciting inventions - I do not mean to
fingerpoint Tor specifically, that just happen to be the one you
emphasize so I feel a need to "de-emphasize" it.
</pre>
</blockquote>
<small><br>
I would have thought that the firewall penetration will be a non issue,
the boxes will presumably be connecting via a IPsec VPN (Strong Swan)?
In tunnelling mode this would allow NAT traversal. Finding your
friend's box (node) is the problem, for which we must rely on some form
of dynamic dns.</small><br>
<br>
<blockquote cite="mid:20130105152049.30178.98668@auryn.jones.dk"
type="cite">
<blockquote type="cite">
<pre wrap="">I know that Tor is well tested, but is this not far too centralized a
way of bootstrapping the networks? More so, is it not going to make a
lot of people nervous about hosting a node on the network. It wouldn't
take more than the mention of silkroad and CP in the mainstream media
to taint the whole project.
Has there been any discussion into alternatives? Such as running a
basic xmpp client on each box, which periodically contacts your
friends' boxes with its current IP address, encrypted and signed using
PGP? The host freedombox JID could be distributed with public keys.
This would allow anyone with a domain name to run a tracker for their
friends' boxes, or to just use a free jabber server.
</pre>
</blockquote>
<pre wrap="">I find your idea interesting.
But please think of FreedomBox not as a box containing unique tools, but
as a box containing uniquely simple access to common tools.
...which means specifically for your idea: Please implement your idea as
a common concept independent of FreedomBox, and when it gains some
traction then suggest that FreedomBox _also_ adopts that concept -
either at its core or perhaps as an option among several.
It is my understanding that those talking about Tor as core platform for
Freedombox, seek a high level of privacy, which is more complex to reach
reliably, and is not yet mainstream even among geeks.
I imagine that those seeking high level of privacy (read: secrecy and
stealth) would find it outragous that their box would act as a beacon
towards (most likely centralized, due to the issue of non-public IPs)
xmpp servers.
What I have not seen (please do correct me if simply I've missed it!) is
discussions taking into account the level of privacy needed, and being
realistic about which privacy levels are achievable at an early stage
reflecting what is already mainstream among geeks - which I call
FreedomBox 1.0.
- Jonas
<small>
</small></pre>
</blockquote>
<small>Tor does provide anonymity, it is also a central record for
all
machines connected<br>
see - <a moz-do-not-send="true" class="moz-txt-link-freetext"
href="https://metrics.torproject.org/exonerator.html">https://metrics.torproject.org/exonerator.html</a>
<br>
<br>
As<big> </big>for unique tools, FreedomBuddy certainly the glue
binding together common tools, but no more so than implementing a
custom xmpp client with say python-jabberbot (in debian repo). XMPP
works on a client server basis, so your IP is not being broadcast
beyond your trusted XMPP and DNS servers. If you really need to be
anonymous, connect to them over Tor.<br>
<br>
Now don't get me wrong, I'm a believer in decentralization, anonymity
and net neutrality, I just don't think that fbx should be rolling out
with Tor enabled in version 1.0 I think that it is a very useful tool,
and when a critical mass is met alongside say, another Arab spring,
then users should be urged to enable the service.<br>
<br>
- Graham<br>
<br>
<br>
<small><br>
<br>
</small></small>
</blockquote>
<small><small></small>See basic example clients here -
<a class="moz-txt-link-freetext" href="http://www.linuxforu.com/2012/06/use-xmpp-to-create-your-own-google-talk-client/">http://www.linuxforu.com/2012/06/use-xmpp-to-create-your-own-google-talk-client/</a><br>
"The code below will connect the client to the server and
authenticate. A base client in just 12 lines of code — can you believe
it?"<br>
<br>
- Graham<br>
</small>
<pre class="brush: python; gutter: true; first-line: 1">
</pre>
<br>
<small><br>
</small>
</body>
</html>