Re: [GNUnet-developers] Why old-school C?

[Ed Baskerville]

On Jul 15, 2015, at 2:20 AM, Jeff Burdges <address@hidden> wrote:

> Just fyi, one should not assume that GNUnet as a whole should run on a
> mobile device anyways, due largely to metadata leakage issues. 

What are these metadata leakage issues?

> We'll make GNS run over Tor with only a few processes.  In practice, the
> processes relevant for a mobile application could perhaps be bundled
> into an application somehow. 

Is there a mechanism for such a mini-peer implied in the GNUnet protocols? What 
pieces would end up actually running on a mobile device?

> Anyways, privacy tools like Tor, GPG/PGP, XMPP+OTR, etc. arrive extremely 
> late on iOS, if they arrive at all.  Just don't use iOS if you care about
> privacy.

From the perspective of an individual user, sure.

From the perspective of trying to build a network: if the attitude is to build 
tools for users that care about privacy, then, yeah, same deal, don't build 
tools for iOS--or OS X or Windows for that matter.

But if the attitude is to try to build a network that helps people that don't 
yet care--or that don't care enough to abandon their comfortable computing 
environment--then you're going to have to build tools for those platforms. 
Non-privacy-obsessed people are important to reach out to, and their perceived 
switching cost is very high. But if all they have to do is use better* apps 
that their friends are using in order to get on the network, you can use it as 
an opportunity to teach more people about privacy and abuse of data by large, 
powerful organizations, and about how free software makes it easier to verify 
that privacy really is being protected.

* along dimensions unrelated to privacy

Ed