Re: [Social-discuss] Future of GNU social

From:

Bjarni Rúnar Einarsson

Subject:

Date:

Mon, 31 May 2010 11:01:06 +0200

On 28 May 2010, at 22:50, Ted Smith <address@hidden> wrote:

> On Fri, 2010-05-28 at 18:45 +0200, Dan Brickley wrote:
> I don't think OStatus can be used for this, because to me, it seems that
> OStatus depends on HTTP as a transport protocol. It might be possible to
> hack in a peer-to-peer network, but it would take a lot of work -
> probably more work than just finding those social networking primitives
> would.

In terms of actual deployments, so-called "pure" P2P networks are just
a transport layer plus a DHT for DNS-independent routing, and a heap
of code for getting around the fact that most users live behind NAT
firewalls.

If you can offer users the ability to serve HTTP requests this way,
then they can absolutely host their own nodes (e.g., with Tor hidden
services). Users with public facing servers can host their own nodes
with no problems whatsoever.

It turns out this isn't really that hard to do, you just need someone with a visible IP to run a reverse proxy for you. Combine this with a dynamic DNS system, and the proxies themselves can be organized in a 'cooperative' p2p-like structure.

Getting this to work is my project for the summer, and I am quite excited about the possibility of every single device out there, be it server or laptop or mobile phone, becoming a potential publisher of content on the web instead of merely a consumer.

However, I am a little more concerned with the usability side of things - I took a look at StatusNet the other day, and was frightened off by the gigantic list of dependencies. That either has to be reduced dramatically or the complexity hidden somehow, for average-Joe to be able to run his own node in a distributed social network.

But on the other hand, it's probably too early to worry about that. Get it working first, then get it working well?

address@hidden

http://bre.klaki.net/

http://bre.klaki.net/bre/contact.shtml