BitTorrent as distributed social network

[Colin Hawkett]

Article looking at the possibility of BitTorrent becoming distributed social network, more-or-less along the lines of Diaspora, but with an already-installed user base in the hundreds of millions.

[Robert Steele]

very nice.  cross-posted to Phi Beta Iota tagged autonomous internet.  thank you for this.

[Venessa Miemis]

that is fascinating

On Thu, May 12, 2011 at 10:04 PM, Colin Hawkett <haw...@gmail.com> wrote:

[Colin Hawkett]

The issues I see -

1. Where is your profile page served from? Where is your personal data? Does it remain available if you are not online? If it does, how is it secured?
2. People see BitTorrent as the means they get.... 'stuff'. Are they comfortable bringing that activity mainstream? Anonymity is a big deal in this space, which can be a bit of a pain for a social platform, depending on how you look at it. I can't see people getting 'stuff' while logged in as themselves.
3. Discovery/Search - for exactly the same reason you need pirate bay or equivalent, you need a central server to locate people, etc. - unless they have some pretty funky new distributed tech.
4. Synchronisation between nodes - e.g. my desktop/laptop/tablet and within the social graph. This is interesting in general, but gets harder when the data is volatile (which it generally is not with BitTorrent in its current form). e.g. I update my status - how does that get reliably to everyone... in near-real-time? Distributed twitter.... tough.
5. Authentication - not clear to me how this is managed.

Essentially the same things Diaspora has to worry about, plus not being perceived as 'social'. The user base is a big deal, but it is the weakest form of user base because they are almost entirely anonymous. One slight upside is that this user base probably has a higher than usual proportion of 'early adopter' psychology.

My take is that a logical network topology based on the social graph is a good thing, and could improve the viability of a P2P physical topology, but it needs to start with a decent model of the social graph, and I'm still not seeing that - it always seems to be tacked on to some other property that people are trying to leverage for existing network effects.

[Samuel Rose]

On Fri, May 13, 2011 at 12:12 AM, Colin Hawkett <haw...@gmail.com> wrote:
> The issues I see -
>
> Where is your profile page served from? Where is your personal data? Does it
> remain available if you are not online? If it does, how is it secured?

This could be solved in numerous ways, but solving it wouldn't negate
the other problems that I think you have nailed...

> People see BitTorrent as the means they get.... 'stuff'. Are they
> comfortable bringing that activity mainstream? Anonymity is a big deal in
> this space, which can be a bit of a pain for a social platform, depending on
> how you look at it. I can't see people getting 'stuff' while logged in as
> themselves.

Exactly. Like it or not, many bittorrent users have been using the
protocol/application to share files. Chances are they won't want that
activity tied to their identity.

> Discovery/Search - for exactly the same reason you need pirate bay or
> equivalent, you need a central server to locate people, etc. - unless they
> have some pretty funky new distributed tech.

I think this part could be done by extending bittorrent client
software so that it would "serve" the identity of the user. This would
need a defined distributed architecture, but a "finger"-type of
protocol could be implemented in the bittorrent protocol.

> Synchronisation between nodes - e.g. my desktop/laptop/tablet and within the
> social graph. This is interesting in general, but gets harder when the data
> is volatile (which it generally is not with BitTorrent in its current form).
> e.g. I update my status - how does that get reliably to everyone... in
> near-real-time? Distributed twitter.... tough.

Not that it is perfect, but diaspora relies on Reddis to distribute
data about updates (as far as I know. Please jump in correct me if I
am wrong there). Erlang (programming language) also has some
long-standing examples of how this can work with Mnesia database. No
matter how it's done, work will be needed to improve reliability and
performance, but this is now within our grasp (even in applications
that run across http )

> Authentication - not clear to me how this is managed.
>

> Essentially the same things Diaspora has to worry about, plus not being
> perceived as 'social'. The user base is a big deal, but it is the weakest
> form of user base because they are almost entirely anonymous.

This is really the biggest obstacle, and it's not a technical problem.

> One slight
> upside is that this user base probably has a higher than usual proportion of
> 'early adopter' psychology.
> My take is that a logical network topology based on the social graph is a
> good thing, and could improve the viability of a P2P physical topology, but
> it needs to start with a decent model of the social graph, and I'm still not
> seeing that - it always seems to be tacked on to some other property that
> people are trying to leverage for existing network effects.

--
--
Sam Rose
Future Forward Institute and Forward Foundation
Tel:+1(517) 639-1552
Cel: +1-(517)-974-6451
skype: samuelrose
email: samue...@gmail.com
http://futureforwardinstitute.com
http://forwardfound.org
http://hollymeadcapital.com
http://p2pfoundation.net
http://socialmediaclassroom.com

"The universe is not required to be in perfect harmony with human
ambition." - Carl Sagan

[Jon Lebkowsky]

Has anybody defined use cases or user stories for "distributed social network"? Maybe we're not clear enough about the problem we're trying to solve.

~ Jon

--
Jon Lebkowsky (@jonl)
Internet Expert and Strategist

Honcho, Polycot Associates, LLC

Work |  Twitter | LinkedIn | Facebook | Wikipedia | Blog | EFF-Austin

[Samuel Rose]

On Fri, May 13, 2011 at 9:27 AM, Jon Lebkowsky <jon.le...@gmail.com> wrote:
> Has anybody defined use cases or user stories for "distributed social
> network"? Maybe we're not clear enough about the problem we're trying to
> solve.
>
> ~ Jon
>

This is generally a great idea. We've created an architecture at
https://docs.google.com/present/view?id=dfxhxcx8_67c87nc9hp that we
believe would allow people to write and run applications as we know
them, but have them work in a distributed fashion. Still, I think it
could be useful to start from the "user story" application perspective
to see what people are thinking of from that end.

[Bryce Lynch]

This reminds me a little of Peerscape, which was implemented as a web
browser add-on. Unfortunately, the project seems to have petered out.

--
The Doctor [412/724/301/703]
http://drwho.virtadpt.net/
"I am everywhere."

[Bryce Lynch]

On Fri, May 13, 2011 at 00:12, Colin Hawkett <haw...@gmail.com> wrote:
> Where is your profile page served from? Where is your personal data? Does it
> remain available if you are not online? If it does, how is it secured?

Seeing as how it's implemented on top of the original BitTorrent
client, it seems as if a user's data is served off of their running
client and cached locally by people who are connected to them. As for
how it's secured, I don't know. I'd think that access to a particular
user's information would be restricted by whether or not the user has
made something public, private (friends-only), or private
(user-defined group only). ACLs on the user's machine would, of
course, apply to the cache files maintained by the client.

> People see BitTorrent as the means they get.... 'stuff'. Are they
> comfortable bringing that activity mainstream? Anonymity is a big deal in

They aren't now?

> this space, which can be a bit of a pain for a social platform, depending on
> how you look at it. I can't see people getting 'stuff' while logged in as
> themselves.

They do it all the time with private and registration-only trackers.
And hybrid BitTorrent/social networking sites like hexagon.cc, for
that matter.

> Discovery/Search - for exactly the same reason you need pirate bay or
> equivalent, you need a central server to locate people, etc. - unless they
> have some pretty funky new distributed tech.

It's possible that they're using DHT search to find users, sort of
like the method used by Gnutella to find files (either searching
actively or searching the cached indices of peers). I'll have to play
around with it to find out for sure.

> Synchronisation between nodes - e.g. my desktop/laptop/tablet and within the
> social graph. This is interesting in general, but gets harder when the data
> is volatile (which it generally is not with BitTorrent in its current form).
> e.g. I update my status - how does that get reliably to everyone... in
> near-real-time? Distributed twitter.... tough.

Tough but not impossible. I think it'll be worth experimenting with
to see how they did it. It would also be worth playing with the
distributed microblogging implementations out there (like Plexus and
rstat.us) to see how well they work.

It might also be worth taking a look at Fossil to see how well its
distributed wiki/bug tracker functionality works.

> Authentication - not clear to me how this is managed.

Nor I, from the article.

[Venessa Miemis]

i'd like to write a short post on the blog about this, but i'm still
trying to get my head around the implications.

what's the relationship between bittorrent as social network &
municipal broadband networks/mesh nets?

- v

[Bryce Lynch]

On Fri, May 13, 2011 at 10:27, Venessa Miemis <veness...@gmail.com> wrote:

> what's the relationship between bittorrent as social network &
> municipal broadband networks/mesh nets?

We're not sure how a BitTorrent-based socnet will work on a mesh
network. It's an experiment we have on the table for Byzantium in the
future, and we'll post our results when we try it.

As for municipal broadband, it appears that those battles have yet to
be really won, given the pushback that the big ISPs are putting on the
municipal projects. Off the cuff, my concern is that the ISPs will
treat BT-based socnets like they treat BT, i.e., with suspicion and
possibly throttling. That could hurt adoption.

[Jon Lebkowsky]

I noticed Adam Fisk on the list of participants at Contact Summit; is he on this list? His feedback on the BitTorrent-based social network idea would be valuable.

~ Jon

[richard adler]

I suspect the real issue with using BitTorrent in this way is not technical (it sounds like there would be many advantages to that approach), but rather political.

Given how the torrents have been demonized as "illegal" and "dangerous" to mainstream users, if the description of a nextnet got out there as "built on a BitTorrent foundation," it seems like that would just be a gift on a golden platter to every potential (and well-financed and PR-savvy) opponent of the project.

Richard

[Venessa Miemis]

On May 13, 10:50 am, Bryce Lynch <virtualad...@gmail.com> wrote:

> On Fri, May 13, 2011 at 10:27, Venessa Miemis <venessamie...@gmail.com> wrote:
> > what's the relationship between bittorrent as social network &
> > municipal broadband networks/mesh nets?
>
> We're not sure how a BitTorrent-based socnet will work on a mesh
> network.  It's an experiment we have on the table for Byzantium in the
> future, and we'll post our results when we try it.
>
> As for municipal broadband, it appears that those battles have yet to
> be really won, given the pushback that the big ISPs are putting on the
> municipal projects.  Off the cuff, my concern is that the ISPs will
> treat BT-based socnets like they treat BT, i.e., with suspicion and
> possibly throttling.  That could hurt adoption.

it makes me wonder if relationships/partnerships could be formed with
tribal councils on Native American reservations. they're considered
sovereign nations, aren't they? it would be interesting to see what
their feeling would be towards running experiments/prototypes of these
types of innovations there, where they have their own rules. i have no
idea of the complexities of that, but it seems like if a bunch of
possible sustainability/thrivability solutions were iterated there and
then shared freely with them, that would be an incentive to allow it.

- v

[Colin Hawkett]

There's a lot of angles to that one :) Here's a few of them -

1. There isn't a relationship. BitTorrent as a social network wouldn't care what the physical network topology was, as long as it could a) find stuff, and b) route data from A -> B. Its performance could probably be affected significantly, but not its ability to operate.
2. If we consider a fully distributed/decentralised internet, where there are no central servers, then the BitTorrent architecture has some problems, since it relies on trackers (central servers) for the discovery process. The basic BitTorrent mechanism goes like this (correct me if I'm wrong) -
1. You grab a torrent file from some website - it contains a list of trackers
2. Your BitTorrent client then contacts those trackers to find out which other connected peers have part or all of the file you want
3. You start grabbing different bits of the file from the different peers identified by the trackers (creating the swarm). You are also part of the swarm - i.e. serving parts for others.
4. The trackers need to be reliably available (i.e. a central server), or they won't know much about what is going on in the network, and more importantly the torrent files you get hold of would go out of date quickly and give you scant information about where to get the file. The point being that BitTorrent is really dependent upon reliable central servers, and uses P2P for download performance more than anything else.
   
3. More abstractly - the impact of a social logical P2P topology (e.g. BitTorrent social, Diaspora, etc.) on potential physical topology (e.g. mesh, backbone etc.) is pretty cool. I'm not convinced this space has been explored so much and is something that interests me greatly. The basic premise is that P2P systems end up having a lot of dependence on trust. Which nodes should I route through? Who should hold copies of my data? Which identity provider do I use? etc. - lots of questions of this sort arise. For a community mesh network, we might consider an architecture where any node on our community's physical network is a trusted replication peer, or that our community network as a whole collaborates for identity provision within that community. At a logical level (e.g. communities which span geographic boundaries), we might define similar or alternate capabilities - there is a lot of intersection, and this provides various 'pollenating' benefits. To put it into concrete BitTorrent speak - a community mesh might contain a tracker node or two for the peers within the community mesh. From the description above, a torrent file just describes a set of trackers (equivalent to a set of communities), and has the ability to link communities by the files they have in common. It is reasonable to see the broader social network being used to manage access control between communities as well (who can use our trackers - who can access which peers within our community - do we trust the community which provides your identity?). This sort of stuff maps really well to real-world structures.

Anyway, that is a pretty average description of a bunch of concepts all loosely pulled together to reach a point about how BitTorrent social (whatever that might end up being) could have relevance to community mesh networks. The broader discussion in this space is very cool... 

Colin

- v

[Curtis Faith]

> Not that it is perfect, but diaspora relies on Reddis to distribute
> data about updates (as far as I know. Please jump in correct me if I
> am wrong there).

I guess I'll use my free dumb question for newcomers card here. What is Reddis? Any project links?

I tried googling various combinations of Diaspora and Reddis to no avail. Too many false matches on the Indian word Reddis and Indian Diaspora.

- Curtis

[Venessa Miemis]

On Fri, May 13, 2011 at 12:42 PM, Curtis Faith <cur...@worldhouse.org> wrote:

> Not that it is perfect, but diaspora relies on Reddis to distribute
> data about updates (as far as I know. Please jump in correct me if I
> am wrong there).

I guess I'll use my free dumb question for newcomers card here. What is Reddis? Any project links?

maybe he meant reddit.com 

[Miles Fidelman]

Venessa Miemis wrote:
> On Fri, May 13, 2011 at 12:42 PM, Curtis Faith <cur...@worldhouse.org
> <mailto:cur...@worldhouse.org>> wrote:
>
> > Not that it is perfect, but diaspora relies on Reddis to distribute
> > data about updates (as far as I know. Please jump in correct me if I
> > am wrong there).
>
> I guess I'll use my free dumb question for newcomers card here.
> What is Reddis? Any project links?
>
>

> maybe he meant reddit.com <http://reddit.com>

>
>
> I tried googling various combinations of Diaspora and Reddis to no
> avail. Too many false matches on the Indian word Reddis and Indian
> Diaspora.
>

diaspora - https://joindiaspora.com/
http://redis.io/ - an open-source key-value store

--
In theory, there is no difference between theory and practice.
In<fnord> practice, there is. .... Yogi Berra

[Samuel Rose]

On Fri, May 13, 2011 at 12:53 PM, Miles Fidelman
<mfid...@meetinghouse.net> wrote:
> diaspora - https://joindiaspora.com/
> http://redis.io/ - an open-source key-value store
>
>

Thanks Miles.

--
--
Sam Rose

[Sepp Hasslberger]

On May 13, 2011, at 3:43 PM, Bryce Lynch wrote:

> On Thu, May 12, 2011 at 22:04, Colin Hawkett <haw...@gmail.com> wrote:
>> Article looking at the possibility of BitTorrent becoming distributed social
>> network, more-or-less along the lines of Diaspora, but with an
>> already-installed user base in the hundreds of millions.
>
> This reminds me a little of Peerscape, which was implemented as a web
> browser add-on. Unfortunately, the project seems to have petered out.

Peerscape sounds like a great idea. That is how I would imagine the distributed net to function.

"Under the hood, your computer stores copies of your data, the data of your friends and the groups you have joined, and some data about, e.g., friends of friends. It also caches copies of other data that you navigate to. Computers that store the same data establish connections among themselves to keep it in sync.

It's all about managing information instead of servers. Peerscape uses public-key cryptographic signatures to encode the relationships among people, groups, and their content."

No need for servers, and the data can be locally stored and redundantly kept available for who would like to access them.

Sepp

[Curtis Faith]

Thank you, I figured it was something like that.

I apologize if I'm saying anything that's been covered here before, I only read back a few months in the archives so far.

Now that I see it's a key-value store, NOSQL variant, I can already see huge scalability problems with that approach. It hints that they do not have the optimal architecture. Thought, I'll have to dig into it further to make sure.

Social networking software performance is a systems architecture problem at heart. There isn't any amount of Moore's Law goodness that can compensate for the sins of a bad architecture.

Social network servers have a data distribution problem not a database problem. Many good lessons come from finance for building trading and exchange data feeds. People pay big money to make sure their data comes in reliably with extremely low latency. The scalability issues have been solved. You know how quickly the volumes blow up when all hell breaks loose in the markets with the high-frequency-trading algorithms running full tilt, right?

You need to use some of the same kinds of tricks in a distributed system to get scalability, but that's not possible to do in a 100% pure peer-to-peer architecture.

There is a reason that Twitter performance sucks despite the fact that some supposedly very experienced people must have been working on it for years. They're optimizing the parts without changing the fundamental architecture to a data distribution architecture. The signs of some wrong architectural decision are written all over my machine every few days when I see Twitter not respond to mundane everyday requests, not throttle back to increase latency to reduce server load, or post a Twitter is Busy message of some sort or another.

Twitter acts like a highway that is tuned to be close to maximum capacity; just before the drop in efficiency hits and you start seeing red brake lights. One little thing out of place on the side of the road and traffic comes to a standstill.

This needs to be even more carefully considered for a true distributed peer-to-peer architecture to work. Efficiency is going to dictate that some peers are going to serve different distribution roles in the network. How and why they do this is critical to acceptance of any system.

Where can I find the best historical thinking on these issues for this list? Any good blog posts? Pointers to web sites or design documents? Who are the people here who consider themselves system-level architects for social networking?

- Curtis

[Charles N Wyble]

On 05/13/2011 11:42 AM, Curtis Faith wrote:
>> Not that it is perfect, but diaspora relies on Reddis to distribute
>> data about updates (as far as I know. Please jump in correct me if I
>> am wrong there).
> I guess I'll use my free dumb question for newcomers card here. What is Reddis? Any project links?

Reddis is a nosql data store. Hmmm. Wow. It's hard to search for. Can't
seem to find a project site for it at the moment.

> I tried googling various combinations of Diaspora and Reddis to no avail. Too many false matches on the Indian word Reddis and Indian Diaspora.

Yeah. They need some SEO love it looks like.

> - Curtis

[Miles Fidelman]

Charles N Wyble wrote:
> On 05/13/2011 11:42 AM, Curtis Faith wrote:
>
>>> Not that it is perfect, but diaspora relies on Reddis to distribute
>>> data about updates (as far as I know. Please jump in correct me if I
>>> am wrong there).
>>>
>> I guess I'll use my free dumb question for newcomers card here. What is Reddis? Any project links?
>>
> Reddis is a nosql data store. Hmmm. Wow. It's hard to search for. Can't
> seem to find a project site for it at the moment.
>
>

partially because it's redis (one "d")

[Isaac Wilder]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I've been hacking on Diaspora for a few months, though I'm far from
understanding all of the technology that has gone into it. Here are my
thoughts.

On 05/13/2011 02:20 PM, Curtis Faith wrote:
> Thank you, I figured it was something like that.
>
> I apologize if I'm saying anything that's been covered here before, I
only read back a few months in the archives so far.
>
> Now that I see it's a key-value store, NOSQL variant, I can already see
huge scalability problems with that approach. It hints that they do not
have the optimal architecture. Thought, I'll have to dig into it further
to make sure.

I agree that Diaspora has not found the optimal architecture for
p2psn, but I do think that they've yet to be topped. I'm hoping that
MondoNet or Project Byzantium will have something to say here.

>
> Social networking software performance is a systems architecture
problem at heart. There isn't any amount of Moore's Law goodness that
can compensate for the sins of a bad architecture.

This is where I've found my head at lately. What if we get a /64 block
of v6 space from ARIN, and put it to use building a non-web
application that uses direct addressing for each peer. This might
solve a lot of the problems introduced in p2p systems via NAT and
other subnetting practices. Perhaps the web-based, server-based,
db-based approached is too encumbered by HTTP packaging. Perhaps we
need to think outside of the browser. Stand-alone software could open
up its own port, instead of having to package and push everything
through 80.

>
> Social network servers have a data distribution problem not a database
problem. Many good lessons come from finance for building trading and
exchange data feeds. People pay big money to make sure their data comes
in reliably with extremely low latency. The scalability issues have been
solved. You know how quickly the volumes blow up when all hell breaks
loose in the markets with the high-frequency-trading algorithms running
full tilt, right?
>
> You need to use some of the same kinds of tricks in a distributed
system to get scalability, but that's not possible to do in a 100% pure
peer-to-peer architecture.

What are the effects when we don't view the architecture as
single-tier peering, but muti-tiered and federated. That is,
peer-to-peer network on a local level connected with other
peer-to-peer networks on a regional level, and so on...
We can introduce incredibly low latency in local-scale interaction at
the cost of introducing some latency in global-scale ones.

>
> There is a reason that Twitter performance sucks despite the fact that
some supposedly very experienced people must have been working on it for
years. They're optimizing the parts without changing the fundamental
architecture to a data distribution architecture. The signs of some
wrong architectural decision are written all over my machine every few
days when I see Twitter not respond to mundane everyday requests, not
throttle back to increase latency to reduce server load, or post a
Twitter is Busy message of some sort or another.
>
> Twitter acts like a highway that is tuned to be close to maximum
capacity; just before the drop in efficiency hits and you start seeing
red brake lights. One little thing out of place on the side of the road
and traffic comes to a standstill.
>
> This needs to be even more carefully considered for a true distributed
peer-to-peer architecture to work. Efficiency is going to dictate that
some peers are going to serve different distribution roles in the
network. How and why they do this is critical to acceptance of any system.

I couldn't agree more. Every system has its trade-offs. Let's consider
the relationship between efficiency, latency, reliability,
distribution, and federation. I'm trying to get the Free Network
Foundation going to have just this type of conversation. Are people
here interested in participating?

>
> Where can I find the best historical thinking on these issues for this
list? Any good blog posts? Pointers to web sites or design documents?
Who are the people here who consider themselves system-level architects
for social networking?

I don't consider myself a system-level architect. Not yet at least.
But I am keenly interested in having this conversation, particularly
with you, Curtis. I've got a lot to learn. We all have. Charles knows
a thing or two about systems architecture, as do many others. Still,
maybe what we need is new ideas. I'm not sure that there really is a
historical conversation to point to, but I think that's exactly why we
should be having the conversation.
>
> - Curtis
>

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)

iQEcBAEBAgAGBQJNzZWDAAoJEA8fUKCD77NLlc0H/jHFZuTtQy61O9U3sYphFlBT
cfoPGem3U0O5CB2fg07jnAXq/R1pET0lwGBucqG0AN6WknK3Vce1RSunMRXLszFv
xg2Y+97pX33u45T544ZQKZLNt+XVbHFzYtWJdtTG32jNn1zbDQqhrWCbpoW1FMhp
rdGI40+PTFAV08MemPOkLFgVOA1vheSOc7PVUH840snbMUhjKVd6vgKm4OlMwhMN
BmSIoF6uTmsdiKiWWDXoEGwORkRQNomGkKx/B7gv4wBYIqeZKqa7c7dUX+RaGek3
mLw+OnsxaPGTk7UOx/1KZcS0+5xsjFCpGIr+yf+/ign9Nm5gSv1x3umbzTYz/z4=
=WMWU
-----END PGP SIGNATURE-----

[Isaac Wilder]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I REALLY like this idea. I wonder if anyone on list has connections
to people in the first nations community. I just really think this is
an idea worth exploring.

imw

>
>>
>> -- The Doctor [412/724/301/703]http://drwho.virtadpt.net/ "I am
>> everywhere."

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)

iQEcBAEBAgAGBQJNzZYIAAoJEA8fUKCD77NL0vsIAKLK2L2UjcQVQqucUvIGhplU
eA0roP8GcmI5PYl56uxpDo/1LH4OhdVkxsC7zN+CcjVnN61emBF1oNhau1RWN6vu
RguLqQDrg5Do4FSW2nnjIsdtrfhnQDH4Rved8rjVyQxCcI/OvdCQg/fpPjQE+Qh6
bhm4hTORLBTLIAm0IKCOkacSmivlsVfg9LMsr3wOzo1ejiC6SMORq8kCx41/c4Ue
JVLT/RUeIRZC/8gzUJSk0C1oPyB/YOU2mglfmfhLnJQ/5VeHE6ps6X/pyBOvEQyH
QURNFbkU5fJNNpwazYch0XhLnBvx5jS9hwmmvGYGqqRM9UidEUUdn1gRhiL2ruk=
=dBus
-----END PGP SIGNATURE-----

[Isaac Wilder]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 05/13/2011 09:55 AM, richard adler wrote:
> I suspect the real issue with using BitTorrent in this way is not
> technical (it sounds like there would be many advantages to that
> approach), but rather political.

I totally agree here, although there are technical aspects to
consider, too. I don't want to underestimate the power of existing
network effects, but I do wonder if BitTorrent's history is more of a
liability than an asset, at least in terms of the social and political
struggle that lies ahead.

>
> Given how the torrents have been demonized as "illegal" and
> "dangerous" to mainstream users, if the description of a nextnet
> got out there as "built on a BitTorrent foundation," it seems like
> that would just be a gift on a golden platter to every potential
> (and well-financed and PR-savvy) opponent of the project.

You said it.

>
> Richard
>
>
> On Fri, May 13, 2011 at 10:50 AM, Bryce Lynch
> <virtua...@gmail.com <mailto:virtua...@gmail.com>> wrote:
>
> On Fri, May 13, 2011 at 10:27, Venessa Miemis
> <veness...@gmail.com <mailto:veness...@gmail.com>> wrote:
>
>> what's the relationship between bittorrent as social network &
>> municipal broadband networks/mesh nets?
>
> We're not sure how a BitTorrent-based socnet will work on a mesh
> network. It's an experiment we have on the table for Byzantium in
> the future, and we'll post our results when we try it.

I'll chime in here, too, and just say that I've been reading up on
Byzantium, and I'm impressed. My one big critique is that I think
social networking should be viewed as a more central functionality of
peer-to-peer infrastructure. How do we build the functionality of
Facebook into a mesh node. I think that once we've got that figured
out, the evo-revo-lution can begin.

>
>
> As for municipal broadband, it appears that those battles have yet
> to be really won, given the pushback that the big ISPs are putting
> on the municipal projects. Off the cuff, my concern is that the
> ISPs will treat BT-based socnets like they treat BT, i.e., with
> suspicion and possibly throttling. That could hurt adoption.

At a certain point, any technology that's put to the end of enabling
resilient peer-to-peer communications is going to be resisted. Perhaps
what we need to do is think about the best way to invite such
resistance, and then defeat it. Let them try to 'traffic-shape' our
efforts, and we'll have our day in courts. Let them try to turn off
our radios. We will power them with the sun and broadcast our freedom
until the stop us by force.

>
>
> -- The Doctor [412/724/301/703] http://drwho.virtadpt.net/ "I am
> everywhere."
>
>

take care,
imw

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)

iQEcBAEBAgAGBQJNzZjPAAoJEA8fUKCD77NL/u8IAMwLBQ/IoTdH7dSmfWOeZrpA
YqQZI0WLqFu0gwYt7O0tYpLMif1nygqlxEknnFTv1sByOFN2rwlitvIMbH85w3+4
oIRBztJO6eGJv7nxMikkSwtofgkY/H1c2HJLfyQG8200rsW/T+LLfLBlZgANfIlf
2N1E94CD43xlLqsaX9dj8pr5hbAlPIOuJqVkZtNiBpVJ2BvsSnVfL9U0Uk6LHUcw
C02DFURVjo3dScaNPZfLBtDDODsT062RBs8fTmLqHPKX6oj4RYLulqWweUWczYMW
siDDzIuPqdy6UzSYl4eE7wm9ORtavmErn9KbzdjpBt9oYvscN7WHWFBT8h/vtWU=
=1Jyn
-----END PGP SIGNATURE-----

[Samuel Rose]

Oops, Miles is right. :)

--
--
Sam Rose

[Melvin Carvalho]

On 13 May 2011 21:20, Curtis Faith <cur...@worldhouse.org> wrote:
> Thank you, I figured it was something like that.
>
> I apologize if I'm saying anything that's been covered here before, I only read back a few months in the archives so far.
>
> Now that I see it's a key-value store, NOSQL variant, I can already see huge scalability problems with that approach. It hints that they do not have the optimal architecture. Thought, I'll have to dig into it further to make sure.

What makes you think that key value does not scale?

>
> Social networking software performance is a systems architecture problem at heart. There isn't any amount of Moore's Law goodness that can compensate for the sins of a bad architecture.
>
> Social network servers have a data distribution problem not a database problem. Many good lessons come from finance for building trading and exchange data feeds. People pay big money to make sure their data comes in reliably with extremely low latency. The scalability issues have been solved. You know how quickly the volumes blow up when all hell breaks loose in the markets with the high-frequency-trading algorithms running full tilt, right?
>
> You need to use some of the same kinds of tricks in a distributed system to get scalability, but that's not possible to do in a 100% pure peer-to-peer architecture.
>
> There is a reason that Twitter performance sucks despite the fact that some supposedly very experienced people must have been working on it for years. They're optimizing the parts without changing the fundamental architecture to a data distribution architecture. The signs of some wrong architectural decision are written all over my machine every few days when I see Twitter not respond to mundane everyday requests, not throttle back to increase latency to reduce server load, or post a Twitter is Busy message of some sort or another.
>
> Twitter acts like a highway that is tuned to be close to maximum capacity; just before the drop in efficiency hits and you start seeing red brake lights. One little thing out of place on the side of the road and traffic comes to a standstill.
>
> This needs to be even more carefully considered for a true distributed peer-to-peer architecture to work. Efficiency is going to dictate that some peers are going to serve different distribution roles in the network. How and why they do this is critical to acceptance of any system.
>
> Where can I find the best historical thinking on these issues for this list? Any good blog posts? Pointers to web sites or design documents? Who are the people here who consider themselves system-level architects for social networking?

Look no further:

http://www.w3.org/DesignIssues/

[Samuel Rose]

On Fri, May 13, 2011 at 3:20 PM, Curtis Faith <cur...@worldhouse.org> wrote:
>
> Social network servers have a data distribution problem not a database problem. Many good lessons come from finance for building trading and exchange data feeds. People pay big money to make sure their data comes in reliably with extremely low latency. The scalability issues have been solved. You know how quickly the volumes blow up when all hell breaks loose in the markets with the high-frequency-trading algorithms running full tilt, right?
>
> You need to use some of the same kinds of tricks in a distributed system to get scalability, but that's not possible to do in a 100% pure peer-to-peer architecture.
>

I believe that diaspora is employing
http://www.amqp.org/confluence/display/AMQP/Advanced+Message+Queuing+Protocol
for addressing this. My understanding is that AMQP is capable of
meeting the challenge.

[Charles N Wyble]

On 05/13/2011 03:57 PM, Samuel Rose wrote:
> On Fri, May 13, 2011 at 3:20 PM, Curtis Faith <cur...@worldhouse.org> wrote:
>> Social network servers have a data distribution problem not a database problem. Many good lessons come from finance for building trading and exchange data feeds. People pay big money to make sure their data comes in reliably with extremely low latency. The scalability issues have been solved. You know how quickly the volumes blow up when all hell breaks loose in the markets with the high-frequency-trading algorithms running full tilt, right?
>>
>> You need to use some of the same kinds of tricks in a distributed system to get scalability, but that's not possible to do in a 100% pure peer-to-peer architecture.
>>
> I believe that diaspora is employing
> http://www.amqp.org/confluence/display/AMQP/Advanced+Message+Queuing+Protocol
> for addressing this. My understanding is that AMQP is capable of
> meeting the challenge.

Queuing systems fall over under high load. Multicast/data fountains are
what a lot of people are using now. Moves the complexity into the
network instead of the servers. Push the ugly around and all that. I
like the multicast approach and I've got a lot of experience with it.
Including building a multicast to unicast converter and back again.

[Miles Fidelman]

Curtis Faith wrote:
> Social networking software performance is a systems architecture problem at heart. There isn't any amount of Moore's Law goodness that can compensate for the sins of a bad architecture.
>
>

Hear, hear!

> Social network servers have a data distribution problem not a database problem. Many good lessons come from finance for building trading and exchange data feeds. People pay big money to make sure their data comes in reliably with extremely low latency. The scalability issues have been solved. You know how quickly the volumes blow up when all hell breaks loose in the markets with the high-frequency-trading algorithms running full tilt, right?
>

Of course near-real-time financial trading has very different
performance requirements for group messaging (which is really what
social networking is all about).

> You need to use some of the same kinds of tricks in a distributed system to get scalability, but that's not possible to do in a 100% pure peer-to-peer architecture.
>

Depends on how you define "pure peer-to-peer." USENET (NNTP) is pretty
good example of a highly scalable multiply connected mesh network.
(Yes, it now has a hierarchy, but it really doesn't have to.) Freenet,
Gnunet, and Gnutella do a pretty good job of scaling.

[Curtis Faith]

Scalability is all about being able to decompose work into parallelizable chunks and minimizing the work that needs to be done on any given node.

General message queuing systems still have queue bottlenecks at the network, disk i/o and CPU level that can be easily exceeded with the wrong overall system architecture.

The key to scalability is to not have any bottlenecks that can't be subdivided and load-balanced on the fly without taking any nodes offline. What matters here is which nodes are talking to which other nodes, and what kind of information they are sending to each other, not the specific implementation of the nodes themselves. That's what I meant by optimizing the parts without changing the fundamental architecture.

Most of the architectural issues associated with the protocols are irrelevant to scalability. You could store data in in-memory lists for high performance, but if the amount of data coming into or out of a node exceeds the network or CPU performance capability, that node is effectively down. If this happens enough, fail-overs merely cascade to other previously under-capacity nodes taking the whole system down to a crawl as users everywhere halt.

So you need to have a way of making sure that this doesn't happen. For example, a mechanism so that nodes start offloading work and user connection responsibility to underutilized server nodes long before any given node reaches capacity. You can't have every peer talk to the busy nodes or when someone with 100,000 followers tweets you'll use up too many resources (individual TCP packets and the associated CPU time to process them). A good architecture will do what Charles describes, it will fan out and distribute the work from updates and change the connection topology based on the characteristics of the communication requests themselves.

Communication data should roll up geographically or topically and then fan out on the same basis. If you have the luxury of collocating a group of servers (for instance, if you are a centralized company, like say Twitter or FaceBook or an stock market exchange data provider) then you can use IP multi-casting to make the network communications very efficient. In a federation of peer servers located all over the world, you won't have this option.

[Miles Fidelman]

Curtis Faith wrote:

Not all of these are axiomatic:

> Scalability is all about being able to decompose work into parallelizable chunks and minimizing the work that needs to be done on any given node.
>

No. It's about being able to decompose and parallelize as necessary.
Minimizing load on given nodes is not necessary.

> The key to scalability is to not have any bottlenecks that can't be subdivided and load-balanced on the fly without taking any nodes offline.
>

It's also nice to be tolerant of offline nodes - failures, upgrades,
flakey connectivity associated with mobile nodes (we are talking mobile
mesh networks here), etc. Scalability and data availability are
somewhat tied at the hip.

> Most of the architectural issues associated with the protocols are irrelevant to scalability. You could store data in in-memory lists for high performance, but if the amount of data coming into or out of a node exceeds the network or CPU performance capability, that node is effectively down. If this happens enough, fail-overs merely cascade to other previously under-capacity nodes taking the whole system down to a crawl as users everywhere halt.
>

Network architecture is a central consideration, and that is driven by
protocol and data architecture. Building around a multi-cast or
broadcast protocol leads to very different architectures and node
designs than peer-wise protocols.

Miles Fidelman

[Charles N Wyble]

On 05/13/2011 05:26 PM, Curtis Faith wrote:
> Communication data should roll up geographically or topically and then fan out on the same basis. If you have the luxury of collocating a group of servers (for instance, if you are a centralized company, like say Twitter or FaceBook or an stock market exchange data provider) then you can use IP multi-casting to make the network communications very efficient. In a federation of peer servers located all over the world, you won't have this option.
>

Why not? You could have multicast on a regional basis and then
unicast-multicast conversion across the WAN. Or build multi cast into
the WAN. MPLSl3VPN and all that jazz. Tis how I built my last
distributed application. *shrugs* I just hack on stuff and get paid good
money for it. I'm not a wizard architect or anything. Just a builder
that sees problems and builds solutions and solves problems as they come
up.

[Samuel Rose]

AMQP supports multicast.

[Colin Hawkett]

On Saturday, 14 May 2011 00:17:31 UTC+10, The Doctor [412/724/301/703] wrote:

On Fri, May 13, 2011 at 00:12, Colin Hawkett <haw...@gmail.com> wrote:
> Where is your profile page served from? Where is your personal data? Does it
> remain available if you are not online? If it does, how is it secured?

Seeing as how it's implemented on top of the original BitTorrent
client, it seems as if a user's data is served off of their running
client and cached locally by people who are connected to them.  As for
how it's secured, I don't know.  I'd think that access to a particular
user's information would be restricted by whether or not the user has
made something public, private (friends-only), or private
(user-defined group only).  ACLs on the user's machine would, of
course, apply to the cache files maintained by the client.

The problem is that if you don't control the client implementation (it just needs to conform to the protocol), then you don't know whether cached versions of your data and the ACL rules will be applied correctly, or if they will be applied at all. No doubt a 'miltant' open faction will produce a client that conforms to the protocol but ignores ACL on philosophical grounds (i.e. everyone should get access to your data, 'your stuff is our stuff, get used to it' type argument). Then there's hackers and crackers etc. In general the security of client node implementations is a pretty tricky problem for distributed systems.

Definitely the social network itself can provide trust knowledge, but whether this is as secure as a central server with lots of devoted security resources is a worthwhile discussion. Obviously the sony debacle shows that when big repositories are compromised, then the damage is massive, but there are also difficult consequences for compromised nodes in a P2P network (e.g. botnet type stuff).

The simplest solutions to this problem devolve to highly available, highly trustable nodes with lots of cached data - (i.e. central servers run by organisations with the resources to supply them...)

> People see BitTorrent as the means they get.... 'stuff'. Are they
> comfortable bringing that activity mainstream? Anonymity is a big deal in

They aren't now?

'mainstream' was probably a poor choice of word - the gist was intended to be aligning their everyday social, personally identifiable activity with their illegal file sharing activity.

> this space, which can be a bit of a pain for a social platform, depending on
> how you look at it. I can't see people getting 'stuff' while logged in as
> themselves.

They do it all the time with private and registration-only trackers.
And hybrid BitTorrent/social networking sites like hexagon.cc, for
that matter.

I wonder how personally identifiable these logins are, or whether those networks have much illegal file sharing activity - i.e. one way or the other real identity and illegal file sharing are things people like to keep separate.

[Curtis Faith]

I haven't been following this super closely but unless I missed something big IPv4 is still the only standard you can count on and multicasting is only supported by certain routers. One can certainly build an abstraction for the architecture that uses multicasting when available, I just don't see how that will be very often in most circumstances. Absent a central authority with access to known hardware or hardware that supports known protocols end-to-end, you need to allow for at least some segments of the communication coming through unicast.

Perhaps we are envisioning different use cases for this. I'm thinking group conferencing with groups located in different parts of the world, not in the same building. I'm also thinking about Twitter-like mixed one-to-many streams of low-bandwidth communications. Unless I'm missing something important (and it wouldn't be the first time this happened) multicasting only provides a benefit if more than one destination (router or server node) is connected to any given multicast capable router (like a Cisco router supporting MPLSl3VPN, for instance).

Your use of the term WAN is perhaps telling. It is common when discussing corporate I.T. infrastructure or private networks. For most individuals, there is just: the internet. Any you can't count on anything existing between you and any other users. No leased-line or ATM or VPN connections or any particular SLA agreement.

The architecture for the next net needs to work with that as a baseline. Any optimizations for special cases should be build on top of that.

I do consider the case of coworking spaces, hacker spaces, maker spaces, etc. to be a growing likely target for this technology, and in these spaces, it seems reasonable to suppose that capable routers will exist if there is compelling reason for their purchase. It also seems reasonable to conclude that many people might be on the same video conference on a regular basis connecting with other hackerspaces, etc. So if there are benefits to be gained through optimizing for multicast within a LAN, I expect these groups to buy the hardware to keep their LAN efficient in that event.

In the end, I think we'd both propose the same solution: as you stated: "multicast on a regional basis and then
unicast-multicast conversion across the WAN." We just differ in our opinion of how often the multicast feature will be used in practice.

Peace

Curtis

[Jon Lebkowsky]

I'm concerned that this conversation addresses how to build rather than what to build. We have a general sense what problem we're trying to solve here, but I'd like to see specifics - which is why I mentioned use cases or user stories a couple days ago. It's meaningless if you build a social network platform without a sense how it'll be used. Replicating Facebook/Twitter streaming of short burst msgs, I submit, is not enough. I think we're already beginning to realize the limitations of those systems, which support shallow and broad messaging, not-quite-conversation, and systems that support deeper conversation and relationship (some of the older forum-based systems, like the WELL). And perhaps we could look at some specific cases, e.g. I'm involved with the Society of Participtory Medicine, which focuses on patient communities, which could be well-supported by powerful networking and communication that includes granular privacy controls.

I advocate for more of a "what to build" conversation before "how to build," though I understand that developers and network administrators are inherently more focused on the how than the why.

~ Jon

--
Jon Lebkowsky (@jonl)
Internet Expert and Strategist

Honcho, Polycot Associates, LLC

Work |  Twitter | LinkedIn | Facebook | Wikipedia | Blog | EFF-Austin

[Miles Fidelman]

Jon Lebkowsky wrote:
> I'm concerned that this conversation addresses how to build rather
> than what to build. We have a general sense what problem we're trying
> to solve here,

Actually, I sure don't. I originally thought the focus was on
distributed, decentralized INFRASTRUCTURE - e.g., wireless mesh
networks that are not dependent on carriers, and are not subject to
disruption or centralized control.

The focus seems to have shifted to decentralized applications - again,
less vulnerable to control and disruption, but dependent on the current
Internet infrastructure.

These are two very different problems, that require very different
solutions.

[Venessa Miemis]

a user-owned
> hardware and software stack that allows communities to self-organize,
> self-govern, and empower themselves to manifest ideas that are good
> for themselves and that build resilience, sustainability, and
> thrivability.

On Sat, May 14, 2011 at 9:24 AM, Jon Lebkowsky <jon.le...@gmail.com> wrote:

I'm concerned that this conversation addresses how to build rather than what to build. We have a general sense what problem we're trying to solve here, but I'd like to see specifics - which is why I mentioned use cases or user stories a couple days ago. It's meaningless if you build a social network platform without a sense how it'll be used. Replicating Facebook/Twitter streaming of short burst msgs, I submit, is not enough. I think we're already beginning to realize the limitations of those systems, which support shallow and broad messaging, not-quite-conversation, and systems that support deeper conversation and relationship (some of the older forum-based systems, like the WELL). And perhaps we could look at some specific cases, e.g. I'm involved with the Society of Participtory Medicine, which focuses on patient communities, which could be well-supported by powerful networking and communication that includes granular privacy controls.

I advocate for more of a "what to build" conversation before "how to build," though I understand that developers and network administrators are inherently more focused on the how than the why.

i've been discussing this with a few people offlist and described it as "a hardware and software stack that allows communities to self-organize, self-govern, and empower themselves to implement ideas that are good for everyone (non-zero sum) and build resilience, sustainability, and thrivability."

this group has been a hub for rich discussion, and i think if we are truly interested in how to communicate/collaborate in practice, not just in theory, then we should just begin.

we'll learn what the better tools/features are that we need/desire by using the existing ones and realizing what's missing. 

thinking and talking about what would be best is a fun exercise, but not always practical when it comes to actual human behavior. i think we should use ourselves as the test case and begin to experiment with what works best for us, and then build the tools to support those existing behaviors.

- v

[Jon Lebkowsky]

On Sat, May 14, 2011 at 8:46 AM, Venessa Miemis <veness...@gmail.com> wrote:

"thinking and talking about what would be best is a fun exercise, but not always practical when it comes to actual human behavior. i think we should use ourselves as the test case and begin to experiment with what works best for us, and then build the tools to support those existing behaviors."

Is this something that people just go off and do? Or are we thinking to organize in some way?  It seems to me that we would benefit from a plan, a set of parameters, a set of models. Maybe some categories of activity, e.g. network infrastructure, network management, use cases, application structure, core technologies, etc.

I think there's been an implicit promise that a project or projects would emerge that would be more focused - beyond loose discussion, data gathering, and curation. We seem to have a set of participants who could take next steps effectively if we can figure out how to organize. Perhaps we focus on finding funding sources so that there can be compensation for doing something in a more organized way.

~ Jon

[richard adler]

Jon,

I'd like to second your call for models and case studies, but I for one would like to see that effort proceed along multiple tracks. These discussions have reviewed a wide variety of technologies and possible infrastructures, and there's no reason to think we couldn't see several models and financing arrangements constructed out of them.

A plurality of models which the community could then review and assess would move these discussions into a more concrete space. But there's also no reason to think we would have to choose only one of those options. Provided the models are designed within a reasonable scope (ie. they don't attempt to link an entire state in the first stage, etc), we could see several networks developed at the same time. Which would also ensure an emphasis on interoperability, open standards, and flexibility.

Richard

[Miles Fidelman]

Venessa Miemis wrote:
> > a user-owned
> > hardware and software stack that allows communities to self-organize,
> > self-govern, and empower themselves to manifest ideas that are good
> > for themselves and that build resilience, sustainability, and
> > thrivability.
>

> i've been discussing this with a few people offlist and described it
> as "a hardware and software stack that allows communities to
> self-organize, self-govern, and empower themselves to implement ideas
> that are good for everyone (non-zero sum) and build resilience,
> sustainability, and thrivability."

Ok... this clarifies things a lot. So, partially to play Devil's
advocate, and partly to understand the purpose here, let me ask two very
specific questions:

1. What does a hardware/software stack - user-owned or otherwise - have
to do with allowing "communities to self-organize,

self-govern, and empower themselves to manifest ideas that are good for
themselves and that build resilience, sustainability, and

thrivability?

2. Why a specific and/or new software/hardware stack, as opposed to the
myriad that already exist (user-owned and otherwise).

Vanessa offers a partial answer with:

> this group has been a hub for rich discussion, and i think if we are
> truly interested in how to communicate/collaborate in practice, not
> just in theory, then we should just begin.
>
> we'll learn what the better tools/features are that we need/desire by
> using the existing ones and realizing what's missing.
>
> thinking and talking about what would be best is a fun exercise, but
> not always practical when it comes to actual human behavior. i think
> we should use ourselves as the test case and begin to experiment with
> what works best for us, and then build the tools to support those
> existing behaviors.

To which, let me elaborate:

There's LOTS of existing hardware/software stacks. At the
hardware/network level these range from fax machines and the Internet to
amateur radio to wireless mesh networks. For that matter, one can
include African drums, smoke signals, and telephone coops started a 100
years ago by Iowa farmers who discovered that strands of barbed wire
worked just fine for supporting telephone pary lines.

At the software level, there were the old computer bulletin boards,
freenets, P2P networks (USENET, FidoNet), and there are today's P2P
applications (FreeNet, GnuNet, Gnutella, IRC, etc.), as well as the
myriad of web sites, blogs, and wikis built from open-source software
and running on user-owned machines. And then there are more covert
technologies, such as onion routers, steganography, botnets, and so forth.

Lots of people (including people on this list) are already using various
combinations of these technologies for applications ranging from
day-to-day business, to social networking, to military operations in
unfriendly territory, to sharing music, to supporting disaster response
(from ham radio operators to Ushahidi), to fomenting revolution in the
streets (fax machines in the Tianamnmen Square days, Twitter in the
middle east), to education, to various collaborative economic projects
(search on the "Public Webmarket," a project I was involved in way back
when), to militia groups and terrorist cells.

All of the technology to date has exhibited both strengths and
weaknesses, and continues to evolve as people identify various
shortcomings. The key point is that the technology has been used, and
continues to be used, successfully in lots of venues - and both usage,
usage patterns, and technology are co-evolving rapidly.

Which leads to the obvious question: What experiences and gaps have led
to this particular set of people having this particular discussion?
What problem are people actually trying to solve - beyond, perhaps,
finding a new and different way of doing things (or more cynically, "not
invented here")?

Personally, my experiences lead me to the conclusion that
self-organization and self-government have very little to do with
technology and lots to do with group process and social infrastructure -
ranging from rules of order to contracts to accounting rules to legal
and cultural context. There's a reason that autocratic, hierarchical
organizations seem to get things done more effectively than ad-hoc
groups of small organizations or individuals; and it has far more with
the ability to make decisions quickly than it has to do with technology
(beyond the basics of having a way to bring people together in a
physical or virtual room). [I was reminded of this recently during the
process of starting a new venture. After starting down the road of
organizing as a cooperative, pretty much everybody else involved said
"every decision is way too painful in a coop" and reminded me of how
many small coops have been reorganizing as LLCs, with clear lines of
authority.]

The Internet is a really interesting example of self-organization on a
global scale - dating back to a few dozen researchers, albeit with the
grand daddy of all funders - which has evolved to become critical global
infrastructure, where nobody owns more than a small piece of the
equipment, and nobody is in charge. It works, continues to evolve, and
has become central to an awful lot of day-to-day life.

Then there's eBay and PayPal - originally an auction site for hobbyists,
it has enabled a whole slew of new economic activity.

I'd be really interested in hearing about real examples of
self-organization and self-government - particularly those that don't
deconstruct to one or two core people actually starting things off. I'd
like to hear about some real examples of emergent self-organization and
self-government, and the ways that technology has enabled/supported
these activities. That would seem to be a good starting point for
looking at what's missing or what needs to be changed.

Miles Fidelman

[richard adler]

All of the technology to date has exhibited both strengths and weaknesses, and continues to evolve as people identify various shortcomings.  The key point is that the technology has been used, and continues to be used, successfully in lots of venues - and both usage, usage patterns, and technology are co-evolving rapidly....

Personally, my experiences lead me to the conclusion that self-organization and self-government have very little to do with technology and lots to do with group process and social infrastructure - ranging from rules of order to contracts to accounting rules to legal and cultural context.

There is truth in this. But a crucial part of these discussions is the example provided by the recent popular uprisings in the Middle East. We saw tremendous ingenuity displayed by activists there, but those innovations also faced serious obstacles because those group processes and social infrastructures were occurring on top of physical infrastructures they did not own and control.

To be sure, people can find ways to route around that lack of control, as indeed they did at critical moments earlier this year. But at least one of the goals of these discussions could be--and should be--models for internet infrastructures that would move that control from corporate or governmental entities and into the hands of the people who use them.

Social infrastructures are a critical topic for discussion, and I welcome it as part of Contact Con. But the technology underlying those social arrangements IS relevant and of central concern here.

Richard

[Miles Fidelman]

richard adler wrote:
>
>
> Personally, my experiences lead me to the conclusion that
> self-organization and self-government have very little to do with
> technology and lots to do with group process and social
> infrastructure - ranging from rules of order to contracts to
> accounting rules to legal and cultural context.
>
>
> There is truth in this. But a crucial part of these discussions is the
> example provided by the recent popular uprisings in the Middle East.
> We saw tremendous ingenuity displayed by activists there, but those
> innovations also faced serious obstacles because those group processes
> and social infrastructures were occurring on top of physical
> infrastructures they did not own and control.
>
> To be sure, people can find ways to route around that lack of control,
> as indeed they did at critical moments earlier this year. But at least
> one of the goals of these discussions could be--and should be--models
> for internet infrastructures that would move that control from
> corporate or governmental entities and into the hands of the people
> who use them.

Agreed, though it keeps striking me as a set of issues that border on
intractable:

1. Shifting to user-owned infrastructure, en masse, requires that there
be really serious incentives for everyone involved. To date, it's been
much easier for people to buy phones, computers, network services from
large organizations that make everything easy and comparatively cheap.
Yes, there are ham radio operators floating around, and people who run
independent wireless networks and such - but most people just don't want
to be bothered until it's too late. Military units, first responders,
militias, terrorist groups will acquire technology in anticipation of
action; most people can't even be bothered to have a first aid kit and
some emergency supplies at home.

2. To the extent that technology is pre-positioned, widely available,
and widely used - it becomes a serious target. What's worked to date is
when people find a new way to take advantage of technology that has
become critical to businesses (fax during the Tiananmen Square
uprisings, twitter more recently); use it until blocked; find
work-arounds in the moment (e.g., Satphones smuggled over the boarder).
It's always going to be an arms race, and it may well be better not to
present an easy-to-anticipate target.

Having said that, there are quite a few technologies already available
that are hard to shut down without shutting down things the
powers-that-be care about:
- onion routers provide for circumenting firewalls
- steganography seems to be widely used by terrorist groups (or so many
would have us believe)
- botnets seem to be pretty hard to shut down (or even track down)
- satphones can be smuggled pretty easily and they can be had from
companies and countries that are on different sides of conflicts
- there's still an awful lot of ham, CB, and emergency radio gear
floating around
- the military spends billions on spread-spectrum, software-defined
radios and ad-hoc mesh networking - with most of the core technology
published in publicly-available journals

Would I like to have a data-capable smart phone / hot spot; that relies
totally on ad hoc mesh networking for connectivity? Of course. Getting
enough other people to adopt them, for them to actually be usable, is a
much harder problem. It's hard enough for a wireless carrier to roll
out a new generation of technology - after investing billions in new
phones, towers, radios, roaming agreements, and marketing to make the
new phones useful on day one.

> Social infrastructures are a critical topic for discussion, and I
> welcome it as part of Contact Con. But the technology underlying those
> social arrangements IS relevant and of central concern here.

Sure... but I expect that useful results are more likely to come out of
MILCON and DEFCON.

[Miles Fidelman]

Miles Fidelman wrote:

> richard adler wrote:
>> To be sure, people can find ways to route around that lack of
>> control, as indeed they did at critical moments earlier this year.
>> But at least one of the goals of these discussions could be--and
>> should be--models for internet infrastructures that would move that
>> control from corporate or governmental entities and into the hands of
>> the people who use them.
>

> Having said that, there are quite a few technologies already available
> that are hard to shut down without shutting down things the
> powers-that-be care about:

> - there's still an awful lot of ham, CB, and emergency radio gear
> floating around
> - the military spends billions on spread-spectrum, software-defined
> radios and ad-hoc mesh networking - with most of the core technology
> published in publicly-available journals
>
> Would I like to have a data-capable smart phone / hot spot; that
> relies totally on ad hoc mesh networking for connectivity? Of
> course. Getting enough other people to adopt them, for them to
> actually be usable, is a much harder problem. It's hard enough for a
> wireless carrier to roll out a new generation of technology - after
> investing billions in new phones, towers, radios, roaming agreements,
> and marketing to make the new phones useful on day one.
>
>> Social infrastructures are a critical topic for discussion, and I
>> welcome it as part of Contact Con. But the technology underlying
>> those social arrangements IS relevant and of central concern here.
>
> Sure... but I expect that useful results are more likely to come out
> of MILCON and DEFCON.
>

I should add that there seems to be an awful lot of ready-to-use mesh
networking gear coming out for the first responder community -- stuff
that's designed to be carried into wilderness areas by firefighters, and
into earthquake zones and other places where existing infrastructure has
been knocked out. That's the kind of stuff that might actually be
useful, and available, when needed.

[Charles N Wyble]

On 05/14/2011 11:03 AM, Miles Fidelman wrote:

> Miles Fidelman wrote:
>>
>
> I should add that there seems to be an awful lot of ready-to-use mesh
> networking gear coming out for the first responder community -- stuff
> that's designed to be carried into wilderness areas by firefighters,
> and into earthquake zones and other places where existing
> infrastructure has been knocked out. That's the kind of stuff that
> might actually be useful, and available, when needed.
>

Yep. I was instrumental in building exactly that for a previous
engagement. A kit that could be dropped into a katrina type situation.
Villagetelco + serval project = dirt cheap, rugged, anywhere comms.

[Miles Fidelman]

Miles Fidelman wrote:
> I should add that there seems to be an awful lot of ready-to-use mesh
> networking gear coming out for the first responder community -- stuff
> that's designed to be carried into wilderness areas by firefighters,
> and into earthquake zones and other places where existing
> infrastructure has been knocked out. That's the kind of stuff that
> might actually be useful, and available, when needed.

And for some really interesting stuff, google "opportunistic
communications" and "pocket-switched networks" (yes, that's pocket, with
an "o"). You'll find stuff like this:
www.cl.cam.ac.uk/~jac22/talks/p2p-keynote-6.9.6.ppt.gz

[Mark Roest]

Hello Colin and all,

Regarding "a decent model of the social graph"

The core underlying reasons for doing all of this are to enable people to:

1.  find those who want to focus on the same changes they want to make, those who can teach them how to be effective, those with resources to share, designers of solutions, and others on a parallel path with whom to collaborate in efforts to scale.

2. do business with those who form a natural, sustainable, empowering economic ecosystem with them.

Given the immensity of the amount of knowledge and information that exists and is searchable today, the question is how to turn a waterfall of data into a kitchen sink faucet of relevant, peer-reviewed information, knowledge, references, and wisdom.

The tools commonly available today do only a little to solve the problem. There are two major reasons for the enormity of the problem: 

1. the fact that the dominant business (and academic) model today seeks to create differentiated branding and to siphon off parts of each market from other businesses (or lines of research) in the efforts to survive, to thrive, and to become dominant. Information is now the primary way to achieve this.

2. the fact that with 827 identified major land eco-regions on this planet, and many  ocean and sea eco-regions, each of which has many local ecosystems, different solutions are feasible for and relevant to different places.

As a lifelong bookworm (using the Dewey Decimal System) and former library page (using the Library of Congress system), I know well that a little categorization goes a long way to achieving focus of information. The filtering solution to the waterfall of miscellaneous data is to tag everything that is truly useful for a specific eco-region, and valuable for a specific culture. Cross-referencing these two tags is what drastically reduces the amount of data that is returned on a search. If it is done in the context of a geographic information system, overlaid on a digital earth imaging platform such as Google Earth or WorldWind, it becomes the most powerful information system the world has ever seen.

The quality control system can consist of both peer review by users, and expert evaluation and modeling of impacts by people in the fields of ecology, bio-mimicry, natural alternatives, and alternative technologies, all working in a supportive, effective information environment.

Since over half of humanity has now been displaced (one way or another) from the countryside to the city, search needs to return any combination of tags that work for the people doing it. They may be people from multiple cultures who come from the same biome (multiple eco-regions in a major land cover type), just getting started in the process of finding ways to return to their homelands and revive their cultures, economies, and local environments. They may be a single village cooperative, or a single individual in a village, seeking solutions for problems that are weakening or killing the people in their village, and leaving children without parents. They may be industrial designers or food garden researchers looking for a way to share their knowledge, and make a living doing it. The may be students in a university in the developing world, seeking both a way to make a living and raise families, and ways to save their people and the natural world around them from devastation and suffering.

While these people are not highly visible on the internet right now (but are increasingly so), they constitute most of the 4 to 5 billion people who have never accessed the internet in their lives, most of whom live on less than $4 per day. The work of this group and other like it, when combined with the open hardware initiatives, and marketing through the diaspora and the Blessed Unrest that Paul Hawken writes about, some of which is in the WiserEarth.org directory, can change both of those facts in a matter of a few (<5) years.

The three 'killer applications' for that combined, synergistic effort, are:

1. basic communication capability (why most people in industrial societies carry cell-phones and have to be warned not to use them while driving),

2. access to the knowledge they need to change their living conditions and save their natural environment, and to create systematically organized, socially just, sustainable local, eco-region, national and continental economies.

3. access to an integrated public health system (using WorldVistA.org's VistA-EHR health information management system)

Multiple specific applications and portals are also needed, on the order of millions of them; people will learn how to develop them down to the village and neighborhood level.

I am copying some people who bring important resources and practices to the effort to change the world. Eventually many strands of work will be woven together, and we will have the platform that empowers all of humanity.

Regards,

Mark Roest

On Thu, May 12, 2011 at 9:12 PM, Colin Hawkett <haw...@gmail.com> wrote:

The issues I see -

1. Where is your profile page served from? Where is your personal data? Does it remain available if you are not online? If it does, how is it secured?

1. People see BitTorrent as the means they get.... 'stuff'. Are they comfortable bringing that activity mainstream? Anonymity is a big deal in this space, which can be a bit of a pain for a social platform, depending on how you look at it. I can't see people getting 'stuff' while logged in as themselves.
2. Discovery/Search - for exactly the same reason you need pirate bay or equivalent, you need a central server to locate people, etc. - unless they have some pretty funky new distributed tech.
3. Synchronisation between nodes - e.g. my desktop/laptop/tablet and within the social graph. This is interesting in general, but gets harder when the data is volatile (which it generally is not with BitTorrent in its current form). e.g. I update my status - how does that get reliably to everyone... in near-real-time? Distributed twitter.... tough.
4. Authentication - not clear to me how this is managed.

Essentially the same things Diaspora has to worry about, plus not being perceived as 'social'. The user base is a big deal, but it is the weakest form of user base because they are almost entirely anonymous. One slight upside is that this user base probably has a higher than usual proportion of 'early adopter' psychology.

My take is that a logical network topology based on the social graph is a good thing, and could improve the viability of a P2P physical topology, but it needs to start with a decent model of the social graph, and I'm still not seeing that - it always seems to be tacked on to some other property that people are trying to leverage for existing network effects.

[Mark Roest]

PS

Check out the archives on KPFA.org for the show on Saturday, 5-14-11, from 9 to 10 a.m.; the last segment is on PulseWire, which is part of the WorldPulse that is  in the Cc-list here, and is another great, distributed marketing channel for real solutions.

M

[Adam Fisk]

Just to quickly chime in, to me Chrysalis is a truly exciting development, filling in many of the glaring technical gaps with things like Diaspora. First off, I don't consider Diaspora to be P2P. It's essentially just federated servers with no P2P distribution as far as I can tell (please correct me if I'm wrong if there are people on the list who have cracked open the code more than I have). P2P on the other hand does a few key things well that I think are vital to keep in mind for any implementation, whether it's a P2P social network or something else. It's a pretty specific tool useful in specific scenarios, to be applied as surgically as possible:

1) Distributing large files

2) Minimizing points of control

Things it does not do well:

1) Distributing small files (latency is too high, multiple peers don't matter)

2) Database-like stuff

The "minimizing points of control" point has all sorts of implications, of course, from limiting liability in the case of things like Gnutella to making it difficult to censor content for governments from China to Iran (see my own forthcoming Lantern project). There are also implications for things like distributed implementations of WikiLeaks and for minimizing control over the world's digital content -- the angle Diaspora is attempting to take.

So why is Chrysalis interesting? It takes one very specific thing that P2P does well -- distributing large files -- and brings that power to a social network. To me, this is what I *thought* Diaspora was originally doing but is not. The reason they're not is that what BitTorrent is doing is really hard. There's all sorts of crazy stuff going on underneath to make that possible, from UTP to NAT/firewall traversal to a DHT and the BT downloading algorithm itself. There are lots of large files on social networks, particularly videos and photos, and Chrysalis properly uses P2P to distribute them. This has all sorts of cool scalability implications, not to mention implications for loosening the grip of centralized authorities over content.

Where this starts to break down is the little stuff, the small files like profiles and the social graph itself that are hard to do in a distributed fashion. The Diaspora federated approach is interesting in that regard. I'm unclear on what approach BT is taking there -- possibly just a centralized database? That would certainly make sense.

Overall, though, to me this whole quest is a matter of properly applying these different technologies to different parts of the architecture. The answer is not a "p2p social network" - that's too simplistic. To me the best architecture involves using p2p where appropriate, possibly federation where appropriate, and centralization where appropriate.

Great to see so many sharp comments on the list. I'd love to delve more deeply with everyone but am way too caught up with some other things (Lantern/Brave New Software) to follow it as much as I would like, so I apologize in advance. The above sketches out the high level issues as I see them.

-Adam

[Kaliya Hamlin]

On May 14, 2011, at 2:46 PM, Miles Fidelman wrote:

Jon Lebkowsky wrote:

I'm concerned that this conversation addresses how to build rather than what to build. We have a general sense what problem we're trying to solve here,

Actually, I sure don't.  I originally thought the focus was on distributed, decentralized INFRASTRUCTURE  - e.g., wireless mesh networks that are not dependent on carriers, and are not subject to disruption or centralized control.

This feels like it is "doable" 

The focus seems to have shifted to decentralized applications - again, less vulnerable to control and disruption, but dependent on the current Internet infrastructure.

I think people here underestimate the scope and scale of this infrastructure - it can't be "replaced" by some in the air radio links...There is a role for "REAL" infrastructure - fiber in the ground on a massive scale. I think the European or at least French model where the cities are wired with fiber optics and that these are owned by the cities and companies then pay to run services over the wires is a very good model. 

You know I guess I first got a real sense of this hanging out on Gordon Cook's list 

http://www.cookreport.com

and even did a small 35 person unconference with him focused on the cable in the ground to homes owned by municipalities/other government levels (New Zealand I think it is the national state) and how this was a viable model.  

This is about the sort of infrastructure we do have...and some of the issues/opportunities

http://www.cookreport.com/images/stories/CookMindMap.jpg

This is interesting on a global collaborative architecture

http://www.cookreport.com/index.php?option=com_content&view=article&id=262:fast&catid=38:current-issues&Itemid=73

Another person who I learned a lot from was Esme Vos is in my network who is basically the world's leading expert on this - http://www.muniwireless.com/

These are two very different problems, that require very different solutions.




--
In theory, there is no difference between theory and practice.
In<fnord>  practice, there is.   .... Yogi Berra

Kaliya Hamlin -  Identity Woman 

Executive Director, Personal Data Ecosystem Consortium - http://www.personaldataecosystem.org

Leader,  Internet Identity Workshop  http://www.internetidentityworkshop.com

Identity Woman Blog - http://www.identitywoman.net

Twitter - @identitywoman

Skype - Identitywoman

Phone - 510 472-9069

IM Handles (best way to "talk" to me)

AIM/IChat:  kal...@mac.com

Yahoo!: earth...@yahoo.com

GTalk: identi...@gmail.com

MSN: identi...@hotmail.com

[Jon Lebkowsky]

Esme would actually be a great addition to this group - also David Isenberg and David Reed; I've learned a lot from them about the network industry and its perspectives.

~ Jon

[Bryce Lynch]

On Fri, May 13, 2011 at 12:04, Colin Hawkett <haw...@gmail.com> wrote:

> You grab a torrent file from some website - it contains a list of trackers
> Your BitTorrent client then contacts those trackers to find out which other
> connected peers have part or all of the file you want

That hasn't been the only way of setting up a torrent for a couple of
years now. DHTs have been part of the BitTorrent protocol since...
let me check... May of 2005?

--
The Doctor [412/724/301/703]
http://drwho.virtadpt.net/
"I am everywhere."

[Colin Hawkett]

On Tuesday, 17 May 2011 01:06:28 UTC+10, The Doctor [412/724/301/703] wrote:

On Fri, May 13, 2011 at 12:04, Colin Hawkett <haw...@gmail.com> wrote:

> You grab a torrent file from some website - it contains a list of trackers
> Your BitTorrent client then contacts those trackers to find out which other
> connected peers have part or all of the file you want

That hasn't been the only way of setting up a torrent for a couple of
years now.  DHTs have been part of the BitTorrent protocol since...
let me check... May of 2005?

Thanks for the info - I wasn't aware of that, although I'm not sure it affects the analysis too much - some points/questions -

1. It seems to me that a DHT implementation still requires a torrent file specifying tracker(s) to kick off the DHT based discovery process, and that populating that torrent file with a 'good' node (i.e. reliable server) is still a recommended strategy. It seems the main difference is that the tracker(s) contained in the torrent file will query the network according to the DHT algorithm, rather than holding info about all of the peers themselves - i.e. every node is a tracker. Am I right in thinking that the reliability/availability of the tracker(s) in the torrent file is just as important with either algorithm?
2. I was having trouble working out how widely the spec. is actually used - are most torrent files following the DHT spec. these days? I assume most clients (regardless of the torrent file format) are using DHT as the primary discovery mechanism?
3. As far as I can tell DHT isn't part of the protocol, and remains in draft status. Where did 2005 come from? I'm probably looking in the wrong place :)
   

Wouldn't be surprised to be wrong on all counts :) Cheers,

Colin

[Colin Hawkett]

On Monday, 16 May 2011 08:40:56 UTC+10, Kaliya - Identity Woman wrote:

The focus seems to have shifted to decentralized applications - again, less vulnerable to control and disruption, but dependent on the current Internet infrastructure.

I think people here underestimate the scope and scale of this infrastructure - it can't be "replaced" by some in the air radio links...There is a role for "REAL" infrastructure - fiber in the ground on a massive scale. I think the European or at least French model where the cities are wired with fiber optics and that these are owned by the cities and companies then pay to run services over the wires is a very good model. 

Agree heartily with this sentiment - the model you describe is a proprietary, centralised one, but is more effective at delivering value to users than corporate owned infrastructure. Surely these alternatives to 'open distributed' should be in the mix for a discussion about the next-net?

It might even be the case that the basic physical architecture isn't such a burning issue, and consequently the problem of creating self-organising, self-governing, empowered communities is mostly a software/cultural/political one. I really like the problem/vision statement (IMO group mechanics are the critical element of a properly wired global brain), but I'll admit I do struggle with the idea that a shiny new hardware infrastructure is of critical importance for this.

I'm wondering in general whether people feel the url (rather than the title) of the group is constraining the next-net discussion to distributed/decentralised, or whether this architecture is in fact considered to be obviously 'correct' going forward? I guess it comes back to defining the problem...

[Bryce Lynch]

On Mon, May 16, 2011 at 13:52, Colin Hawkett <haw...@gmail.com> wrote:
> It seems to me that a DHT implementation still requires a torrent file
> specifying tracker(s) to kick off the DHT based discovery process, and that

It doesn't. I've found a couple of torrents out there that only
supported DHT, but the downside is how long they take to get running
due to the discovery process.

> populating that torrent file with a 'good' node (i.e. reliable server) is
> still a recommended strategy. It seems the main difference is that the

To get a swarm moving fast, it is. Kademlia (which is the DHT
implementation used by BitTorrent) can bootstrap itself from only one
node in the swarm. The identifiers of each node in the swarm are
hashes, just as the files handled by a particular torrent are hashes.
Each node, after computing its identifier, then tries to find other
nodes by trying to contact permutations of its identifier (one up and
one down, ten up and ten down, until it finds a) a node, and b) a node
that knows something about the swarm it wants t join).

> tracker(s) contained in the torrent file will query the network according to
> the DHT algorithm, rather than holding info about all of the peers
> themselves - i.e. every node is a tracker. Am I right in thinking that the
> reliability/availability of the tracker(s) in the torrent file is just as
> important with either algorithm?

To the best of my knowledge, this is not the case. DHT aims to
replace trackers entirely, if I'm reading the specs correctly.

> I was having trouble working out how widely the spec. is actually used - are
> most torrent files following the DHT spec. these days? I assume most clients

It depends.

Some trackers explicitly forbid distributed torrents to keep the files
private, and when torrents are uploaded DHT support is stripped off.
Others are pretty much DHT only these days (like the Pirate Bay).
Most trackers I've seen support both.

Most modern torrent clients support DHT, and add the necessary
information to the .torrent files they generate automatically unless
you tell them not to.

> (regardless of the torrent file format) are using DHT as the primary
> discovery mechanism?

I would say it's probably about half-and-half, in my experience.

> As far as I can tell DHT isn't part of the protocol, and remains in draft
> status. Where did 2005 come from? I'm probably looking in the wrong place :)

https://secure.wikimedia.org/wikipedia/en/wiki/BitTorrent_%28protocol%29#Distributed_trackers

http://torrentfreak.com/bittorrent-2005-part-6-torrent-clients/

http://azureus.sourceforge.net/changelog.php

[Colin Hawkett]

On Wednesday, 18 May 2011 01:25:17 UTC+10, The Doctor [412/724/301/703] wrote:

On Mon, May 16, 2011 at 13:52, Colin Hawkett <haw...@gmail.com> wrote:
> It seems to me that a DHT implementation still requires a torrent file
> specifying tracker(s) to kick off the DHT based discovery process, and that

It doesn't.  I've found a couple of torrents out there that only
supported DHT, but the downside is how long they take to get running
due to the discovery process.

Ok, that aligns with my understanding of distributed discovery in a generalised sense - i.e. performance and/or reliability downsides. Also the blind seeking mechanism can generate a bit of traffic, and in a system where a lot of searches are done for things which don't exist at many nodes it can get quite noisy. So the population of the torrent file with the good node would not only serve the purpose of making it faster, but also deliver a fairly significant reduction in overall network traffic.

The other problem is reliability - e.g. if the one node that you use to bootstrap doesn't have a connected path to the nodes you want to find, then the search fails.

> populating that torrent file with a 'good' node (i.e. reliable server) is
> still a recommended strategy. It seems the main difference is that the

To get a swarm moving fast, it is.  Kademlia (which is the DHT
implementation used by BitTorrent) can bootstrap itself from only one
node in the swarm.  The identifiers of each node in the swarm are
hashes, just as the files handled by a particular torrent are hashes.
Each node, after computing its identifier, then tries to find other
nodes by trying to contact permutations of its identifier (one up and
one down, ten up and ten down, until it finds a) a node, and b) a node
that knows something about the swarm it wants t join).

Ok but it needs to get that one node from somewhere - it seems that if most people's use case is to click on a torrent file to launch their client, and that torrent doesn't identify any nodes, then the system doesn't work - i.e. it bootstraps as a network of one node (itself). It could use a persistent cache from its last execution, but that carries the risk of being stale. Either way it needs to get the seed from somewhere. Agree that in a social network scenario, the click-to-open use case is probably less common, and a more connected system is probably the norm, but in that system the noise from all the distributed searches would be fairly large without decent seeds.

implementations aren't specs - but I agree adoption of the draft seems pretty broad with a long history :)

anyway thanks for the info, appreciated.

[Isaac Wilder]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Wanted to jump in here.
Comments inline.

On 05/17/2011 09:10 AM, Colin Hawkett wrote:
> On Monday, 16 May 2011 08:40:56 UTC+10, Kaliya - Identity Woman wrote:
>
>> The focus seems to have shifted to decentralized applications -
>> again, less vulnerable to control and disruption, but dependent
>> on the current Internet infrastructure.

I agree, and while I think that talking about which apps will be
necessary for a distributed network to thrive is a worthwhile pursuit,
I do think that we would do well not to focus quite so heavily on the
application layer. Category killers will emerge, as Charles likes to
say. The talk of a distributed social knowledge engine is fascinating,
but I think that ultimately that tool would be of little use if it
were implemented as a software solution on top of the existing
hierarchical network stack.

>
>
> I think people here underestimate the scope and scale of this
> infrastructure - it can't be "replaced" by some in the air radio
> links...There is a role for "REAL" infrastructure - fiber in the
> ground on a massive scale. I think the European or at least
> French model where the cities are wired with fiber optics and
> that these are owned by the cities and companies then pay to run
> services over the wires is a very good model.

It's true: the scope and scale of telecommunications infrastructure
can be staggering. The struggle to own and operate that infrastructure
cooperatively can sometimes seem infinitely uphill, but as I see it,
it is the only way. I, too, look at the municipality as an emergent
locus of human sovereignty and determination. I don't have quite the
same faith in local government, but I do believe that we need to
organize and construct infrastructure on the local scale. I recognize
that we can't simply replace fiber infrastructure with radio links,
but I think that it is necessary to wrest the control of terminal
nodes away from the telcos. When users start to talk directly to one
another, they will realize also that they are stronger together, and
capable of constructing their fair share of the global network.

>
> Agree heartily with this sentiment - the model you describe is a
> proprietary, centralised one, but is more effective at delivering
> value to users than corporate owned infrastructure. Surely these
> alternatives to 'open distributed' should be in the mix for a
> discussion about the next-net?

There is no harm in discussing alternatives to the open and
distributed model, but it would take some real effort to convince me
that proprietary technologies can compete with free ones, in the long
run. I'm willing to have the discussion, but just know that I'll be
pulling for freedom in every case.

>
> It might even be the case that the basic physical architecture isn't
> such a burning issue, and consequently the problem of creating
> self-organising, self-governing, empowered communities is mostly a
> software/cultural/political one. I really like the problem/vision

> statement (IMO group mechanics are /the/ critical element of a

> properly wired global brain), but I'll admit I do struggle with the
> idea that a shiny new hardware infrastructure is of critical
> importance for this.

Again, no harm in having the discussion, but I disagree vehemently.
Physical architecture, actual co-ownership, the physical locations of
the bits are at the root-essence of the struggle. There is a
software/cultural/political component to the problem, no doubt, but
all the distributed software in the world is no good if it can be shut
down at the behest of a powerful government or MNC. I also wish to
facilitate the creation of communities that are self-organizing,
self-governing, and empowered, not to mention resilient and
sustainable. I believe that most effective principle around which to
organize is the cooperative construction of infrastructure that gives
the power back to the people. I believe that if we can get people to
assent to this project, we can get them to assent to their humanity,
and thereby to what is good for humanity. I think it is the only way.

>
> I'm wondering in general whether people feel the url (rather than
> the title) of the group is constraining the next-net discussion to
> distributed/decentralised, or whether this architecture is in fact
> considered to be obviously 'correct' going forward? I guess it comes
> back to defining the problem

> <https://groups.google.com/d/topic/building-a-distributed-decentralized-internet/ZWo7ffAfCNM/discussion>...
In fact, I wish that it would constrain the discussion a bit more.
Well... that's not exactly true. I'm glad that people are talking
about distributed applications, because it's a necessary part of the
movement, I just feel that there's a danger in addressing only the top
layer of the network stack. Sure, it is the layer that is most
visible, with which we usually interact, but it is not the one that
determines the fundamental politics of the network. If we truly want
to live sustainably, if we truly want to help humanity flourish, if we
truly want to be free, then we must set our sights on the physical
layer, and we must not miss the mark.



take care,
imw
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)

iQEcBAEBAgAGBQJN01RHAAoJEA8fUKCD77NLnAkH/iBIBsMIXeVf/sw5PgXnyOyQ
LHCVkL25VROGbow1+ZNQF3im/VePR9EIphwhv/6wkNGnIDOsktvM2I2i77TV0Ey6
4buAlgECLo/jNoKSOQP8WZqF7acgJXFD8FHeTUc+RACjZ0lnftC1WM9xe4HtP6K3
s7e3MXKn/Un+hLXc3h5UBpA7Gv0PHtpMCAB6gBOMTeKYy0fS6ksFXL0GLFDrzIwV
iL+SGtdtZ+yrQ7JNCFu7kRLCjf/HKRare91JM+Wz6Zs0amwmL4U72zubt/y4HvC5
Lb32IVvVeueJwioD/SWW7cjq564Y5bGZPj7msBNp1j44VmlcXb+pCcV8xFTokFM=
=8+Ch
-----END PGP SIGNATURE-----

[Colin Hawkett]

Thanks Isaac - I think this is a highly worthwhile discussion to have, and perhaps gets to the core of some occasional polarisations.

I think we (and most here) more-or-less agree on the problem, and the desired outcome, but see different paths. The problem seems to be that the organisations (corporate, government) which control the infrastructure don't do so in the public interest. From the corporate perspective, I'm not even sure it is legal for them to do anything but act in the interests of their shareholders. I'd say these corporations which we revile are simply the product of a capitalist system designed to produce them. 

The desired outcome, I think most here agree, is that infrastructure should be managed in the common/public interest. So our options are to either create new infrastructure with the goal that it either has good management (or is unmanageable? or can't be corrupted?), or get good management of what we have. Stepping back from the infrastructure issue a little, I think we would mostly agree that it is a specific manifestation of something we see repeated throughout society - corporate interest trumps public interest.

Government is supposedly the entity that represents the interests of the people, and has the power to legislate and create policy that constrains the capitalist (or whatever) system for this purpose. To me it is fairly clear that this part of the system is broken - the people's interests are not represented, and the capitalist system, with no ethical basis, controls the controller - like some quasi-sentient nuclear reactor that manages its own control rods. And it hasn't the slightest interest or capacity to understand that role from any perspective but the interests of shareholders, or increasingly just the head honchos.

It seems clear to me that the cause of the problem is dodgy government, not dodgy infrastructure. From this perspective there are a few issues with the new infrastructure approach -

a) if it ain't broke don't fix it.

b) we'll still have dodgy government, and they're gonna make the new stuff annoyingly difficult to run if we expect them not to control it (some stuff [1][2] around bitcoin is already showing this behaviour).

c) we'll still have dodgy government, so corporate interest trumps public interest in all the other things we have yet to build new ones of.

d) building new ones 'cos they buggered up the old ones' seems like an unnecessarily hard slog.

So I guess my logic is that if infrastructure isn't the problem, then how about the dodgy government thing? What can we do from a next-net perspective?

[Miles Fidelman]

Colin, a very nice way of putting things! A few comments follow below
the quotes:

Colin Hawkett wrote:
> The desired outcome, I think most here agree, is that infrastructure
> should be managed in the common/public interest. So our options are to
> either create new infrastructure with the goal that it either has good
> management (or is unmanageable? or can't be corrupted?), or get good
> management of what we have. Stepping back from the infrastructure
> issue a little, I think we would mostly agree that it is a specific
> manifestation of something we see repeated throughout society -
> corporate interest trumps public interest.
>

<snip>

> It seems clear to me that the cause of the problem is dodgy
> government, not dodgy infrastructure. From this perspective there are
> a few issues with the new infrastructure approach -
>
> a) if it ain't broke don't fix it.
> b) we'll still have dodgy government, and they're gonna make the new
> stuff annoyingly difficult to run if we expect them not to control it
> (some stuff [1]

> <http://launch.is/blog/l019-bitcoin-p2p-currency-the-most-dangerous-project-weve-ev.html>[2
> <http://launch.is/blog/l020-is-bitcoin-the-wikileaks-of-monetary-policy.html>] around

> bitcoin is already showing this behaviour).
> c) we'll still have dodgy government, so corporate interest trumps
> public interest in all the other things we have yet to build new ones of.
> d) building new ones 'cos they buggered up the old ones' seems like an
> unnecessarily hard slog.
>
> So I guess my logic is that if infrastructure isn't the problem, then
> how about the dodgy government thing? What can we do from a next-net
> perspective?

I'd also add "how about this "dodgy corporate thing?" :-)

A couple of thoughts present themselves:

1. There are quite a few models of infrastructure ownership to draw on:

- user ownership of infrastructure (works well for large corporations,
not so well for individuals)

- condominiums and cooperatives (works well for houses, apartment
buildings, housing complexes, farmers, and a surprising number of
electric and telephone companies)

- municipal ownership (e.g., of waterworks, electric utilities; with
municipal telecom utilities among the few that offer gigE network
services) -- one can argue that a local government is essentially a
condo association writ large

- various forms of user-owned financial institutions (credit unions,
mutual banks and insurance companies, etc.) - along with a growing "move
your money" campaign (moveyourmoneyproject.org)

- various forms of purchase aggregation (i.e., multiple people or groups
buying things together - in order to increase leverage in the market,
such as insurance purchased through a professional association)

- various forms of affinity programs (e.g., discounts negotiated by
AARP, AAA, etc., certified products like Free Trade Coffee) - again,
ways to shape a market

2. There are examples of each of these applied to telecom.
infrastructure (e.g, corporate and academic networks, CREDO telephone
service, municipal telecom. utilities, etc.)

3. Put one or more of these together, coupled with a grass-roots
"move-your-telecom" campaign, and something might be doable.

Miles Fidelman

[Colin Hawkett]

Ha - so the thinking is that if we have a sound government, then thats the end-game - their job is to create policy and legislate for the public interest, ergo the dodgy corporate thing will get sorted. Just about everything gets sorted if we can trust our government to act in our interest.

A couple of thoughts present themselves:

1. There are quite a few models of infrastructure ownership to draw on:

- user ownership of infrastructure (works well for large corporations,
not so well for individuals)

- condominiums and cooperatives (works well for houses, apartment
buildings, housing complexes, farmers, and a surprising number of
electric and telephone companies)

- municipal ownership (e.g., of waterworks, electric utilities; with
municipal telecom utilities among the few that offer gigE network
services) -- one can argue that a local government is essentially a
condo association writ large

- various forms of user-owned financial institutions (credit unions,
mutual banks and insurance companies, etc.) - along with a growing "move
your money" campaign (moveyourmoneyproject.org)

- various forms of purchase aggregation (i.e., multiple people or groups
buying things together - in order to increase leverage in the market,
such as insurance purchased through a professional association)

- various forms of affinity programs (e.g., discounts negotiated by
AARP, AAA, etc., certified products like Free Trade Coffee) - again,
ways to shape a market

So I'm thinking here too that if we trust government, and it represents our interests, then through the policy making process, which is more or less a constantly changing re-assessment of the landscape for the good of the people, the right forms of infrastructure ownership will fall out at the places where they need to. It seems to me that the most obvious way to make this work is to have the policy making proces publicly crowd-sourced, and transparent so corporations don't have the ability to corrupt the system by corrupting a small number of individuals.

[Venessa Miemis]

On Wed, May 18, 2011 at 2:34 PM, Colin Hawkett <haw...@gmail.com> wrote:

So I'm thinking here too that if we trust government, and it represents our interests, then through the policy making process, which is more or less a constantly changing re-assessment of the landscape for the good of the people, the right forms of infrastructure ownership will fall out at the places where they need to. It seems to me that the most obvious way to make this work is to have the policy making proces publicly crowd-sourced, and transparent so corporations don't have the ability to corrupt the system by corrupting a small number of individuals.

saw this on liberationtech listserv today. use google chrome to translate from italian. 

"the web site http://www.comunalimilano2011.it/ that uses the OpenDCN platform, others on this list may not have seen it. It's an outgrowth of the Milan Civic Network which has been helping to cultivate community and civic life in Italy for 15 or so years."

[Miles Fidelman]

Colin Hawkett wrote:
> Ha - so the thinking is that if we have a sound government, then thats
> the end-game - their job is to create policy and legislate for the
> public interest, ergo the dodgy corporate thing will get sorted. Just
> about everything gets sorted if we can trust our government to act in
> our interest.

<snip>

> So I'm thinking here too that if we trust government, and it
> represents our interests, then through the policy making process,
> which is more or less a constantly changing re-assessment of the
> landscape for the good of the people, the right forms of
> infrastructure ownership will fall out at the places where they need
> to. It seems to me that the most obvious way to make this work is to
> have the policy making proces publicly crowd-sourced, and transparent
> so corporations don't have the ability to corrupt the system by
> corrupting a small number of individuals.

Well that's not what I was suggesting. I was suggesting that money =
power.

It's pretty hard to change government without money and power -
particularly when a lot of what we buy funds the status quo (every time
you drink out of a Dixie Cup, or wipe your rear with a Georgia-Pacific
paper product, you're funding the Koch Brothers and the Tea Party!).

I'm suggesting that rather than tilt against windmills to change
government, and wait for government to solve telecom problems, that we
take more direct action - become our own corporations and/or their
equivalent (cooperatives, local governments, etc.).

In banking, that means taking money out of Bank of America accounts and
putting it into community banks and credit unions. In telecom, it's
about municipal networks.

Change the economics first, then change Government.

[Jon Lebkowsky]

On Wed, May 18, 2011 at 2:04 PM, Miles Fidelman <mfid...@meetinghouse.net> wrote:

I'm suggesting that rather than tilt against windmills to change government, and wait for government to solve telecom problems, that we take more direct action - become our own corporations and/or their equivalent (cooperatives, local governments, etc.).

In banking, that means taking money out of Bank of America accounts and putting it into community banks and credit unions.  In telecom, it's about municipal networks.

Change the economics first, then change Government.

That's the most realistic way to go.  I often remember how someone once asked Gary Snyder what to do about "the establishment," and he said "ignore 'em."  What we're talking about here, generally, is evolution, and some people don't want to evolve. That's okay, evolution is not for everybody; nobody should be forced to evolve.

~ Jon

--
Jon Lebkowsky (@jonl)
Polycot Associates

[Patrick Anderson]

Colin Hawkett wrote:
> From the corporate perspective, I'm not even
> sure it is legal for them to do anything but act
> in the interests of their shareholders.

Ah, but what if the Users *were* the shareholders?

And what if the ROI for those owners was the Product itself?

And, for those who are not yet part of that organization,
what if the Profit charged against them was treated as
though they were making a late investment?

A corporation structured in this way would not have the
traditional conflict between shareholders and users because
those two groups would be one and the same!

We can start by organizing Users to pre-pay for the
Product (internet access in this case), but treat those
funds as investment and real co-ownership in that corp.

We, the Users, *already* pay all the Costs of operation
anyway, we are just paying them late, and are therefore
also required to pay Profit.

When a corporation is User Owned, and the return for
those investments is the Product itself, then those
owners do not buy the Product because they own it
already as a side-effect of their co-owing the Sources.

Those Owners must still pay all the Costs, but cannot
pay Profit, for who would they pay it to?

[Miles Fidelman]

Patrick Anderson wrote:
> Colin Hawkett wrote:
>
>> From the corporate perspective, I'm not even
>> sure it is legal for them to do anything but act
>> in the interests of their shareholders.
>>
> Ah, but what if the Users *were* the shareholders?
>

Umm... they're called:
- mutual banks and insurance companies
- condominiums
- cooperatives
- associations
- .....

[Charles N Wyble]

Sounds like a great idea. Go build it! Let us know when we can join up.

[Isaac Wilder]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

>>

>> When a corporation is User Owned, and the return for those
>> investments is the Product itself, then those owners do not buy
>> the Product because they own it already as a side-effect of their
>> co-owing the Sources.
>>
>> Those Owners must still pay all the Costs, but cannot pay Profit,
>> for who would they pay it to?
>
>
> Sounds like a great idea. Go build it! Let us know when we can join
> up.
>

Does sound like a great idea, but I disagree with the second
sentiment. We're building the same thing.

Hey Patrick, you wanna build with us?

FNF could be exactly the corporation that you're talking about.

take care,
imw
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)

iQEcBAEBAgAGBQJN1HRZAAoJEA8fUKCD77NLuhAH/0g36tCcFD6gSwwz07EhtBAC
DS3iUTKSUIcIGCxtboPOV6TNTc0mDpkVlQa8vnI6oSBEvC31B8HY03FRSBENTT/9
M0+YHuEwK21BZLYeqw6jPqXzd9JepwO3MUNTXAnog9CS2b/iUerLVW54K/rwAVEr
QOLsvyhA4kvfWon4I8odRITR7nDAQmWEIDLAIT1SUaI9No3eDOgfpzvSxZ2DyGpa
KFuA6MREIlaPXTCSPmobJ7ek67gb8x2EqBLKmwGiwHW8ZbzlWGScuPBUtkwkd6w0
4Mbt2WvauaTf/6WtcAozSTLz2YypaqBQC+sCI4KsWB44u9PR9j8goIbiM3teIJw=
=6/kh
-----END PGP SIGNATURE-----

[Charles N Wyble]

On 05/18/2011 08:37 PM, Isaac Wilder wrote:
>

> Sounds like a great idea. Go build it! Let us know when we can join
> up.

Does sound like a great idea, but I disagree with the second
sentiment. We're building the same thing.

Of course. You are correct. I meant I would love to see the fruits of efforts
around making this happen. All of the documentation, policies, frameworks
etc to make it an operational reality.

Hey Patrick, you wanna build with us?

Yes. Please. Sorry for sounding elitist or anything. I look forward to having the
FNF be exactly that model. Or potentially a separate utility like organization
that we all would pay into to maintain the infrastructure.

FNF could be exactly the corporation that you're talking about.

Indeed it could be.

take care,
imw

[Patrick Anderson]

Miles Fidelman wrote:
> Umm... they're called:
> ...
> - cooperatives

A conusmer cooperative is very different from what I propose in a few
ways, but maybe the most important is the way the Consumer/Owners
*buy* the Products back from the group instead of being the owners
already.

This means the cooperative actually charges Price above Cost (Profit)
against those Consumer/Owners, and then tries to get rid of that
hot-potato in a variety of ways, but most usually end up causing the
cooperative to suffer from consolidation of control into the hands of
the originators who gain ownership of all the growth caused by those
overpayments.

The User Ownership model I propose is more like what would happen in
the smallest of scenarios.

For example, imagine you and your neighbor run a network wire between
your houses. Obviously you must collectively pay the Costs of
purchasing the equipment and the Costs of the labor to install it, and
the Costs of supplying the electricity, but you would not and COULD
not pay Profit because you are not each buying the Product back from
the collective two of you -- that would be nonsense, though that is
exactly what we will face if we give control to a municipality, and
cooperative, an association, and probably nearly any other
organizational form currently in existence...

[Jeetu Golani]

Hi,

First, an excellent discussion, a lot to learn here :)...some quick thoughts:

- As Jan and others here have pointed out a better understanding of the crux of the problem that is to be addressed is important.

- As I understand it the need seems to fall under two broad categories within the social networking scenario :

  - Anonymity of individuals.
  - Desire for there to be no central body or government control, monitoring of data,etc.

- Anonymity of individuals can be addressed technologically but the problem is deeper and more socially rooted. To demonstrate one of the simplest mechanism for anonymity is using nicknames as has been used on IRC. This coupled with a project like Tor (https://www.torproject.org/)  could make it theoretically tougher for a snoop to link up the handle with originating IP and then retrace via ISP records the real identity of the user.

- The reason why this is still not enough and is more a social issue is that while this may make it tougher  to directly identify the user, as we establish social connections  online - friends, aquaintances, business connections
and communicate more about our thoughts, ideas, pictures, etc we give out a fair amount of information inadvertently. Moreover we wish to give out this information.

- Using the information we so willingly give out, it would not be very tough to have a fair assessment of who the individual is and a psych profile of the individual, his beliefs,etc More importantly how to get to that person through one of his social connections who may not be as security conscious.

- Now to take this one step further on a technological solution to counter the above. We could throw in encryption into the mix above and have all communication and interaction limited to only a set of people that we trust and that have for .e.g keys to be able to read each others posts over such a social network.

- The problem of the above is that this restricts the interactions to be only with individuals in the closed and trusted group of people "you already know" forging new connections needs you to interact in open view of others.

- A distinction between Facebook and Twitter is that while Facebook fosters and is primarily used to keep in touch with people you already know, Twitter makes it easier to follow and exchange ideas with people you do not know but have a common train of thought, it eases forging temporary relationships where you follow a person that seems interesting and get off the bandwagon just as easily.

- I suspect mechanisms such as the above and any model wherein you are restricted to communicating with only a certain subset of people will be rejected by the masses who are looking towards the kind of interactions Facebook, Twitter, Blackberry Messenger, etc currently provide. Since humans are social and want to reach out and this is the basic premise of on which most social networking services are formed. In short, we are blabbermouths and most of us want to shout out our opinions and accomplishments - that awesome holiday we took, that accolade we won,etc.

- Finally there is also classic real life problem that any such trust based model is only as strong as the weakest link. All it takes is getting to one individual and we have access to all of his/her social connections.

- By stating the above I do not mean to say that pursuing anonymity or mechanisms to break free from a central control body are futile. Au Contraire. The above is simply a call to better understand the problems that can indeed be solved using decentralized networks and some splendid ideas people have presented in this forum and those that may not work as well.

While Facebook/Twitter kind of scenarios may or may not benefit as much but applications such as file sharing using Bittorrent are definitely scenarios where this decentralized model do indeed work. I am hoping that my own open source project eBrainPool that allows people to directly run software from resource pools of multi-architecture and operating system mesh networked devices (eventually over the Internet) would also hopefully fall under the categories of applications that would benefit from these decentralized models (the first iteration and presentation of proof of concept of this project did in fact piggy back on the Limewire P2P network :) )

I am sure there are many such applications and scenarios that remain hitherto unknown or unexplored where these models would have maximum impact and would not be rejected solely due to our own social and psychological needs.

Finally forgive me for I know not what I speak and for the loosely knitted string of thoughts :) I'm just having a good time learning from all the awesome people here and ideas that have already been presented here :)

Thanks,

Bye for now
Jeetu
twitter: @0topcat0
http://ebrain.in | Beehive Computing
Discover and run software from any device near you - an open source (GPL) project

[Colin Hawkett]

On Thursday, 19 May 2011 11:19:38 UTC+10, Patrick Anderson wrote:

Colin Hawkett wrote:
> From the corporate perspective, I'm not even
> sure it is legal for them to do anything but act
> in the interests of their shareholders.

Ah, but what if the Users *were* the shareholders?

That model is definitely one worth considering in many situations, but with a government that we trust and acts in our interests, we get it at the macro level (and fractally I hope) by proxy - the people are the shareholders in government, and government makes the rules/policy/legislation. From my perspective, capitalism isn't actually broken - it is acting exactly as designed - a purring machine. It should be expected to seek control of anything that increases shareholder value - it has no basis in ethics. That job is for government, and a properly running government may even decide that capitalism is am unacceptable model even with regulation in the public interest. Or it may not. Whatever - once we have a government we trust and acts in our interests, I'm not so bothered. The only constant is change, and what is right today won't be right tomorrow (which highlights another feature of good government - agility). This discussion also goes with Lucas' statement that the underlying mechanism should be politically agnostic.

One thing I think the winners in the current broken system will be very happy with is the 'rage against the machine' being directed at corporations. They're happy to take the flack while we mostly ignore the fact that government is the broken/corrupted part. We can rile at corporate behaviour until we're blue in the face, they won't care. Run a marketing campaign, spin of a new company under a new name, change the product without changing it, publicly sacrifice a CEO or 2, move operations to another country - their bag is choc full of tricks to handle criticism of unethical behaviour. Total red herring. While policy and legislation support their behaviour it will continue, end of story.

[Colin Hawkett]

Cool - we're talking about the same end-game : changing government. The discussion is then various paths/possibilities for doing that. I reckon efforts to change the economics are excellent in parallel with creating a collaborative system for change - indeed such a system would facilitate and feedback with the economic endeavours, and more besides. I'd contend that a collaborative system to enhance and focus the energy for change would be very close to the top of the list of things we need to do - i.e the new mechanics of government.

[Patrick Anderson]

Colin Hawkett wrote:
> with a government that we trust and acts in our interests,

Governments cannot act in our (the Users') interest because they are
controlled by corporations.

Corporations cannot act in our interest because they are controlled by
investors that expect Profit as reward.

Treating Profit as reward causes conflict with Users because Profit
only occurs when Users lack control.

Profit is *undefined* when Users own and control the Sources of
Production because the Product is not sold - since it is already in
the hands of those who need it!

> the people are the shareholders in government,

I wish that were true, but Governments are composed of humans that
respond to the power consolidated into the hands of the owners of the
Corporations.

> capitalism isn't actually broken

Capitalism is broken because we have mistaken Profit as a reward
instead of understanding it as a measure of the Payer's dependence
upon those current Owners.

When we finally realize what Profit really is, we will treat it as an
investment from the Consumer who paid it - causing a negative feedback
loop that will cause Profit to safely approach zero as each
User/Consumer slowly gains the ownership they need to finally stop
paying tribute to another.

[Colin Hawkett]

On Thursday, 19 May 2011 15:21:11 UTC+10, Patrick Anderson wrote:

Colin Hawkett wrote:
> with a government that we trust and acts in our interests,

Governments cannot act in our (the Users') interest because they are
controlled by corporations.

That is the point - our end game should be to change this. As noted earlier, I think the key is to crowd-source government right up to the policy level - the target for corruption becomes extremely fuzzy.

Corporations cannot act in our interest because they are controlled by
investors that expect Profit as reward.

Treating Profit as reward causes conflict with Users because Profit
only occurs when Users lack control.

Profit is *undefined* when Users own and control the Sources of
Production because the Product is not sold - since it is already in
the hands of those who need it!

> the people are the shareholders in government,

I wish that were true, but Governments are composed of humans that
respond to the power consolidated into the hands of the owners of the
Corporations.

 Again, this is the point - I agree with you it is not true - this is the problem statement. This is what needs fixing.

> capitalism isn't actually broken

Capitalism is broken because we have mistaken Profit as a reward
instead of understanding it as a measure of the Payer's dependence
upon those current Owners.

When we finally realize what Profit really is, we will treat it as an
investment from the Consumer who paid it - causing a negative feedback
loop that will cause Profit to safely approach zero as each
User/Consumer slowly gains the ownership they need to finally stop
paying tribute to another.

I'm don't think you're describing capitalism there. Surely redefining profit as investment changes the system to something else? Can you point me to any widely accepted definition of capitalism in these terms?

[Mark Roest]

Hello Curtis,

I think we will really need to have sufficient capacity in each major industrial country, and in each, say, biome, to do what you say, "you can use IP multi-casting to make the network communications very efficient." Once we arrive at sustainable economies, most communication will be within the divisions I mention. It has been said that 85% to 90% of communication is local. In industrial / internet society today, that has been changed to interest groups and industry clusters. But a lot of it is chatter trying to understand larger patterns.

Once we use digital earth imaging and GIS effectively to truly understand the larger patterns that really mean something important, we will be able to focus more of our life energies on working with the patterns we discover, in the natural world that supports our existence. Then the real followings will be more regular and predictable (like Steve Gaskin's Monday Night Class, but on the Net instead of at the Avalon Ballroom).

If we build sufficient information feedback loops into the systems we build to do "you can use IP multi-casting to make the network communications very efficient" (or other designs that do roughly the same thing), we can see trends in demand and provision ahead of them, within major markets that do have lots of long-distance traffic, and within other regions that simply are big enough that we know we are going to have do plan to scale the systems as people come into the net and start solving their poverty issues. But for those places, I think we will see much more local focus, with communities and their representatives (and of course, teens looking for partners) doing a lot of the communication within regions or even between them, for purposes of securing key resources that are scarce in their eco-regions.But that is less likely to look like the surge when a popular tweeter tweets in today's consumer culture.

Regards,

Mark Roest

On Fri, May 13, 2011 at 3:26 PM, Curtis Faith <cur...@worldhouse.org> wrote:

Scalability is all about being able to decompose work into parallelizable chunks and minimizing the work that needs to be done on any given node.

General message queuing systems still have queue bottlenecks at the network, disk i/o and CPU level that can be easily exceeded with the wrong overall system architecture.

The key to scalability is to not have any bottlenecks that can't be subdivided and load-balanced on the fly without taking any nodes offline. What matters here is which nodes are talking to which other nodes, and what kind of information they are sending to each other, not the specific implementation of the nodes themselves. That's what I meant by optimizing the parts without changing the fundamental architecture.

Most of the architectural issues associated with the protocols are irrelevant to scalability. You could store data in in-memory lists for high performance, but if the amount of data coming into or out of a node exceeds the network or CPU performance capability, that node is effectively down. If this happens enough, fail-overs merely cascade to other previously under-capacity nodes taking the whole system down to a crawl as users everywhere halt.

So you need to have a way of making sure that this doesn't happen. For example, a mechanism so that nodes start offloading work and user connection responsibility to underutilized server nodes long before any given node reaches capacity. You can't have every peer talk to the busy nodes or when someone with 100,000 followers tweets you'll use up too many resources (individual TCP packets and the associated CPU time to process them). A good architecture will do what Charles describes, it will fan out and distribute the work from updates and change the connection topology based on the characteristics of the communication requests themselves.

Communication data should roll up geographically or topically and then fan out on the same basis. If you have the luxury of collocating a group of servers (for instance, if you are a centralized company, like say Twitter or FaceBook or an stock market exchange data provider) then you can use IP multi-casting to make the network communications very efficient. In a federation of peer servers located all over the world, you won't have this option.

On May 13, 2011, at 5:00 PM, Charles N Wyble wrote:

> On 05/13/2011 03:57 PM, Samuel Rose wrote:
>> On Fri, May 13, 2011 at 3:20 PM, Curtis Faith <cur...@worldhouse.org> wrote:
>>> Social network servers have a data distribution problem not a database problem. Many good lessons come from finance for building trading and exchange data feeds. People pay big money to make sure their data comes in reliably with extremely low latency. The scalability issues have been solved. You know how quickly the volumes blow up when all hell breaks loose in the markets with the high-frequency-trading algorithms running full tilt, right?
>>>
>>> You need to use some of the same kinds of tricks in a distributed system to get scalability, but that's not possible to do in a 100% pure peer-to-peer architecture.
>>>
>> I believe that diaspora is employing
>> http://www.amqp.org/confluence/display/AMQP/Advanced+Message+Queuing+Protocol
>> for addressing this. My understanding is that AMQP is capable of
>> meeting the challenge.
>
> Queuing systems fall over under high load. Multicast/data fountains are
> what a lot of people are using now. Moves the complexity into the
> network instead of the servers. Push the ugly around and all that. I
> like the multicast approach and I've got a lot of experience with it.
> Including building a multicast to unicast converter and back again.
>
>
>

[Mark Roest]

Hello Jon,

If you do the necessary minimum to harness the skills evident in this thread, and go for it, you can get compensated later, by creating a company like Red Hat -- you do not have to wait for a grant from someone who validates you from your words, ahead of time. Even though it feels good, it's unnecessary, if it is not readily available. If it is available, just going for it, but with the requisite process for the coding to go well, will get you there sooner. Go for it, as a collective!

Regards,

Mark Roest  (no programming skills yet, just whole systems perspectives)

On Sat, May 14, 2011 at 7:06 AM, Jon Lebkowsky <jon.le...@gmail.com> wrote:

On Sat, May 14, 2011 at 8:46 AM, Venessa Miemis <veness...@gmail.com> wrote:

"thinking and talking about what would be best is a fun exercise, but not always practical when it comes to actual human behavior. i think we should use ourselves as the test case and begin to experiment with what works best for us, and then build the tools to support those existing behaviors."

Is this something that people just go off and do? Or are we thinking to organize in some way?  It seems to me that we would benefit from a plan, a set of parameters, a set of models. Maybe some categories of activity, e.g. network infrastructure, network management, use cases, application structure, core technologies, etc.

I think there's been an implicit promise that a project or projects would emerge that would be more focused - beyond loose discussion, data gathering, and curation. We seem to have a set of participants who could take next steps effectively if we can figure out how to organize. Perhaps we focus on finding funding sources so that there can be compensation for doing something in a more organized way.

~ Jon

--
Jon Lebkowsky (@jonl)
Internet Expert and Strategist

Honcho, Polycot Associates, LLC

Work |  Twitter | LinkedIn | Facebook | Wikipedia | Blog | EFF-Austin

[Mark Roest]

Hello Miles and Richard,

Here is a slightly different twist: rather than "relies totally on ad hoc mesh networking for connectivity", which does have issues of making it work all the time right off the bat, how about working using WiFi levels of power wherever and whenever it works (in industrialized areas with high penetration), and switching to the existing cellular network when it doesn't, as in certain long-distance calls, until the new system reaches critical mass?

So suppose you go for it, but include both technologies -- the close-to-ideal, and the locally dominant one -- in the phones, and let them choose on the fly.

This is potentially a killer ap for one reason: the high prices charged and unfair contracts imposed by the dominant telcos. It enables people to get by fine most of the time on a minimum rate plan, and people will tell you that they want to get out from under their $300 a month phone bills! In fact you could support that strategy explicitly by showing the user how much time has been used on cellular since the last time they paid the bill and hit the counter reset button on the display. (A little trickier than that -- need to coordinate with when the telco starts a new count.)

Besides being a killer ap commercially, this fosters camaraderie and collaboration -- a feeling of membership in a new paradigm, even before the real paradigm shift is complete.

In non-industrialized parts of the world, the wi-fi power level will simply provide the fastest route to being able to communicate at all, for lots of people! If their phones also operate as cellular, they will be able to operate wherever they travel, especially in the early years.

Please consider this compromise, and provide a 'multi-carrier' design as you change the world! It will change it faster!

Regards,

Mark

On Sat, May 14, 2011 at 8:27 AM, Miles Fidelman <mfid...@meetinghouse.net> wrote:

richard adler wrote:

   Personally, my experiences lead me to the conclusion that
   self-organization and self-government have very little to do with
   technology and lots to do with group process and social
   infrastructure - ranging from rules of order to contracts to
   accounting rules to legal and cultural context.


There is truth in this. But a crucial part of these discussions is the example provided by the recent popular uprisings in the Middle East. We saw tremendous ingenuity displayed by activists there, but those innovations also faced serious obstacles because those group processes and social infrastructures were occurring on top of physical infrastructures they did not own and control.

To be sure, people can find ways to route around that lack of control, as indeed they did at critical moments earlier this year. But at least one of the goals of these discussions could be--and should be--models for internet infrastructures that would move that control from corporate or governmental entities and into the hands of the people who use them.

Agreed, though it keeps striking me as a set of issues that border on intractable:

1.  Shifting to user-owned infrastructure, en masse, requires that there be really serious incentives for everyone involved.  To date, it's been much easier for people to buy phones, computers, network services from large organizations that make everything easy and comparatively cheap.  Yes, there are ham radio operators floating around, and people who run independent wireless networks and such - but most people just don't want to be bothered until it's too late.  Military units, first responders, militias, terrorist groups will acquire technology in anticipation of action; most people can't even be bothered to have a first aid kit and some emergency supplies at home.

2. To the extent that technology is pre-positioned, widely available, and widely used - it becomes a serious target.  What's worked to date is when people find a new way to take advantage of technology that has become critical to businesses (fax during the Tiananmen Square uprisings, twitter more recently); use it until blocked; find work-arounds in the moment (e.g., Satphones smuggled over the boarder).  It's always going to be an arms race, and it may well be better not to present an easy-to-anticipate target.

Having said that, there are quite a few technologies already available that are hard to shut down without shutting down things the powers-that-be care about:
- onion routers provide for circumenting firewalls
- steganography seems to be widely used by terrorist groups (or so many would have us believe)
- botnets seem to be pretty hard to shut down (or even track down)
- satphones can be smuggled pretty easily and they can be had from companies and countries that are on different sides of conflicts
- there's still an awful lot of ham, CB, and emergency radio gear floating around
- the military spends billions on spread-spectrum, software-defined radios and ad-hoc mesh networking - with most of the core technology published in publicly-available journals

Would I like to have a data-capable smart phone / hot spot; that relies totally on ad hoc mesh networking for connectivity?  Of course.  Getting enough other people to adopt them, for them to actually be usable, is a much harder problem.  It's hard enough for a wireless carrier to roll out a new generation of technology - after investing billions in new phones, towers, radios, roaming agreements, and marketing to make the new phones useful on day one.

Social infrastructures are a critical topic for discussion, and I welcome it as part of Contact Con. But the technology underlying those social arrangements IS relevant and of central concern here.

Sure... but I expect that useful results are more likely to come out of MILCON and DEFCON.

[Mark Roest]

Very good! "The emperor has no clothes!"

Do aikido on the old one -- see earlier post tonight about the killer ap -- go local wi-fi with mesh when you can, to cut your cell-phone bill. That, because it will sell enough 'open hardware' phones to pay to manufacture in staggering volumes and allow the (generic) manufacturer to sell at $10 per phone (or $20 or more --  but to the end user or retailer -- not to a telco who marks it up to $300 and locks you into a 2 year contract that takes more than that out of you in excess fees) -- IN THE DEVELOPING WORLD! THAT CHANGES THE GAME HERE, TOO!

If we empower the developing world's people on the ground, in rural villages this way, we will enable them to rebuild rural economies that have been devastated by the traders who buy food from farmers for about the same money it took them to grow the crop in the first place. And charge more there than in the cities for city goods. Also by US crop subsidies for exporters.

If rural economies are rebuilt, many of the people who used to live there and participated in the largest mass migration in human history can return home.

That takes pressure off the cities in their nations. That takes some of the pressure off the industrialized nations to provide jobs for people from the developing ones. It solves the issues around immigration and competition for jobs in the industrial countries. It even enables people in cities in developing nations to organize in labor unions and fight effectively for decent wages, which means that US and European corporations can no longer get such a bargain for labor abroad, which means that their nations' citizens have more power to restore their own wages and benefits.

It also means that they can become stronger politically, and reverse what Ronald Reagan started -- wrest control of the economy and compensation back from the corporations and dominant individuals. And fund education and include the history (see Howard Zinn's A People's History of the United States, and Project Censored, and KPFA.org) so we never let it happen again!

Regards,

Mark Roest

[Mark Roest]

Hello Colin and all

I think that last is the point -- we have a system dominated by those who control capital now, hence it's name. If we use collective intelligence and collaboration to change that fact, the system will have fundamentally changed; it will no longer be truly a system dominated by those with capital, and therefore it will be worthy of a new name.

If we build a distributed-decentralized-internet like this, we can also use it for managing the political and economic decision process from the grass roots up.

This has been a truly remarkable discussion! We now have a set of reasons for action, which can help it to maintain its overall purpose and direction over time. Technologists, serve and empower the process, as true Stewards!

Regards,

Mark Roest

[Curtis Faith]

Mark,

Interesting idea. I like it. It recalls an idea I got while living amidst a lot of closely spaced high rises in Las Cañitas, Buenos Aires, Argentina a few years back.

In that neighborhood, and I suspect many others, there are perhaps 20 to 30 wifi networks within reach at any given time, I thought it would be really cool to develop an open-source hardware wifi DSL/Cable modem router that could act like a normal one, or that could mesh together to bypass the frequently broken and unreliable internet service where necessary. Or for secret untraceable communications. Or for bypassing the local cell phone networks to make mobile phone calls that were able to link into the mesh network somehow.

If you added some computing power to the box using very low power CPU chips like ARM, it could become a server node in the Diaspora, Twitter, Social Networking collaboration and run a Linux variant.

The key would be to offer benefits for the buyer who only wanted to use it as a wifi router. That would drive ubiquity and make it much more likely that the mesh would exist for any given router.

An open-source hardware and software wifi router that was also a node not only in the mesh network but in the server network for replacing Twitter and FaceBook and other goodies would be very useful.

We could even distribute out the data using encryption techniques and automatic replication to nearby nodes (as well as ones in key other countries) so that backups were unnecessary and so that local police or governments wouldn't be able to get enough information even if they happened to confiscate all the nodes in an entire city unless they happened to know and have power over the nodes in the other countries.

I'd imagine some countries would become the Switzerland's of privacy. I had heard that Costa Rica was heading in this direction a few years back.

Having a known set of smarter routers would also allow one to offer better performance as a group. Frequencies could be dynamically negotiated to minimize interference between routers. Right now, there's got to be a lot of packet loss because of noise from other routers in most cities.

Peace

Curtis

[Jon Lebkowsky]

Admittedly I could only skim, there was so much verbiage here since I last looked. However I have some points that I think are relevant based on what I see here...

Re a corporation's responsibility to shareholders, I think the legal obligation is contractual, and a corporation can structure an agreement that tells shareholders, for instance, that sustainability will have priority, or that some percentage of profits will be committed to environmental stewardship, that sort of thing. I.e. (noting that I'm not a lawyer) I think there's no law that says a corporation has to put profit first, just a decree from Milton Friedman... but it could be seen as a contractual issue and I assume shareholders could create legal grief if a corporation wasn't behaving according to their expectations. I think that's the basis for some class action suits.

Re all this talk about "government": You don't get anywhere by conceiving of government as a black box or monolithic entity that can be described as any one thing. What we call "government" is an incredibly complex ecosystem of legislatures, courts, officials.agencies, laws, regulations, etc. The components don't act in concert or with singularity of purpose. Any attempt to address government as a "thing" and not as a complex system is a fantasy and won't get anywhere. I did see something recently from someone who was disillusioned after starting to work in a legislative office, and finding that all the work there is driven by greedy corporations. It's either big greedy corporations wrangling with other big greedy corporations, or with small greedy corporations, and sometimes small greedy corporations wrangling with other small greedy corporations. There's no time or energy left for the kinds of civic issues we would hope they'd be dealing with.  I have said for many years that citizens' best bet is to swarm legislators - show up in real numbers with coherent proposals and arguments, but numbers and facts won't necessarily trump money in that world. Where I've seen lobbying work, it was because corporations friendly to our side of the issue stepped in (this in the muni wireless battle in Texas, where Intel, Dell, and Texas Instruments were instrumental in making the bad legislation go away).

My advice to this list is that we should avoid talking about "government" and focus on actions we can take that don't require that conversation, at least for now.

Also, to the point about going to third world countries and empowering people on the ground, just know that the politics there may be more difficult than the politics here, and more brutal.

~ Jon

--
Jon Lebkowsky (@jonl)
Polycot Associates

[Patrick Anderson]

Colin Hawkett wrote:

> Patrick Anderson wrote:
>> Governments cannot act in our (the Users') interest
>> because they are controlled by corporations.
>
> That is the point - our end game should be to change this.

I'm just saying, if we change the corporations (one corporation
at a time - by creating new corps that do the right thing), then
we don't need to separate the goals of gov from corp.

> I think the key is to crowd-source government

If we have crowd-owned corps, we will have crowd-controlled gov.

>> I wish that were true, but Governments are composed of humans
>> that respond to the power consolidated into the hands of the
>> owners of the Corporations.
>
>  Again, this is the point - I agree with you it is not true - this is the
> problem statement. This is what needs fixing.

We can change the corporation's goal of keeping Price above Cost by
attracting Consumers to Invest for the purpose of receiving at-cost
Product. At that point, we will be free to redirect the special value
called Profit (which is defined as Price above Cost) to be treated as
though the Payer of that special value had just made an investment
and is the real owner of that growth. This auto-distributes control
into the hands of those willing to pay for growth and removes the
unnatural and dangerous drive toward scarcity and destruction that
treating Profit as reward tends to cause.

>> Capitalism is broken because we have mistaken Profit as a reward
>> instead of understanding it as a measure of the Payer's dependence
>> upon those current Owners.
>

> I'm don't think you're describing capitalism there. Surely redefining profit
> as investment changes the system to something else? Can you point me
> to any widely accepted definition of capitalism in these terms?

I'm not describing the currently operating Capitalism that is raping our
planet, I am describing why Capitalism is broken, and how it must change
if we are to continue to exist as a species.

Yes, redefining Profit as Payer Investment changes the system dynamics,
and that is a good thing because the current system is extremely wrong.

[Patrick Anderson]

Colin Hawkett wrote:
> with a government that we trust and acts in our interests,

Colin, why does government currently not act in our interest?

[Colin Hawkett]

On Thursday, 19 May 2011 23:58:35 UTC+10, Patrick Anderson wrote:

Colin Hawkett wrote:
> Patrick Anderson wrote:
>> Governments cannot act in our (the Users') interest
>> because they are controlled by corporations.
>
> That is the point - our end game should be to change this.

I'm just saying, if we change the corporations (one corporation
at a time - by creating new corps that do the right thing), then
we don't need to separate the goals of gov from corp.

Ok - my first reaction to that approach is that it would be an incredibly hard route - so hard that I struggle to see how it would be possible. The systems of government, many with constitutional rights, and 'democratic' process already exist for the role of civic interest. It seems a little strange not to consider leveraging what we do have to fix the system designed to represent out interest.

I do see your high level logic though - corporations control government, so lets control the corporations... strikes me like a dog chasing its tail - I'll give it some more thought.

I think there is confusion here between your philosophical position on capitalism from an ethical perspective, and the mechanics of the system of capitalism. It seems to me you are just saying that you disagree with capitalism, and prefer a different system. That's cool - I'm not arguing against that sentiment at all. I'm just saying its not broken - in the same way that a gun is not broken when it kills people with the safety off. Its a system operating according to specification.

If the consensus is that capitalism sucks, then we should choose something else, by all means. That would be a lot easier to do if we trusted our government and it acted in our interest.

[Miles Fidelman]

Mark Roest wrote:
>
> Here is a slightly different twist: rather than "relies totally on ad
> hoc mesh networking for connectivity", which does have issues of
> making it work all the time right off the bat, how about working using
> WiFi levels of power wherever and whenever it works (in industrialized
> areas with high penetration), and switching to the existing cellular
> network when it doesn't, as in certain long-distance calls, until the
> new system reaches critical mass?

We're actually pretty close to that now:

You're essentially describing the capability built into the One Laptop
per Child computer - they establish IP over WiFi mesh connections with
each other, and through each other, until a connection reaches an
Internet POP.

A lot of cell phones support WiFi as well as 3G, and the IP stack will
select WiFi if you're in range of an access point (sort of - there's a
lot of proprietary stuff in those stacks, and carriers like to bill).

And then there are cellular microcells - when you walk into your house,
your phone ends up connecting to the local microcell, which in turn is
connected to a broadband wired network.

It pretty much comes down to configuration at the routing level - what
routes your local node publishes, and peering configuration between
various networks. The original Internet was essentially a mesh
network. Private peering arrangements are what route things through
specific large broadband networks at the edge, and large backbone
networks at the core. (Note that it's a little more complex than that -
full mesh routing works well in a small network, but in a network the
size of today's Internet, routing tables and calculations become
intractable for a fully peer-to-peer mesh.)

It also comes down to a systems administration issue - configuring and
tuning routing systems is not an easy task, and the state of the art
does not come close to this being a completely automated process
(witness the various Internet outages when a corrupted routing table
gets propagated).

There's an awful lot of research and development needed (and going on,
mostly with military funding) to get to a point where this kind of
capability can be deployed on a large scale, to non-technical users.

Just one man's opinion of course, but 40 years in the industry, dating
back to year 2 of the ARPANET (1971), and a long stint at BBN working on
things like network management for large pieces of the early Internet,
plus a lot of years working on municipal networks (among other things)
provides some amount of perspective.

[Miles Fidelman]

You're kidding, right?

[Colin Hawkett]

On Thursday, 19 May 2011 23:11:42 UTC+10, Jon Lebkowsky wrote:

Admittedly I could only skim, there was so much verbiage here since I last looked. However I have some points that I think are relevant based on what I see here...

Re a corporation's responsibility to shareholders, I think the legal obligation is contractual, and a corporation can structure an agreement that tells shareholders, for instance, that sustainability will have priority, or that some percentage of profits will be committed to environmental stewardship, that sort of thing. I.e. (noting that I'm not a lawyer) I think there's no law that says a corporation has to put profit first, just a decree from Milton Friedman... but it could be seen as a contractual issue and I assume shareholders could create legal grief if a corporation wasn't behaving according to their expectations. I think that's the basis for some class action suits.

Re all this talk about "government": You don't get anywhere by conceiving of government as a black box or monolithic entity that can be described as any one thing. What we call "government" is an incredibly complex ecosystem of legislatures, courts, officials.agencies, laws, regulations, etc. The components don't act in concert or with singularity of purpose. Any attempt to address government as a "thing" and not as a complex system is a fantasy and won't get anywhere. I did see something recently from someone who was disillusioned after starting to work in a legislative office, and finding that all the work there is driven by greedy corporations. It's either big greedy corporations wrangling with other big greedy corporations, or with small greedy corporations, and sometimes small greedy corporations wrangling with other small greedy corporations. There's no time or energy left for the kinds of civic issues we would hope they'd be dealing with.  I have said for many years that citizens' best bet is to swarm legislators - show up in real numbers with coherent proposals and arguments, but numbers and facts won't necessarily trump money in that world. Where I've seen lobbying work, it was because corporations friendly to our side of the issue stepped in (this in the muni wireless battle in Texas, where Intel, Dell, and Texas Instruments were instrumental in making the bad legislation go away).

I doubt that anyone here, whether they were interested in the governance and policy angle or not, perceives government as anything but the complex system you describe - I took it as a given - how could they not, really? Nevertheless this complex system is tasked with producing policy and legislation, and with the mechanics of enforcing that legislation. These are the constraints under which society runs, under which corporations run. As you say this complex system is riddled with corporate interest rather than civic interest, which was exactly the point I made here. To say 'its a complex system, forget about it' seems... odd. It is what it is - we are participants in it. Whatever path we choose, the ultimate resolution of its problems is surely the goal? All paths are intended to lead away from this mess and toward a solution where the common interest is represented, yes?

More or less agree with the approach of swarming legislators with high quality stuff, but I think we now have within our grasp the technology to actually realise the complex collaboration that vision would require. We seem to be saying something pretty similar at this level. I agree that numbers and facts won't necessarily trump money, but the democratic process has the ability to. We do actually have the means to get control of government - we just don't have the collective will to use it (for lots of reasons, and for another post). The means are there, for now. Perhaps you disagree with this?

Another angle, perhaps that is more helpful - assume we find a path that makes existing systems irrelevant. How do we govern in their absence? Surely even in this scenario we must deliver a system for government in the common interest?

 

My advice to this list is that we should avoid talking about "government" and focus on actions we can take that don't require that conversation, at least for now.

Like what?

[Patrick Anderson]

> Patrick Anderson wrote:
>> Colin, why does government currently not act in our interest?

Miles Fidelman wrote:
>
> You're kidding, right?

My hope was to have Colin answer "Because corporations control governments".

Then the question becomes "Why do corporations not act in our interest?"

The answer to that is "Because shareholders (currently) require Profit
as payment."

Then the question becomes "What if we could pay shareholders with
Product instead?"

The answer to that is "We can only treat Product as reward if
shareholders are Users, and hold exactly as much ownership as they
intend to use of the Product."

Then the question becomes "How do we insure ownership is distributed
to the new Users who come on board as the org grows?"

The answer to that is "We can treat the Profit those new Users pay as
though it were an investment from those Payers."

[Charles N Wyble]

On 05/19/2011 09:30 AM, Miles Fidelman wrote:

> Mark Roest wrote:
>>
> We're actually pretty close to that now:
>
> You're essentially describing the capability built into the One Laptop
> per Child computer - they establish IP over WiFi mesh connections with
> each other, and through each other, until a connection reaches an
> Internet POP.

Actually when I tested this a year or so ago it was broken. The mesh
didn't actually route traffic. :( I hope it's been fixed since.

Giant +1 to the previous paragraphs. The nextnet/Free Network Foundation
roadmap that I and others are working on, addresses a lot of these
issues. I expect to publish it by the end of May. I think that will
really bring a lot of this discussion, thinking etc to a common point.

--
Charles N Wyble cha...@knownelement.com @charlesnw on twitter
http://blog.knownelement.com
Building alternative,global scale,secure,
cost effective bit moving platform
for tomorrows alternate default free zone.

[Miles Fidelman]

Patrick Anderson wrote:
>> Patrick Anderson wrote:
>>
>>> Colin, why does government currently not act in our interest?
>>>
> Miles Fidelman wrote:
>
>> You're kidding, right?
>>
> My hope was to have Colin answer "Because corporations control governments".
>
>

Well yeah :-)

> Then the question becomes "Why do corporations not act in our interest?"
>
> The answer to that is "Because shareholders (currently) require Profit
> as payment."
>
> Then the question becomes "What if we could pay shareholders with
> Product instead?"
>
> The answer to that is "We can only treat Product as reward if
> shareholders are Users, and hold exactly as much ownership as they
> intend to use of the Product."
>
> Then the question becomes "How do we insure ownership is distributed
> to the new Users who come on board as the org grows?"
>
> The answer to that is "We can treat the Profit those new Users pay as
> though it were an investment from those Payers."
>

That doesn't really help a lot when faced with entrenched monopolies,
backed by government regulations and enforcement that help them stay
entrenched.

[Patrick Anderson]

Miles Fidelman wrote:
> That doesn't really help a lot when faced with entrenched monopolies, backed
> by government regulations and enforcement that help them stay entrenched.

Are you saying it is impossible to start new businesses that:

1.) Are funded and Owned by the Users.

2.) Treat Product as reward for those investments.

3.) Treat Profit against latecomers as Payer investment.

Are you saying the governments won't allow such an
organizational form?

Maybe you are right, and we should prepare for that
resistance. I wonder what machinations they will try...

[Miles Fidelman]

Not completely - we certainly have enough cooperatives and municipal
utilities to point to, as well as various forms of joint ventures,
partnerships, and so forth - all of which are examples of user-owned
businesses.

What I am saying is that:

- entrenched monopolies control a lot of the resources needed to succeed
in a telecom. business (e.g., access to rights of way, ownership of
wireless spectrum, etc.)

- there are plenty of cases of entrenched monopolies throwing up
roadblocks to things like municipal utilities (one of the more effective
forms for a community to own infrastructure) -- there are more than a
few states where it is illegal for a local government to go into the
telecom. business; and additional states where municipalities have to
"compete on an even footing" with the private sector (i.e., even if the
city already owns a wireless tower, they have to include its cost in
prices charged)

- there are lots of other regulatory games that entrenched monopolies
can play - for example: want to put wire or an antenna on a power pole,
then you need to both negotiate access, and you need people with the
right certifications to do the work (actually, not a bad idea - working
near live electric wires is pretty dangerous)

- competing with an entrenched monopoly is hard and costly, and they can
afford to operate at a loss to drive you out of business --- it's pretty
hard to generate a critical mass of users when the entrenched
competition is giving away service

- and then, if you're relying on any portion of a carrier's
infrastructure, they are going to do things like block your packets,
impose contractual restrictions (it's a violation of your terms of
service to share your WiFi with your neighbor), and so forth

- by all rights, government should be imposing anti-trust restrictions
as a counterbalance to these kinds of things, but I haven't seen a lot
of anti-trust enforcement since the telco breakup a few decades ago

In short, building and operating infrastructure is a complicated and
expensive proposition that involves competing with 800 point gorillas.
It's doable if you want to play hardball, but it's not a game for
theoreticians.

[Patrick Anderson]

Miles Fidelman wrote:
> we certainly have enough cooperatives and municipal
> utilities to point to, as well as various forms of
> joint ventures, partnerships, and so forth - all of
> which are examples of user-owned businesses.

Those organizational forms do not comply with:

2.) Treat Product as reward for those investments.

3.) Treat Profit against latecomers as Payer investment.

> it is illegal for a local government

> to go into the telecom business;

Yet another reason to incorporate in a GNU way instead
of handing it over to a poorly structured city government
that will overcharge for access and never deliver what we
really need since it will once again be 'us' against
'them' - for we, as citizens, do not have real ownership
of the cities even though we pay taxes into slush-funds
that are then doled out by well intentioned tyrants of
the majority.

> competing with an entrenched monopoly is hard and costly,
> and they can afford to operate at a loss to drive you out
of business --- it's pretty hard to generate a critical mass
> of users when the entrenched competition is giving away service

I agree this will be a problem, but I think we can address this
issue by beginning small and growing slowly.

Of course no business cannot operate at a loss indefinitely.

We will have another advantage in that we will be able to operate
at "at cost" indefinitely - since we will not be trying to
perpetuate Profit, but will instead be paying investors with the
Product itself - and so will only need to collect the Costs of
operation instead of continually trying to charge ourselves
more than it really Costs to sustain that network.

> by all rights, government should be imposing anti-trust
> restrictions as a counterbalance to these kinds of things,

So you want a government corrupted by corporations to turn
against those corporations?

How do you propose we do that?

I doubt you have enough money to purchase such legislation,
and even if you could, it would only be a brief win in a
small skirmish that would soon be washed away by all the
other corporate pressure to work against we, the Users.

[Curtis Faith]

I have a lot of concrete specific ways and associated elaborate but simple plans we can use to fix the plutocracy and corporate greed/power problems, and divert resources from them to good projects.

My background in trading and finance is very strong and I've been thinking about this for years.

The key is to get enough people to synchronize their small actions.

That's why a collaboration environment is the first step. The rest of the problems are actually much easier. Corporations respond to well-defined external stimuli. We can create the ones we need to make the changes we want happen in the most effective humane way possible if we collaborate together.

Peace

Curtis

[Miles Fidelman]

Curtis Faith wrote:
> I have a lot of concrete specific ways and associated elaborate but simple plans we can use to fix the plutocracy and corporate greed/power problems, and divert resources from them to good projects.
>
> My background in trading and finance is very strong and I've been thinking about this for years.
>
> The key is to get enough people to synchronize their small actions.
>
> That's why a collaboration environment is the first step. The rest of the problems are actually much easier. Corporations respond to well-defined external stimuli. We can create the ones we need to make the changes we want happen in the most effective humane way possible if we collaborate together.
>

Perhaps a silly question: What's wrong with existing collaboration
environments?

There's an awful lot of collaboration being done via everything from
email to usenet Drupal to special-purpose sites like Ushahidi. And
there are plenty of examples of ad hoc collaboration in response to
special circumstances (e.g., some of the family matching sites that
cropped up after New Orleans). There are the larger, more formal
efforts such as those around various open source software projects.
There are the flash mob performance art groups like the Banditos
Mysteriosos. And of course there are examples of boycotts organized
across the network (remember "New Coke?").

Not that I'm arguing against new, better, broader technology and
infrastructure; just that there are a lot of tools already available.

Which brings us back to the question: What's missing for your purposes?

[Suresh Fernando]

Miles wrote...

Perhaps a silly question:  What's wrong with existing collaboration environments?

Not a silly question at all... My thoughts on this are the following:

Yes, there are lots of collaboration 'environments' with a lot of people doing very cool things together. There are also a myriad of platforms that can be used. What has not yet happened, however, and what seems to be possible is true mass collaboration towards a socially beneficial end...OK, this is also not entirely true since we have Linux and Wikipedia... To be more precise there is no mass collaboration environment that can scale where people can, for example, come together to build muni broadband networks across the US, fight climate change... whatever...

The point is that for the first time in the course of human history it is actually conceivable that 1,000,000 (100,000,000?) could coordinate their activity...

We are faced with unprecedented challenges, thankfully, we have unprecedented possibilities for coming together...

An environment that would make this possible would have, at the least, the following characteristics:

• It would be Open... permeable boundaries
• It would be Scalable... it could accommodate lots and lots of people getting involved in the 'same project'... where of course there would be many connected sub-projects
• It would be Decentralized... leadership would be distributed across the 'network' of nodes in the project
• It would rely on the principle of Emergence to stimulate and catalyze projects... the open infrastructure would allow anyone to introduce a new project idea (possibly as a sub-project)
• It would rely on the principle of Visibility... all activity in the system would be visible to others in the system. This would be very analagous to the way that Facebook makes the network of social relations widely visible thereby stimulating further social relations... why not make collaboration relations highly visible?
• It would be Intelligent... it would include recommendations algorithms's similar to what online dating services use to connect people that are working on related projects... 

The online dating model and the Facebook model are proven. Why not closely examine those models and import their features to create an open and highly scalable environment that connects people who want to collaborate?


Note that all that I have described above could be sufficient to bring people together on a massive scale. It does not, however, solve the problems that I have already identified and that we are discussing:

The Founding Governance Structure: in the absence of any predetermined set of rules and no-one empowered to make the rules, how do you get going?

Leading and Following: the question as to when to join someone else's project instead of feeling the need to only drive your own. At what point do you realize that someone else's project is actually also your own?


To Sam and Paul's points... since I am talking about an open infrastructure that can massively coordinate activity and this infrastructure will be net based, there is a sense in which this activity is swarm like...

I am also talking about an infrastructure to support a revolution... a cultural one... this, of course, is swarm like ;-)



Suresh

--
Suresh Fernando
Partner
Vicinus Group
http://www.vicinusgroup.com
778-709-7526

[Colin Hawkett]

On Friday, 20 May 2011 05:11:40 UTC+10, Curtis Faith wrote:

I have a lot of concrete specific ways and associated elaborate but simple plans we can use to fix the plutocracy and corporate greed/power problems, and divert resources from them to good projects.

My background in trading and finance is very strong and I've been thinking about this for years.

The key is to get enough people to synchronize their small actions.

That's why a collaboration environment is the first step. The rest of the problems are actually much easier. Corporations respond to well-defined external stimuli. We can create the ones we need to make the changes we want happen in the most effective humane way possible if we collaborate together.

Bingo! :) 

Peace

Curtis