[Freedombox-discuss] What Do You want to use the FreedomBox for?

[Jay Sulzberger]

On Sat, 26 May 2012, Jonathan Wilkes <jancsika at yahoo.com> wrote:

> ----- Original Message -----

> From: Jay Sulzberger <jays at panix.com>
> To: freedombox-discuss at lists.alioth.debian.org
> Cc: 
> Sent: Friday, May 25, 2012 10:50 PM
> Subject: Re: [Freedombox-discuss] What Do You want to use the FreedomBox for?
> 
> 
> 
> On Fri, 25 May 2012, Joshua Spodek <joshuaspodek at yahoo.com> wrote:
>
>>  Thanks for asking. I find it interesting to see everyone's different
>>  goals. I think my needs overlap with typical non-geek users wanting to
>>  avoid faceless corporations owning my private data.
>>
>>  * Diaspora
>>  * Skype replacement
>>  * Host my own email, blog, photographs
>>  * Seamless backups to friends' Freedomboxes
> 
> Suppose we have two people in front of home computers connected
> to the Net via Time Warner Cable in Manhattan.  We assume further
> that the two people are in their own houses, and that they own
> their computers, ordinary sense of "own a computer".
> 
> If the two people are Unix sysadmins, then they can arrange to
> easily move files from one box to the other.  Say they have set
> up a system so that with the push of one button, and the
> indication of a file on their computer, the file gets sent to the
> other computer.  Such an arrangement would serve, I claim, as a
> foundation for what we want.
> 
> I think today the main obstacle for non-Unix-sysadmins to running
> such a file transfer utility is setting up the "home router",
> that is, the router behind the Time Warner "cable modem".  If
> there were no router in the way, then it is not hard to set up a
> system which could be used by two non-Unix-Sysadmins.  (Not hard
> as long as we have some method for getting the Grand Net facing
> address of one box to the other.  And if we allow dependence on a
> third party then whatismyip.com serves; if we wish to avoid third
> party dependence at this level, likely we will have to set
> something up on the router; and there are other methods too.)
> 
> The home router is today usually:
> 
> 1. a box separate from the home computer
> 
> 2. with a difficult to understand method of programming, that is,
> the ridiculous "web based" fill in the incomprehensible form,
> 
> 3. which form is non-standard
> 
> These three things are, I think, mainly what makes direct
> connection over the Net so hard for most people.  Thus we must
> repair these deficiencies:
> 
> 1. whether the box is grossly physically separate from the home
> computer, its setting up to allow direct comunication with the
> other box cannot require more than putting the name of the owner
> of the other box; likely we should have the router be contained
> in a joint "home computer router" thingie

Just to make sure we're both talking about the same problem-- the 
main problem in #1 is that for Bob's Freedombox to talk to Alice's 
Freedombox, Bob must traverse the stock wifi router/dsl modem by 
poking a hole using port forwarding or some other mechanism to 
allowing two-way communication between the Freedomboxes.  
(I'm assuming here that either Alice doesn't have a NAT traversal to 
worry about or has already magically dealt with it.)

There's no workable "one-click" way to do this as I see it-- some 
routers are open wifi routers, some are not, some are password 
protected-- and of those that are password protected, some have 
a default hardware password, some have a default provider password, 
some have a custom one hidden from the person paying for the account,
and some have a backdoor to allow the 
network owner to push "updates" to the router.  Furthermore some
ISPs allow custom changes through the router's web-interface, some 
have a TOS that disallow but are practically lazy about it, some 
will throttle you, some will "repair" the problem, and probably all would 
become aggressive if there were a large increase in home users
setting up their router for an always-on, multi-service, internationally 
reachable server.

I hope I'm wrong about the difficulty of a one-click solution, but if you 
look at the burgeoning privacy-aware network overlays out there right 
now and realize that those protocols would grow exponentially if any 
of them were to implement a one-click solution, it becomes obvious that 
this isn't a problem that the Freedombox can fix.  (In fact throwing 
hardware at the problem would make it more difficult, as a cross-platform 
software solution would be much cheaper.)

The approaches I can think of are:
a) pagekite et al, which would then become a central point of attack/failure
b) supernodes, which run the danger of de facto centralization (like Diaspora's 
main pod) because there hasn't been enough privacy education for the 
common user to be able to gauge the difference in risk level between entrusting 
data to a close geeky friend vs. a stranger with a less-than-evil TOS.

-Jonathan

Thanks, Jonathan, for reading and responding.

I will, as soon as I have one full day free, attempt an answer.

In my first post in response, I will argue that the old PSTN
provides a partial "existence example".  It exists today (I do
not speculate how long it is likely to continue to exist.): I can
call a person's number, which number was given to me by the
person, in person, and expect the connection to be made.

I will discuss the mechanisms, political/economic, electrical,
and electronic, by which the old PSTN accomplishes this ease of
connection.  I will suggest various lines of attack:

1. replacements built by us, similar in design for some of the old PSTN mechanisms

2. new mechanisms built by us for other old PSTN mechanisms


I will also argue against this statement:

   "it becomes obvious that this isn't a problem that the Freedombox can fix"

We agree with your starting point: Today there are, at the level
of the home router, several different, partly incompetent and not
smoothly interoperable, modes of connection.  But this bad
situation is not ordained of Heaven.  The various routers and
protocols and practices stand in our way because they have been
built, advertised, sold, bought and installed.  And now there
they sit: radical electro-mechanical suppressors of the Net
itself.  So let us build new standard "home routers".  Of course,
we will advertise our new things and seek to get people to use
them.

Note that the central problem here is a problem of standards.  We
would have no TV without standards, no electricity in our houses,
no cars, no running water out of faucets, no toilets, indeed, no
computer hardware, without standards.

So let us standardize hardware and protocols for Proper Net
Connection.

Once pictures and sound were hard to send across the Net.  Now we
have http and html and httpds and http/html browsers.  We got the
Web subsystem of the Net by invention, standardization, and
popularization.  We can do the same for Proper Net Connection.

oo--JS.


> 
> 2. no "web form" which asks such questions as "What is the IP
> address of your nameserver?" or requests "Enter fibroblast count
> E4 and Dunning-Kruger osteoclast rate, EUMED units (not ISO
> units!), for your six top friends."
> 
> 3. the button is standard, the same for every proto Freedom Box^W^W^Wstandard 
> box
>
>