Bitcoin Forum

In many areas of the world, there is not just limited connectivity, there is no Internet connectivity.   But does that mean those people would be shut out from using Bitcoin?

Is there a way that Bitcoin transaction can execute via sneakernet?   (i.e., the villager and his wallet stay home, but the transaction to pay was transferred to an agent which is then taken to a location that has Internet connectivity)?

Theoretically we could all memorize the chain and shout transactions to each other, the computers are just here for our convenience. So, yes, possible for sure, and it shouldn't be that hard.

I use bitcoin on a thumbdrive, but have not yet found a way to import or export a transaction via a file.  I'm sure that it's possible to do so, and will be done for email if nothing else, but as far as I know there is no way to sneakernet just a transaction.  But it is possible to sneakernet the entire Bitcoin client on a thumbdrive, which is exactly what I do because of a restrictive firewall.

Once a proper lightweight Android client is completed, and with a cell phone that has NFC/Dash7 networking, regular access to the Internet will become as optional as regular access to a normal computer.  Of course, that assumes that you have a smartphone.

Bitcoin needs a mesh network and a way for people to earn bitcoins for supplying bandwidth.

Dash7 on smartphone and bitcoin could pay for "priority mail stamps" for Jabber text message forwarding.  Without mesh forwarding, Dash7 networking can hit 2 clicks on a milliwatt transceiver, and would make for one awesome "Citizen's Band" version of p2p instant mobile messaging.  Dash7 could also handle broadcasting the transactions of smartphone based lightweight clients without the need for direct Internet access.  Anyone within 2 clicks of the sender would immediatly see the new transaction, presumedly including the receiver if it is a IRL transaction.  Sooner or later, someone will encounter an open hotspot, and the transaction will make it into the blockchain.

So the hotspot owner would earn the bitcoin? That would be a really viral way to grow the network and help provide a massive buffer against a network split.

No, the bitcoin stamp concept was related to the Dash7 Jabber client.  In which case, either the owner of the cellphone that provided the Dash7-to-Wifi gateway would earn some percentage of the stamp's bitcoins, or Dash7 intermediaries who were part of the Dash7 mesh required to reach (open hotspot|recepient's cellphone directly) would share the stamp in some fashion.   I don't know if it's possible to script such a division without knowing the addresses of the ad-hoc mesh in advance, but it would function much like a wireless mesh version of Unix-to-Unix-copy-protocol, but rapidly.  Text messaging via the current cell phone pay services would be faster and more reliable, but this would work even in places that wireless infrastructure doesn't exist for the phone itself as well as an emergency communications method for an entire city when the lights go out.  However, Jabber has presence information as part of the message protocol, and a Dash7 version would likely  require announcing your GPS location to the world in the clear, in order to establish a vector that a message would have to travel to reach it's intended recepient.  Fortunately, there is nothing that says that a Dash7 network ID has to have any connection to the user's real ID, and would be pseudo-anonymous in the same fashion that Bitcoin is, except that sniffers would know where the user was in real time.  Message bodies would likely be encrypted end-to-end, and Dash7 supports device-to-device encryption, so maybe there is still a way to establish a message vector without announcing your precise position in the clear, but your general position would be impossible to conceal.


As for the Bitcoin transactions themselves, there would be no need to establish a mesh, as the transaction would be broadcast in the clear to every Dash7 device that could hear it.  It would be in the self interest of business establishments that accept bitcoins via smartphones to have a passive Dash7|Internet gateway device somewhere on the property to forward any bitcoin transactions that it sees to the Bitcoin network, since it is highly likely that the vast majority are transactions that directly benefit the establishment anyway, and any free riders within range wouldn't justify the cost of filtering.  Even if they didn't, any bitcoin client that heard the transaction would forward it to it's peers upon the next opprotunity provided it's queue of stored transactions wasn't already full.

The only reason wifi has such a short range is because government regulations artificially limit transmission power to 50 mW. This is tiny. Even a mobile phone can go up to 2000 mW.

Technically it should be trivial to modify transmission power of a wifi device, and government regulations don't get enforced much anyhow in "offline" parts of the world like Somalia, if they exist at all.

Then it should be possible to construct a wireless mesh network with several tens of kilometers between each node, bringing even the most isolated villages online.

All you need is one person with a smartphone in each village acting as the local banker.

i think the very last thing that Somalia villager wants is bitcoin  :)

He probably does want to send money to a relative in a different village.

In many remote parts of Africa this is currently handled by SMS.

This system is better than nothing, but it still relies on capital-intensive mobile phone infrastructure, and the transaction costs are huge, in the region of 10-20%.

SMS transfer costs are oppressive, and I can certainly understand wanting to use Bitcoin for that reason alone, but Bitcoin also requires it's own infrastructure.  Eventually, however, some group similar to the OLPC movement could develop a cheap device capable of managing a lightweight Bitcoin client and a Jabber client using some future wireless standard, without the need for all the other stuff on a cell phone that adds costs.

Hmmm, I'm a bit of a hardware geek, I wonder how cheaply I could make such a device?

I'm also a ham, and a former military radio operator, and know that the reasons for the limitation on output power are not arbitrary.  In the case of WiFi in particular, the only reason that such devices are in that band is because it's unlicensed; it's unlicensed because it's the same band as microwave ovens, microwave ovens use it because it's the resonant frequency of water, which is why it can heat up food at all.  It is an extremely bad idea to uncap the output power of devices that use this band, as if any frequencies can harm human life, it's this one.  Also, Wifi is range limted because it's a particularly wide spectrum technology and because of it's frequency choice.  It was never designed to cover great distances, and the fact that it can be done at all is due to the talent of the geeks to do it and not the nature of the technology.

And even if Wifi was uncapped, it is unlikely that mobile users could turn up the watts anyway, as they are battery limited and just turning on the wifi radio eats the battery.


It's not trivial at all, but mostly because the functions of a wifi device are encoded into the hardware because that's much cheaper in volume than a software defined radio.  SDR devices that can act like a standard (or moded) wifi device within that band do exist, but they are too expensive to use as a base node for any kind of mesh network.


Anything is possible with enough money, but if that kind of money were available, tens of kilometers of multiwire, direct burial phone cable (buried or laid directly upon the wilderness dirt) to connect these villages, and a computer  in each village with a few standard modems and a wifi card would be much cheaper and easier to maintain in the long run.  Hell, even a null modem might work.  To this day, that is exactly what the USMC does to connect commanders of base camps using huge rolls of POTS cabling with quick connectors and extra tough armored jackets.  (they do have extra wires connected to special devices that can detect tampering, and even tell where in the cable a cut is.)  In fact, the best place to get the stuff to do what you are talking about is military surplus, since this kind of stuff tends to not sell very high as there is very little value in it outside of military or aid missions.


That might work, but you would need at least two, otherwise all you have done is move the monopoly power from some faraway company to the one guy in the village who owns the cellphone.

If i'm not mistaken, except for mining, you could network two machines and have them exchange transfers data, then pull one of them and connect it alone to the Internet, then connect it back to the other lone one and things would work just as if  they have all been connected to the Internet, except for the huge lag in transaction confirmations etc, and this would still work no matter how many steps you place between one machine and the Internet as long as it gets connected to the Internet no matter how indirectly in a way both things sent and to be received eventually reach their destination.



Mining can't be trusted in such a scattered network 'cause the Internet probably found the solution before one of the sneakernet nodes did and the coins from the sneaker node will be considered invalid and wiped.

Why not have a bank issue notes backed by bitcoins? Users will have to trust the bank, but at least you won't need electricity to conduct a bitcoin transaction.

So, if your network got isolated, and you didn't want to fork the block chain, you could just conduct transactions without confirming them?

Btw, regarding WiFi ranges; google "super wifi"

I've actually been tossing that idea around in my head too. In fact, I've seen it mentioned here a few times already. I like the idea of a lightweight client though. With physical representations of currency, you start to have problems with counterfeiting. And then it's "oops, goodbye deflationary model!"

Could work well. I'm thinking though if you can afford to run a bank and print currency etc etc you can probably afford some type of sat connection to the internet and move bitcoins. I could see a lot of person to person transfers in the future though where neither person on either end had a clue what btc was and just paid a comission to a guy that sent the money to the other end and another guy handed cash to the receiving party. Basically like any other transfer.

Upon further reflection I think that sort of thing might start getting us into trouble. After all, if the bank still has the bitcoins, and if they figure a certain percentage of the bitbucks they issue will be out of their hands at any given moment, they can start doing things like:

1) Using the bitcoins behind the "missing percentage" to speculate. When they lose the bitcoins they will effectively be inflating the bitbucks economy.

2) Printing more bitbucks than they have bitcoins to cover. Let's say the bank is short some money and needs to pay off a debt *now.* Why not print off some bitbucks to represent that "missing percentage?" After all, it's not like there's ever going to be a run on the bank such that they'll have to actually account for that missing percentage.

By creating a physical currency issued by a single entity we will have re-created the Fed and the fiat system we are desperately trying to escape. Let's worry about the local stores that don't accept bitcoins before we worry about the villages that don't have access to the bitcoin network. If we need a physical representation of value, we may as well use one of the established fiat currencies, or else numismatics. Both of these currency classes can be bought with bitcoins and can be used to represent bitcoins physically.

If the time comes when the isolated African village *is* a worry for us, then let one of us who were here in the beginning bring the network to *them.* As society struggles on and the standard of living rises globally, the Internet will become almost as ubiquitous as gasoline or plastics.

I agree with what you're saying but I think it's almost inevitable they're going to do that. You don't need to give them permission.
It would be just how it is now. Bank has 1000 of your BTC in it and a guy comes in for a 100btc loan they'd give him 100 of yours. Now if you go in the next day and want your 1000btc all at once you've got a problem. Obviously on a larger scale more people etc. Also there probably wouldn't be any regulation / reserve rate so who knows maybe they take your 1000 and feel like loaning 50,000 out. It could be worse than the existing system in some ways. Also you know the average person won't understand. They'll see it as "I bought 10k btc and lost it all when it was in the bank, it's not safe".


Maybe being pessimistic but oh well, I'm just looking at how the banks can use it to benefit because they will.

The price of freedom is eternal vigilance. We'll have to work to keep people educated about such possibilities.

We could also potentially pre-emptively create an "open bank" based on the same decentralized principles underlying the Bitcoin network to handle basic bank functions like lending and borrowing. People are going to want loans, and if we don't give them a *good* way of filling that need then some bank-like entity will take advantage of that giant, gaping hole in the market.

Loans, of course, will be tricky. In a deflationary system any sort of interest rate will act as a constantly rising variable interest rate even if it's fixed. It'll be a completely different way of thinking for people since most come from inflationary systems. And even if there is no interest rate on the loan, the rate of deflation will act as an interest rate. If that rate is too high, it could result in bankruptcy for a lot of unwary people. It might be prudent to implement a negative interest rate for such loans so that the rate of deflation plus the negative interest rate works out to be something sane and manageable.

Wow, this is a fun thought experiment. I'm not used to thinking in terms of deflation...

I've mentioned elsewhere that paper bitcoin notes could each have a public key that you could use to check that the bill was backed. The bill could also have a feature that would reveal the private key upon voiding the bill.

As for interest, don't charge it! Why would you, just because you can?

No, you charge interest because it is a key component of the coordination between present and future desires.

I don't see why people keep calling Bitcoin deflationary, it's not. Prices today are chosen based on the amount of fiat currency some number of Bitcoin can buy. As Bitcoin becomes stronger with regard to fiat currency, it seems as if there is deflation, but there really is not. The money supply is increasing at a rate of 300 BTC/hour. Once prices are set primarily in BTC (when producers can buy raw materials with BTC), this "problem" will disappear. Bitcoin has a steady amount of monetary inflation, known by all in advance, which approaches zero. This will allow interest rates to be chosen that take this known inflation into account. I expect these rates will be much lower than today, if not negative.

Technically yes, we are in the inflationary stage of the bitcoin deployment. And yes, the number of bitcoins to be made is set at 21 million. However, we must assume that once all the bitcoins have been deployed there will be a slight rate of deflation as wallets are lost to the destruction of the media on which they are stored. Someone stores their wallet on a personal computer without an off-site backup and a natural disaster or accident destroys the data? All the bitcoins in that wallet are gone. So yes, bitcoins are ultimately deflationary. Once we hit 21 million we will only be able to destroy them.

That is true, but the rate of loss will be inversely proportional to their value. This deflation is quite mild compared to inflation via quantitative easing, interest rate manipulation, and fractional reserve banking.

Well, yes. Obviously! I don't think we'd all be here if that weren't so. :-)

In another thread I was making the analogy that you'd have a local network inside a submarine and the crew wanted to transfer bitcoins to eachother but there might be an occasion when the crew would want to buy a shipment of stuff from a city so ever so often the sub would emerge onto the water surface where the sub would them be in radio range to send the bitcoins. From the feedback that I got it sounded like you can't have a segregated bitcoin network like in my example as it would cause double spending? In any case I did come up with a solution for this. Instead of having one bitcoin network you'd have bitcoin networks. The bitcoin Internet and bitcoin intranets. To connect them to eachother people acting as bitcoin to bitcoin traders would handle the transactions that were off of each bitcoin network so for example bitcoin_submarine would have a bitcoin - bitcoin trader who would collect bitcoins from a crew member and then the bitcoin trader would call another bitcoin - bitcoin trader on the shore and tell that person that they collected 10btc and that the 10btc would go to some address in the city which then would become the responsibility of the bitcoin - bitcoin trader on the shore to make that transaction through the network. I guess what I'm saying is that if bitcoin must rely on the Internet for it to work then this might be an alternative solution so in the case that you make, someone in Africa who does not have a link to the Internet, that village would setup a bitcoin - bitcoin trader, someone who can be trusted by another bitcoin - bitcoin trader and then through a computer or a few computers setup in a LAN configuration that village would be able to make local BTC trades while also making global trades around the world without needing a 100% dedicated Internet connection.

Not really true in this case.  The risk of double spending is if there is a person on the sub trying to send coins from a wallet to another person on the sub, while another person in the greater Internet has a copy of that same wallet trying to spend those same bitcoins.  This excludes the possibility that the sub was trying to generate while under water, which would be futile and therefore should be suspect anyway.  There would still be a near zero risk that coins could be double spent on the sub itself.  A disconnected client is at risk of a double spend, but a disconnected client can certainly itself spend.

Yes, but the sailors can't process the transactions by generating blocks, right? Otherwise, they'd have to fork the chain.

Let's say they fork the chain. When they reconnect, they find out that their chain is shorter (less combined difficulty). What happens to the transactions in their chain? If they are rebroadcast, that might solve the problem. The only thing I see is that they would be unable to generate blocks at current difficulty. Perhaps their clients are smart enough to know when they are disconnected from the main network. When disconnected, blocks are generated at a lower difficulty, which is readjusted every N blocks where N is significantly less than 2016.

No, and if they tried then their blocks would be wiped out each time that they reconnected to the majority network, and those transactions would have to be included in another future block anyway.  The chain split that it would cause wouldn't be a problem for anyone outside of the ship.

Basicly, yes.


A client could detect when it is on the minority side of a network split, although none presently do, by tracking the average time interval between blocks.  If the interval is more than twice the average for several blocks, or more than 6 times the average once, that client is almost certainly disconnected from the majority network.

Still, the clients cannot just decide to lower the difficulty.

Just as a bookend to this thread:
  - http://github.com/bitcoin/bitcoin/issues/271

Also:
Transactions when only one party is online
  - http://bitcointalk.org/index.php?topic=77608.0
 

Update: And these as well:
Desert island economy on Bitcoin without being connected to the internet?
 - http://bitcointalk.org/index.php?topic=106302.0

Creating transactions offline:
 - http://brainwallet.org/#tx