bitcoincard.org

[MoonShadow]

So now this uber light and slim card needs a dongle to function properly? The vapor thickens.

And the maximum range of the dongle is 100 metres according to the website. Presumably the card-to-card mesh communications will have, if anything, an even shorter range.

I suspect that I know the tech being presented here, and if I'm correct, the video is likely understating the practical ranges.  I know that I'm guilty of speculation, but I suspect that the mesh tech alluded to is dash7.

http://en.wikipedia.org/wiki/DASH7

Note that Dash7 devices are inherently location aware, in a relative sense; i.e. they know where there are relative to other dash7 devices.  The customer location tracking features mentioned in the video are one of the encentives for vendors to adopt a set of 'dongles', which would (presumedly) be aware of their own positions in an absolute sense (GPS).

Also note that Dash7 devices can have theortical ranges up to 2km, which practially means about 1km; and that it's max data rate is 200kbps.  However, it cannot do both max range and max speed at the same time, so the video's downgrading of the noted ranges might be due to a concern of overpromising or it might be due to the intent of the card designers fixing the data rate at the max and otherwise living with whatever range remains.

Dash7 isn't good for data that streams or large files, but for mesh propogation of bitcoin data it's practially ideal.  Also the ISM band that DAsh7 uses (433Mhz) is a 32nd harmonic of that which NFC uses as a standard, so both techs can coexist on the same tranceiver and share much silicon, making such features cheaper to produce.

[EhVedadoOAnonimato]

It's not practically possible for a disconnected device (light or otherwise) to be able to use the change from a previous transaction as an input in a new transaction because all other devices should require that the change transaction have 6 confirms or at least a copy of the block's merkle tree as evidence that it's a valid transaction and not a double spend.  Since the device is disconnected, there is no way for it to have that evidence even if some other device has already forwarded the transaction to the network on it's behalf.  In practice, such a device is going to have to keep a set of seperate transaction inputs of varied sizes in order to make multiple payments between online sessions.

Didn't understand this.
AFAICT, this device cannot validate any transaction while offline. It would be a task to the gateway.
And why can't a bitcoincard A send a series of linked-by-change transactions to B, and then B uploads them all to his gateway?

[Ente]

/watching

[MoonShadow]

It's not practically possible for a disconnected device (light or otherwise) to be able to use the change from a previous transaction as an input in a new transaction because all other devices should require that the change transaction have 6 confirms or at least a copy of the block's merkle tree as evidence that it's a valid transaction and not a double spend.  Since the device is disconnected, there is no way for it to have that evidence even if some other device has already forwarded the transaction to the network on it's behalf.  In practice, such a device is going to have to keep a set of seperate transaction inputs of varied sizes in order to make multiple payments between online sessions.

Didn't understand this.
AFAICT, this device cannot validate any transaction while offline. It would be a task to the gateway.

An offline device can create and sign a transaction, it cannot validate it.  However, (in theory, don't know about this device yet) any disconnected device with local copies of it's inputs could also keep the merkle trees that show where those input transactions fit into the blockchain.  This is about as close to an offline validation as it likely possible.

And why can't a bitcoincard A send a series of linked-by-change transactions to B, and then B uploads them all to his gateway?

First the 'gateway' is just another mesh radio connected to the Internet, there is no reason to expect that the gateway itself would be a bitcoin node.  A counter-party to a change based transaction needs to be able to see that the previous transaction fits into a block, otherwise the next transaction is indistingishable from a double spend attempt.  This would be possible if the counter-party were a connected full node & the prior transaction made it back to the internet by another path,but such an event cannot be assumed.

[EhVedadoOAnonimato]

An offline device can create and sign a transaction, it cannot validate it.  However, (in theory, don't know about this device yet) any disconnected device with local copies of it's inputs could also keep the merkle trees that show where those input transactions fit into the blockchain.  This is about as close to an offline validation as it likely possible.

I see. It would require having all block headers, sort of what BitcoinJ does, wouldn't it?
I don't think such a small device would have such ability...

First the 'gateway' is just another mesh radio connected to the Internet, there is no reason to expect that the gateway itself would be a bitcoin node.  A counter-party to a change based transaction needs to be able to see that the previous transaction fits into a block, otherwise the next transaction is indistingishable from a double spend attempt.  This would be possible if the counter-party were a connected full node & the prior transaction made it back to the internet by another path,but such an event cannot be assumed.

Why can't the gateway be a computer with a bitcoin node running?
And even if it isn't, if the entire chain of change-based transactions is provided to the counter-party, where's the issue? It would try to send them all at the same time to the network and only mark as confirmed once they are all in some block. AFAIK you can have an entire chain of dependent transactions confirmed at once in the same block, can't you?

[MoonShadow]

An offline device can create and sign a transaction, it cannot validate it.  However, (in theory, don't know about this device yet) any disconnected device with local copies of it's inputs could also keep the merkle trees that show where those input transactions fit into the blockchain.  This is about as close to an offline validation as it likely possible.

I see. It would require having all block headers, sort of what BitcoinJ does, wouldn't it?
I don't think such a small device would have such ability...

I don't know why not.  Thus far the headers total about 16megs.

First the 'gateway' is just another mesh radio connected to the Internet, there is no reason to expect that the gateway itself would be a bitcoin node.  A counter-party to a change based transaction needs to be able to see that the previous transaction fits into a block, otherwise the next transaction is indistingishable from a double spend attempt.  This would be possible if the counter-party were a connected full node & the prior transaction made it back to the internet by another path,but such an event cannot be assumed.

Why can't the gateway be a computer with a bitcoin node running?

It can, it's just not a requirement.  It's unlikely that most gateways are going to be nodes.  More likley a couple hundred gateways port forwarding or ssh tunneling to one node.

And even if it isn't, if the entire chain of change-based transactions is provided to the counter-party, where's the issue? It would try to send them all at the same time to the network and only mark as confirmed once they are all in some block. AFAIK you can have an entire chain of dependent transactions confirmed at once in the same block, can't you?

No, you cannot.  The followuup transactions are not 'valid' for a block until the transaction prior to it is in a block.  I don't even think that they can propogate past the first full node until the first one is in a block.

[EhVedadoOAnonimato]

No, you cannot.  The followuup transactions are not 'valid' for a block until the transaction prior to it is in a block.  I don't even think that they can propogate past the first full node until the first one is in a block.

Really? An address can't be credited and debited in the same block?
Why not? I mean, why impose such limitation?

(I think we're getting off-topic... I'll open another topic about this)

EDIT: Here it is: https://bitcointalk.org/index.php?topic=86377.0

[genjix]

No, you cannot.  The followuup transactions are not 'valid' for a block until the transaction prior to it is in a block.  I don't even think that they can propogate past the first full node until the first one is in a block.

Really? An address can't be credited and debited in the same block?
Why not? I mean, why impose such limitation?

(I think we're getting off-topic... I'll open another topic about this)

This is false. It happens fairly often in fact that someone sends to an address and then debits in the same block. The only restriction is that the transactions have to be chronologically ordered in the block (due to the way bitcoin blocks are processed)

[genjix]

Also: I saw bitcoincard, and the device is legit. It has only minor kinks that are easy to work out.

I'll avoid posting more as they're quite sensitive to exposing their secrets right now, but the device is really cool. I'm hopeful for it.

First off:
- It's tiny. Credit card sized and < 1 mm thick
- Low cost (cheap device)
- Low power (runs off solar)
- Theoretically it should be secure, but I need to know more of their setup first to be sure their implementation is secure. I'm guessing there may be a couple of vectors they haven't thought of yet, but with proper help they should resolve this.

For merchants, it's trivial and safe to allow Bitcoin payments from these. They also look pretty nifty.

I can imagine one day, Intersango giving away a basket of these for free branded with our logo (or something). They are really cheap and simple devices.

[World]

http://siliconangle.com/blog/2012/06/08/bitcoincard-puts-cryptocurrency-in-your-wallet-with-a-mycelium-smartcard/
"The Bitcoincard moves bitcoin economic interaction offline, which significantly expands both turnover and the target audience. But, most importantly, it makes the clustered local growth of a new free economy possible. Everything starts from the small and mundane, from the things that we encounter on the street everyday. All we need to do is to try to live while using each others’ services directly, without the participation of outside agents such as the dollar, and everything will be smooth sailing."

[Yankee (BitInstant)]

New blog post:

BitcoinCard in Vienna Day 1: Coffee, missing ATM's, and some testing! (Pictures)

http://blog.bitinstant.com/blog/2012/6/8/bitcoincard-in-vienna-day-1-coffee-missing-atms-and-some-tes.html

[xurious]

Haven't read all pages in this post, but simply creating a faraday cage around your card (simple copper sleeve) would eliminate a store "snooping" on where you have been in the store, outside, etc. That is... until you go to purchase something.

[matthewh3]

Haven't read all pages in this post, but simply creating a faraday cage around your card (simple copper sleeve) would eliminate a store "snooping" on where you have been in the store, outside, etc. That is... until you go to purchase something.

Yeah it should have the option to turn the card off or at least the tracking off.  Yes simple cover sleeve to act as a Faraday cage would also be a good idea to protect the card.

[TraderTimm]

Sorry if this has already been mentioned, but if the cards can 'talk' wireless protocols, I suppose you could have a simple bridge-app on your phone to relay the data through the cellular network. Many people who have jailbroken iPhones and Android handsets would be familiar with this.

Nice product, by the way.

[cbeast]

I suspect that this device is so far ahead of its time, that they will not be able to keep up with demand. I kinda hope they just license out the technology to anyone willing to market up-scaled versions and use the licensing fees to make the cheapest version possible for developing countries.

[cypherdoc]

if they can build them for as cheap as they say these will be a real winner.

[Mushroomized]

I want to wear one as a watch!

[genjix]

I have a question to people who think these are a winner.

The issue seems to be a social issue not a technological one. Do people prefer a secure dedicated device over a less secure mobile phone which offers a far superior featureset?

Random thoughts here:

- Devices are so low spec that it makes it difficult to implement any of the blockchain scalability optimisations that can be used for Electrum. Much of the processing is moved server side which assuming it ever did the volume that is mass usage, would be a serious problem.

- I’m not sure that the hardware on these devices can generate enough entropy. A friend (niekt0) told me that ECDSA (unlike other encryption schemes) requires entropy for each signing, not only for generating keys.

- I’ve seen an addon device for mobile phones that allows secure key signing to be possible via an attached USB or similar device. The functionality for a mobile phone is far more advanced that these low spec devices.

- Even if dedicated devices are more secure, mobile phones have more popular adoption. For small amounts of spending cash, their security is adequate. For larger amounts, larger devices or services are fine.

[cbeast]

I have a question to people who think these are a winner.

The issue seems to be a social issue not a technological one. Do people prefer a secure dedicated device over a less secure mobile phone which offers a far superior featureset?

Random thoughts here:

- Devices are so low spec that it makes it difficult to implement any of the blockchain scalability optimisations that can be used for Electrum. Much of the processing is moved server side which assuming it ever did the volume that is mass usage, would be a serious problem.

Have you seen the specs? I'm not sure what processing you are referring to?

- I’m not sure that the hardware on these devices can generate enough entropy. A friend (niekt0) told me that ECDSA (unlike other encryption schemes) requires entropy for each signing, not only for generating keys.

- I’ve seen an addon device for mobile phones that allows secure key signing to be possible via an attached USB or similar device. The functionality for a mobile phone is far more advanced that these low spec devices.

It is difficult to believe that these can work independently without docking at some point to dump data and build a new wallet.

- Even if dedicated devices are more secure, mobile phones have more popular adoption. For small amounts of spending cash, their security is adequate. For larger amounts, larger devices or services are fine.

That's why I think these are ideal for developing nations. They will not only appreciate them more, but will find more creative ways to use Bitcoin.

[Dusty]

The issue seems to be a social issue not a technological one. Do people prefer a secure dedicated device over a less secure mobile phone which offers a far superior featureset?

Usually different persons have different needs.

And while talking generally, at least initially the easier the tech, the better (remember iphone 1 couldn't even receive an MMS!).