BitClip: Now with concept drawings!

[SgtSpike]

And don't you need a full alphanumeric keyboard for typing out bitcoin addresses?

IMHO, people would (should?) not type in bitcoin addresses. They would either load them from a camera or from bluetooth. The device would be better, simpler and cheaper with only a number pad and a few more buttons for navigation and control.

You're making the assumption that both persons involved in a transaction would have a BitClip.  What happens when I want to pay someone for an item on CL, we meet in a public place, I show up with my BitClip, and he shows up with his address printed out, and I have no way to pay him because he doesn't have bluetooth or a way to generate a QR code?

[EhVedadoOAnonimato]

You're making the assumption that both persons involved in a transaction would have a BitClip.  What happens when I want to pay someone for an item on CL, we meet in a public place, I show up with my BitClip, and he shows up with his address printed out, and I have no way to pay him because he doesn't have bluetooth or a way to generate a QR code?

But why can't the camera recognize the printed address the same way it does with the QR code? I was considering it as part of the "expected features".

[Phinnaeus Gage]

Do it!  Given the interest in other novelty projects for bitcoin-related items (bitcoins in resin, bitcoin rings, etc), I wouldn't be surprised to see people jump on something like this like a dog on a... hot dog.

Perhaps the most important item of business is a security audit of the distro used.  People would like assurances that nothing funny was going on in the code in the background.  Not sure how you would be able to provide such an assurance, but it's definitely something to consider.

Prosper.com is a great place to get a microloan too BTW.  Funding took about a week to reach my bank account after the loan was approved and fully funded.  I think it's a great project anyway...

Do it!  Given the interest in other novelty projects for bitcoin-related items (bitcoins in resin, bitcoin rings, etc),

By the end of this week, I'll reveal a prototype of a Bitcoin theme, hand tooled leather wallet, based on this image (but better):



Follow the link in my sig to learn more.

[natman3400]

You should really consider why you would need a full keyboard, in what scenarios are people going to need to type in rediculous bitcoin addresses?

And why not make those very few people use a small usb keyboard instead of increasing the cost of hundreds of units?

In what scenarios would you NOT need to type in an address?  QR codes are the only one I can think of, and not everyone is going to have one of those handy...

How about I just add a button to the QR scanner interface that uploads it to my server for some OCR, and sends back the address. As long as you can get a clear picture, you would probably only have to change one or two characters at most (hell, my 5 year old scanner has some good OCR in its free software suite, it can defiantly convert a bitcoin address).

I've only skimmed the thread , so I appologise if this has no use or was mentioned earlier

Could the technology in development on this website be adapted somehow ?? The raspberry pi , a miniature computer that runs Linux

 http://www.raspberrypi.org/?page_id=11

As I have said, I already have a base hardware, but thanks anyway.

*image here*

Okaithen

Another thing I'd like to strongly suggest: the device should come with an embedded Tor proxy. This is getting particularly more important now that some folks are considering writing software to link IPs to bitcoin transactions.

I'm planning on using VPN to control what the devices can access and vise versa, and as I will probably either not release the logs, or just not keep any, it would be nearly impossible to tell which bitclip was sending the transaction from behind that IP.

You're making the assumption that both persons involved in a transaction would have a BitClip.  What happens when I want to pay someone for an item on CL, we meet in a public place, I show up with my BitClip, and he shows up with his address printed out, and I have no way to pay him because he doesn't have bluetooth or a way to generate a QR code?

But why can't the camera recognize the printed address the same way it does with the QR code? I was considering it as part of the "expected features".

This takes decent CPU power. A luxury this device won't need or have.

[kjj]

You're making the assumption that both persons involved in a transaction would have a BitClip.  What happens when I want to pay someone for an item on CL, we meet in a public place, I show up with my BitClip, and he shows up with his address printed out, and I have no way to pay him because he doesn't have bluetooth or a way to generate a QR code?

But why can't the camera recognize the printed address the same way it does with the QR code? I was considering it as part of the "expected features".

QR was invented specifically because this is hard to do.

[natman3400]

You're making the assumption that both persons involved in a transaction would have a BitClip.  What happens when I want to pay someone for an item on CL, we meet in a public place, I show up with my BitClip, and he shows up with his address printed out, and I have no way to pay him because he doesn't have bluetooth or a way to generate a QR code?

But why can't the camera recognize the printed address the same way it does with the QR code? I was considering it as part of the "expected features".

QR was invented specifically because this is hard to do.

The post office has some pretty powerful OCR that I think there is an open source version of, but its not like you can just plug it into your cheap-a-de-cheapo Pentium 2 PC. I think that the uploading the image might work, but we still need a failsafe entry method.

[DiamondPlus]

The distro would probably just consist of the Linux/BSD kernel, X,  and the bitcoin client (probably custom). What we could do to provide an audit is have someone (probably using bitcoins coming from me) use the device from home for a length of time, and monitor the content of every packet it sends. A specialized device wouldn't use much bandwidth.

[SgtSpike]

You should really consider why you would need a full keyboard, in what scenarios are people going to need to type in rediculous bitcoin addresses?

And why not make those very few people use a small usb keyboard instead of increasing the cost of hundreds of units?

In what scenarios would you NOT need to type in an address?  QR codes are the only one I can think of, and not everyone is going to have one of those handy...

How about I just add a button to the QR scanner interface that uploads it to my server for some OCR, and sends back the address. As long as you can get a clear picture, you would probably only have to change one or two characters at most (hell, my 5 year old scanner has some good OCR in its free software suite, it can defiantly convert a bitcoin address).

Might be worth a shot.  I still like the idea of a keyboard, but it looks like I'm outvoted. 

[natman3400]

The distro would probably just consist of the Linux/BSD kernel, X,  and the bitcoin client (probably custom). What we could do to provide an audit is have someone (probably using bitcoins coming from me) use the device from home for a length of time, and monitor the content of every packet it sends. A specialized device wouldn't use much bandwidth.

And what was the point of quoting me, but not putting it in quotes, and then not saying anything else?

idea of a keyboard, but it looks like I'm outvoted. 

I too like the idea of somekind of keyboard, but we'll have to see how it turns out.

[EhVedadoOAnonimato]

I'm planning on using VPN to control what the devices can access and vise versa, and as I will probably either not release the logs, or just not keep any, it would be nearly impossible to tell which bitclip was sending the transaction from behind that IP.

Oh, you plan to make a client-server solution, not a light-weight bitcoin node. Yeah, makes more sense and will probably be cheaper. Just one thing then: add a way to extract the wallet to a USB device or something, so people wouldn't be entirely dependent on your server in order to spend the coins. Never know, you might get shut down...
Well, I suppose you are already foreseeing something like this for backup purposes anyway.

But why can't the camera recognize the printed address the same way it does with the QR code? I was considering it as part of the "expected features".

This takes decent CPU power. A luxury this device won't need or have.

Ok. But still, I think the full keyboard is not necessary. I wouldn't type in an address. I'd always require a QR code at least. Today, with smartphone apps capable of generating them, it's not that hard.


Thank you for your work, I really hope you manage to conclude it soon.

[natman3400]

Oh, you plan to make a client-server solution, not a light-weight bitcoin node. Yeah, makes more sense and will probably be cheaper. Just one thing then: add a way to extract the wallet to a USB device or something, so people wouldn't be entirely dependent on your server in order to spend the coins. Never know, you might get shut down...
Well, I suppose you are already foreseeing something like this for backup purposes anyway.

No, the devices will still have there own wallet. I'm just funneling the connection so I can make sure nothing bad gets to talk to the devices. You would of course be able to turn this off and either input the nodes or use the bitcoin irc for connections. It also inadvertently doubles as a way to remain more anyonomus. I might have to keep records for legal reasons (read: The FBI requires any "internet service provider" to have some level FBI accessible logging, not sure if I would fall under that), but I wont release them unless hell freezes over and the skies burn red.

[EhVedadoOAnonimato]

No, the devices will still have there own wallet. I'm just funneling the connection so I can make sure nothing bad gets to talk to the devices.  You would of course be able to turn this off and either input the nodes or use the bitcoin irc for connections.

I had understood they would have their own wallets, but I didn't thought they would be "p2p nodes through a VPN" or something. Well, between that, and having a client-server app connected only to your server which is the only true bitcoin node, I think the second solution would be better for you. There's no need for the device to store the blockchain for ex., nor relay bitcoin messages for other nodes and so on. The device only needs to store the private keys locally and then contact your server to know how much money each address has, and to send transactions. Just like the iPhone app is doing.

Of course, there should be a way to save the wallet keys somewhere else, like in USB device. That's both for backup purposes and for not fully depending on your server. If it ever goes down, you export a wallet.dat to a standard bitcoin client. That also eliminates the need for trust, since anyone can verify that your server is not lying about the balance.

I might have to keep records for legal reasons (read: The FBI requires any "internet service provider" to have some level FBI accessible logging, not sure if I would fall under that), but I wont release them unless hell freezes over and the skies burn red.

Please, keep your server in a jurisdiction which doesn't require such thing, and don't keep logs. Or, do everything anonymously. ;-) (see https://bitcointalk.org/index.php?topic=35203.0 )

[natman3400]

No, the devices will still have there own wallet. I'm just funneling the connection so I can make sure nothing bad gets to talk to the devices.  You would of course be able to turn this off and either input the nodes or use the bitcoin irc for connections.

I had understood they would have their own wallets, but I didn't thought they would be "p2p nodes through a VPN" or something. Well, between that, and having a client-server app connected only to your server which is the only true bitcoin node, I think the second solution would be better for you. There's no need for the device to store the blockchain for ex., nor relay bitcoin messages for other nodes and so on. The device only needs to store the private keys locally and then contact your server to know how much money each address has, and to send transactions. Just like the iPhone app is doing.

Of course, there should be a way to save the wallet keys somewhere else, like in USB device. That's both for backup purposes and for not fully depending on your server. If it ever goes down, you export a wallet.dat to a standard bitcoin client. That also eliminates the need for trust, since anyone can verify that your server is not lying about the balance.

I might have to keep records for legal reasons (read: The FBI requires any "internet service provider" to have some level FBI accessible logging, not sure if I would fall under that), but I wont release them unless hell freezes over and the skies burn red.

Please, keep your server in a jurisdiction which doesn't require such thing, and don't keep logs. Or, do everything anonymously. ;-) (see https://bitcointalk.org/index.php?topic=35203.0 )

I live in the united states, and I don't trust any company to do hosting for me.
I want the devices to have the complete and total ability to function on there own. This is both a failsafe and because not everyone is going to trust me with the server position. And who said anything about storing the full block chain? You don't have to do that. And also no need to relay transactions. Just a lightweight, but full, client. And iv'e got encrypted SD cards in the works.

[buttcoin]

I can setup the wiki if you guys need it

[jago25_98]

 I suggest keeping everything on the commandline, such as scanning QR codes. Then, if need be, building it all into an Android device first of all, as a step towards something independent.

 You can already do lots of this with a cheap Android device (like http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=160630685772&fromMakeTrack=true&ssPageName=VIP:watchlink:top:en ) or a phone you already own.

 Once you got that stage you can then scale down the size for something independent and therefore more secure because we planned ahead and the previous developments can run on embedded linux rather than simply Android.

 So I suggest starting with just making Bitcoin on Android really good first - QR codes and everything really easy to use. Just make sure to keep it portable to a smaller form device later on - keep your QR code readers commandline etc. Also I suggest making the whole thing portable/adaptable to work with any future bitcoin-like projects.

[natman3400]

 I suggest keeping everything on the commandline, such as scanning QR codes. Then, if need be, building it all into an Android device first of all, as a step towards something independent.

 You can already do lots of this with a cheap Android device (like http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=160630685772&fromMakeTrack=true&ssPageName=VIP:watchlink:top:en ) or a phone you already own.

 Once you got that stage you can then scale down the size for something independent and therefore more secure because we planned ahead and the previous developments can run on embedded linux rather than simply Android.

 So I suggest starting with just making Bitcoin on Android really good first - QR codes and everything really easy to use. Just make sure to keep it portable to a smaller form device later on - keep your QR code readers commandline etc. Also I suggest making the whole thing portable/adaptable to work with any future bitcoin-like projects.

There is a inherent security risk in using Android, Unix isin't immune, and Android has a fair market share.
And who the hell would but something that only used a command line? Im actually thinking of ways to try and rip as much command line out as i can, just one less thing to infect. And we do have hardware to base it off of. Can somebody help out and find the right name for the pengachu project?

[RodeoX]

Is this it?
http://pubs.media.mit.edu/pubs/papers/pengachu.pdf

[natman3400]

Is this it?
http://pubs.media.mit.edu/pubs/papers/pengachu.pdf

Why yes, yes it is. Thank you.
Considering how little you actually need to run a dinky os with just the bitcoin client (I've tested on a 133mhz computer), we can strip this down quite nicely. While this is based on old technology, I am aware of 133mhz arm processors that cost only about 25-50 cents in bulk.

[DiamondPlus]

Also, there concepts. Not the final word. I would also think that just a numpad would be bad. I might see about getting a base58 keypad made or other. I was thinking we could put a camera in the back that was perma-focused at about 5 inches for scanning qr-codes (maybe even upload images for some sort of OCR? Probably not, but its an idea. I actually based the design off a blockified blackberry, reduced to a numpad and with a d-pad instead of a trackpad. I've personally seen some people type pretty fast and accurate with a numpad though, i've even seen computer keyboards that were just a numpad.

[natman3400]

Also, there concepts. Not the final word. I would also think that just a numpad would be bad. I might see about getting a base58 keypad made or other. I was thinking we could put a camera in the back that was perma-focused at about 5 inches for scanning qr-codes (maybe even upload images for some sort of OCR? Probably not, but its an idea. I actually based the design off a blockified blackberry, reduced to a numpad and with a d-pad instead of a trackpad. I've personally seen some people type pretty fast and accurate with a numpad though, i've even seen computer keyboards that were just a numpad.

If this happens one more time, I'm going to find a way to call copyright .