Bitcoin Forum

Hi there! Today I'm releasing the first version of a simple Bitcoin point of sale application. It was mainly written to demonstrate the use of the green address feature, but can also be used independently of that. It is written in Python and works in combination with a standard Bitcoin client. It is targeted at standard PC hardware, e.g. laptop (used by merchant) + external monitor (facing the customer).

Here is a video of the system in action: http://www.youtube.com/watch?v=o84SfChQ-S8 . (Update: Here is another, shorter video of just the payment being performed: http://www.youtube.com/watch?v=DNpcf9rSBIk .)

More details and explanations, the code and a number of screenshots can be found on the project's page on GitHub: https://github.com/javgh/greenaddress-pos-tools .

Looking forward to your feedback! :-)

Update (Jan 2014): The system has been extended to use NFC and Bluetooth. See this thread for details: https://bitcointalk.org/index.php?topic=395469.0 .

looks sweet :)

I'd like to point out, that I made an effort to make it fairly straightforward to run this application. Granted, it's only tested on Linux for the moment, but it has very few dependencies and can run in combination with a standard Bitcoin client (so no patches needed or anything). You can use any Bitcoin mobile client, unless you want to test the green address feature. Then you will need BitPay with my patch applied (see the GitHub link for more info about that). But hopefully, the version of BitPay in the Android market will soon be updated to make this step easier as well.

What do you think, how cheap could a hardware solution based on this be put together? I'm thinking old, used laptop (maybe an Eee PC) plus smallish external monitor. Maybe all in all for $150? Maybe less? .. I think we could reach the point, where this is a straightforward and fairly inexpensive way to get a brick and mortar merchant set up for Bitcoin payments. I think maybe the only missing piece in the puzzle would be automatic exchange to the merchant's local currency (e.g. USD) to have it be a complete worry-free (regarding exchange rates) solution. It might be interesting to interface with a Bitcoin exchange here.

+ very nice!

Jav, your work is really exciting. but the space is a problem. If a merchant has to find a place for another PC or Laptop in their store, it will not be the ultimate solution.

Awesome!!!! I love how easy that is. Really well done.

The issue of table/counter space is an issue though... but wow I'm impressed.

Jav is a genius product manager.

I hate to say this but this still suffers from all the issues that previous suggestions for point of sale stuff has, in that it takes a great deal of time comparatively to confirm the transaction is actually valid. People aren't going to wait around for a quarter of an hour after buying a stick of gum to wait for a transaction to verify.

People also aren't going to bother spending hours of their time or thousands of dollars trying to execute a successful double spend just to steal a pack of chewing gum.  Accepting without any confirmation in such a circumstance is probably lower risk than letting people pay with credit card.

This is a great step forward and nice video.  When the time is right, we'll want to integrate this (or something like this) with bit-pay.com.

Green addresses. Look it up. Confirmation can be instantaneous.

There's green addresses, but also all other online wallet solutions seem to have instantaneous and free transfers within their system (flexcoin does, I think vibanko does as well, infamously mybitcoin used to use the feature as a selling point).

Now if someone could come up with a "clearing house" kind of solution that includes all online wallets, as well as exchanges and regular wallets... they would make life a lot easier, no?

This could be great for automatic point of sale eg. vending machine because there would be no need to store and transport cash, and goods selling price could be lower!

Hey! Thanks for your great comments! :-)


I'm not sure how much of an issue the space requirement really is. I mean, a cash register can be pretty clunky as well. But it all depends on the situation I guess, and I agree, that a smaller solution would be much nicer. One reason this is open source is to be able to iterate towards a better solution. I think there are many options to be explored: Maybe a Sheevaplug could be used in combination with a USB touchpad and a USB display. Or maybe even a cheap Android phone to which a USB display is somehow hooked up (not sure how feasible that is, but I have seen some hacks in that direction). The latter would require to port the GUI parts to Android.


Instantaneous transfers among users of a specific service isn't really that special (Instawallet has it as well) and doesn't - in my opinion - really lead anywhere. Take the example of Flexcoin: Nothing against it, but merchants won't be limiting themselves to a small subset of possible Bitcoin customers. Unless that subset is close to all of Bitcoin customers and that is simply not going to happen. Bitcoiners are much too wary - and rightfully so - to let a single payment processor get so powerful.

So what we need is an open protocol that allows to split this responsibility over multiple payment processors. It should have a fairly low barrier of entry for new payment processors and it should be easy to get rid of established payment processors should they start to become annoying. The green address technique is an attempt at formulating such a protocol. It could be implement by Flexcoin and Vibanko and the merchant would then see "Verified by Flexcoin" or "Verified by Vibanko" and it's up to the merchant to decide whether that's good enough.

Hi Jav,

Well done.

I'm interested in completing the missing puzzle (automatic exchange) and adopting your solution into our structure, id be interested to know if this would work with World Bitcoin Exchange, please PM me when you get sometime.

Andre

This is the most promising short-term solution for instant payments IMO.
A web of trust between payment processors could make it better. For example, the merchant trust flexcoin but not instawallet. If flexcoin trusts instawallet, the transaction could be like this:

1) Payer spends its funds in instawallet
2) instawallet pays to flexcoin
3) flexcoin pays to the merchant.

I don't remember if I've already suggest you this. Step by step, first having this working.

Congratulations!

Would it not be possible to make that option accessible to the customer with a message that says: Do you have an account with Vibanko, Instawallet, MtGox, Tradehill (insert whatever new solutions might come here)? Then the customer can choose that particular QR code and the transaction is instantaneous.

The merchant's QR-code can contain the valid options and the smartphone software can know your preferences in advance and pay with the valid option you prefer automatically.
Going back to my earlier proposal, at this point, if you don't have an account with any of the valid options, you could send them to your payment processor and he can ask a partner processor to do the payment for them. If the partner can't do the transaction neither, he can ask to other processors. In the end, payment processors will establish an effective ripple network.
The bitcoin community could gain a lot by supporting the decentralized implementation of Ripple.

Again, this green addresses idea is very promising.

The only thing I don't like about green addresses is that as a recipient if I refuse to take the money and send it back, it is not charged back to the person that sent it to me.

I had a discussion about donations to a political party as it might turn out to be very funny with the laws in place in Germany:
Parties are not allowed to take anonymous donations above 500€.
Immediately returning the money counts as not taking it.
If they can't refuse to take them (as they went to them cash in an envelope/in a mail with a bitcoin wallet attached/sent from a green address) they have to give the money to the president of the German Bundestag.
I would love to see what this guy dues with bitcoins :)

For donations you don't need green addresses nor instant payment.
It would be nice to see the president of the German Bundestag with bitcoins, but 500€ is an expensive joke.

Very cool!  I really like the idea of green addresses.

If you actually watched the video you would have noticed that the actual payment was recognized within seconds, and if you did any cursory research on the subject you would have realized it's infeasible to do a double spend for a small amount and as such it is very safe to accept small transactions without confirmations.

Many stores are already switching to cash registers with displays for the customer, at least in the Netherlands, so things like this could easily be integrated into existing POS platforms.

This seems like a great solution for smaller stores, but larger stores or chains that use custom POS platforms would benefit a lot from a standardized 'POS API' (as opposed to the usual Bitcoin API) that also allows for communication with a remote Bitcoin client (which will be the case in a store with multiple payment terminals/counters, where everything has to be handled by one central 'cash register server'), and that can be used without a GUI frontend - effectively making it very easy to implement Bitcoin into an existing POS solution without any knowledge about the underlying system.

Also, a small shop doesn't need a pc nor a laptop. Smartphones with dual core are sold today. You don't even need dual-core for this. A small screen and a smartphone is something any business can afford today. Two years from now this can easily and cheaply be mainstream.

I would say, though, that it'd be rather strange to see a smartphone mounted onto a counter, displaying a QR code :)

Actually, you could even get a cheap Android tablet or ebook reader (the latter will work better) from some random Chinese store, and run a simple application on it that will wirelessly connect to your 'payment server', and show the appropriate QR code and payment information. You can get a low-spec Android tablet for <$100 from China, and some simple ebook readers with epaper sell for about $100-$150 (with WiFi capability).

Yeah I think an Android tablet app is a good way to implement this. The bonus is, it's a touch screen, so you can have the user interact with it in interesting ways. Like, after the payment, you could have "And now scan this to collect a coupon!" or some way to sign up for news from your favorite merchants, get a link to their website, select a delivery address using a touchscreen map, etc. Lots of possibilities.

But this video is very cool.

I was thinking in an smartphone with hdmi connected to a screen, but yes, I tablet would be more powerful.

Some thing like this would not take up much room. http://www.hengyu-tech.com.cn/db_picture/pro1/201010310840081320.jpg (http://www.hengyu-tech.com.cn/db_picture/pro1/201010310840081320.jpg)

This forum supports embedding images. I bet you find the right button to do it.
http://www.hengyu-tech.com.cn/db_picture/pro1/201010310840081320.jpg

Thank You, LOL

Do you have a update on this project?

Thank You

No, I'm afraid there hasn't been much happening regarding this project. Although I have recently started to collaborate with someone (user ne1), who would like to combine this with an Arduino, to get some hardware projects going. Should be interesting to see what comes out of that.

To the tablet comments here: I agree, a cheap Android tablet seems like a good hardware choice.

Before I ask my question I must first admit I'm not a programmer, but a business man.

Could a idea like yours be integrated into, lets say, a open source POS software system?

The reason I ask is because the only way that Bitcoin becomes widely accepted (in my opinion) is if it can be accepted along with credit/debt cards without the need of additional software or equipment.

It is easier to sell someone a POS system, then it is a new currency 

I think that should be possible.


I am not sure what exactly you have in mind here. Your requirement is no additional equipment as well as no additional software? That seems tricky (?). But I'm not very familiar with current POS software systems and what kind of setups are in use today.

 
Sorry, i was referring to the idea that some have about building "credit card" processing machines that process just bitcoin transactions.

It seems easier to sell a small business a POS system then a $200 piece of equipment just to process bitcoins. 

I'm curious: why do you think that would be easier? That "credit card" processing machine will cost something as well, won't it? And if the solution here can be deployed on a cheap Android tablet where it might be possible to hit a price point of around $100 it seems to me those two options are pretty much the same in terms of up front cost.

And would a small business really be that concerned about a one time cost of $100 if they wanted to try Bitcoin? Actually, it might be interesting to explore alternative businesss models here. Maybe if - for example - Bit-Pay would offer an affiliate model, you could give a tablet like this for free to a small shop, help them get set up and everything and try to get your money back through affiliate earnings. It might be a little early for this kind of setup (probably not enough people would pay with Bitcoin at this point to make enough in affiliate earnings), but it might be an option at some point.

Even cheaper.

http://venturebeat.com/2011/10/26/aakash-android-tablet-exclusive/

$60 without subsidies.
And we know these things tend to get cheaper...
I still think that a smartphone could be enough.

This project had been dormant for a while, but I reused it now for a technology demo on how a Bitcoin payment via NFC might look like. See this thread for more details and a video of the solution: https://bitcointalk.org/index.php?topic=395469.0 .