[BOUNTY CLOSED] Open Source (CC) Paper Wallet Kit for safe offline coin storage

[aantonop]

The final agreed template has two blocks in the backup stub so that it can be printed in the "Basic with Human-Readable Backup Stub" layout.

All the designs will accommodate printing that mode.

My personal preference is that be the default layout - the backup stub has no public key (you can generate it from the private if you need), but has TWO versions of the private key - one as a QR and one as human-readable. That offers the greatest recoverability, with three possible places to recover: the main note QR, the backup QR and the backup-backup human readable.


The final template is the one that supports what you ask for, I think we're good!

[acorn]

Sorry again to those of you waiting for the final template.ai file. Illustrator is unavailable to me at the moment (time to get some new computers for the office..) but I'll be getting to a suitable computer after work to post it. I will save it so that is compatible with earlier versions of CS, since it's come to my attention that my earlier version is optimized for CS6 and some of you may be a few versions earlier.

ETA: 6:00PM PST!

[canton]

Heavy paper, hard to shine light through

Here's some really interesting reading:
http://www.patentstorm.us/patents/6106932/description.html

In particular:

The substrate, or base, of many instant-win lottery cards is cut from cellulosic board stock. A metallic foil is then laminated to the board stock. The metallic foil is coated with a primer to minimize oxidation and to provide a surface that is receptive to ink. After symbols have been printed atop the primer in pigmented ink, a sealant is frequently applied over them to protect the printed symbols.

Three cheers to everything we can learn from the long and venerable practice of cheating the lottery! I suspect that no matter how thick the card stock, "candling" with lasers or bright scanners might be a problem... So I wonder if the final printed product could contain a reflective yet inkjet/laser-printable square where the QR codes and addresses go? Might add more security, waterproofing, *and* look shiny and pretty & valuable... though doubtless more expensive.

[jonls]

The following are relative to the NOTE’S top left corner.

A PLACEMENT: 0.75, 0.685 inches
A DIMENSION: 1 x 1 inches

Thanks for the offsets! Something seems to be wrong here though, when compared to your templates. E.g A is placed almost as far into the note as it is wide. Should I subtract the margin (0.5in) from these numbers?

[acorn]

The following are relative to the NOTE’S top left corner.

A PLACEMENT: 0.75, 0.685 inches
A DIMENSION: 1 x 1 inches

Thanks for the offsets! Something seems to be wrong here though, when compared to your templates. E.g A is placed almost as far into the note as it is wide. Should I subtract the margin (0.5in) from these numbers?

D'oh! You're right. Those offsets are from the page's top left corner! Yes, subtract the margins from the offsets.

[aantonop]

There are special holographic stickers that project a woven image over barcodes if photographed or scanned, making them impossible to read. I plan to include some in the "Pro" kit I will sell.

Overall, the security of paper wallets is as much a matter of process as it is a matter of technology. A big part of this project will be the instructions for users that will help them implement the process of paper wallets securely. These instructions should address issues of off-line software, printer "memory", suitable ink, storage options (wall-safe etc), considerations for fire/water, suggestions for backup storage etc. These instructions are what will take the security technology provided and elevate it to a secure process.

By the way, the more bitcoin rises, the worse this problem gets... I need more paper wallets to spread the value, so as to spread the risk of loss. I can't have a 10BTC wallet - that's insane ($2000 on one piece of paper? No thanks). I need 100 x 0.1 BTC instead. A year from now, I will need to spread the same risk to 1000 x 0.01 BTC. More notes means more opportunities for errors and accidental disclosure or loss, which in turn makes the process itself more important.

[aantonop]

The following are relative to the NOTE’S top left corner.

A PLACEMENT: 0.75, 0.685 inches
A DIMENSION: 1 x 1 inches

Thanks for the offsets! Something seems to be wrong here though, when compared to your templates. E.g A is placed almost as far into the note as it is wide. Should I subtract the margin (0.5in) from these numbers?

D'oh! You're right. Those offsets are from the page's top left corner!

If Adobe Illustrator does Measurement Guides, we should use them for the template. If not, I can do it in AutoCAD. They look like this, showing the exact position from which you are measuring.

|<---- 5 3/8 in ---->|

[jonls]

D'oh! You're right. Those offsets are from the page's top left corner! Yes, subtract the margins from the offsets.

Ok, seems to work much better I've updated the generator to fit your provided measurements. In addition I have tried to fix printing multiple pages but I'm not entirely sure that the margins are generated correctly. I have switched to setting the margins using a CSS @page rule which means that in Chrome margins now simply have to be set to "default" and in firefox this method should also work. However, the generated PDFs seem a little off and I can't do a test print at this moment. Testing and bug reports of this version are very welcome: https://github.com/jonls/openpaperwallet/tree/update-template

[aantonop]

Pricing Considerations

I am putting together some price estimates for a retail paper wallet "kit", containing the following:

- Instructions
- Software for printing the keys
- 10 sheets of pre-printed, perforated notes (offset printing, heavy archival paper)
- 4 sheets of test-templates on plain laser paper for calibration
- 120 x security stickers
- 40 x plastic sleeves

The kit would allow you to print and secure 40 paper wallets. Estimated wholesale COST of the kit: ~ $10 for production runs of less than 5000 sheets.

Questions for the thread:

1) Is this the right "size" for a starter kit, or should it be smaller/bigger?
2) Would you buy such a kit, if yes, how much would you pay? (min/max range please)
3) What are the top 3 reasons you would buy rather than DIY?
4) Would you want to sell such a kit, if yes, how much would you charge retail to make it worthwhile? (range please)


Any and all feedback welcome!

[acorn]

LATEST TEMPLATE FILE:
http://m-ato.me/bitcoin/templatepreferredfinal.ai

I will remove links in previous posts to the older templates just in case.

[superbit]

Being able to pre-order the printer paper is a great idea.  What about supplying consumers with pre-printed waterproof-tear proof paper to print their keys on?  Very affordable when buying 5000 sheets or so at a time.

[aantonop]

Being able to pre-order the printer paper is a great idea.  What about supplying consumers with pre-printed waterproof-tear proof paper to print their keys on?  Very affordable when buying 5000 sheets or so at a time.

Exactly: easy, secure and beautiful paper wallets, pre-printed so you can add the keys.

How much would you pay?

[canton]

LATEST TEMPLATE FILE:

Thanks Acorn! Here's my first round of design. I've taken the liberty to adjust the template a little bit. Most significantly, I wanted to make it possible for the backup/tear-off stub to have QR codes AND alphanumeric codes for both the public and private keys so we don't have to pick. The only font family I use here is Ubuntu (the font used for the traditional bitcoin logo.)



Here's a photo so you can see the actual size including the scratch-off sticker size.



The one concern I have about the template right now is that it will require a great deal of cleverness to make a full-bleed design that accommodates only four horizontal perforations. My design is likely impossible because we'd need to give the printing service approx 1/8th inch wiggle room for inaccurate perforations.

Similarly, it's ambitious to expect that end users will be able to properly calibrate their printers beyond 1/8th inch accuracy.

Should we instead consider having eight horizontal perforations and plan for a 1/8th inch bleed to the design, like this?

[aantonop]

I don't think a professional offset printer will have any problem handling 1/8 inch accuracy on printing or the perforation die-cut. They can achieve 1/64 on the color alignment and at least 1/16 on the cut.

Users may have problems aligning printers as accurately, which is why I was planning to include a few calibration sheets.

Any changes to the template have to be weighed against:

a) The work acorn has already done and will need to re-do
b) The work the other designers have done and will need to re-do
c) The relative value of the change.

If changing the template will make this easier or better, let's do it. I cannot accurately judge how much work it will be, but you give good reasons to consider.

I think rather than making changes, we should first test out some of these issues. You see to like testing the hypotheses (which is great!).

Who is able to run a couple of tests to check if a regular home printer can align a print onto a specific area accurately enough to avoid problems with the "bleed" margins?

One last point: I personally believe there is no need or purpose for the PUBLIC key on the stub, for two reasons. Firstly, you don't need to load a wallet often, It's better to make a new one for security. Secondly, you can recreate the public key from the private key import. So having the private WIF is enough to get the receiving address back. I'd rather have two copies of the private WIF on the stub to double my chances of recovery, than waste one space for a public key I won't use. I have about 100 paper wallets right now and I have never used the public key twice. I have however, had difficulty with one QR code already from a smudge. So my personal experience says two private keys, no public on the stub.

[Timbo925]

Used the new template to edit my designs. Will be all adding them to this shared folder. (https://drive.google.com/folderview?id=0B1-OuUkg4oEBRC0tVTFIM0ZuSUU&usp=sharing)
If you see a something wrong or have a good idea to change something, let me know (color scheme can always be changed)


Current Front Designs (More in the link):


(Also without the BTC)


(Also without the BTC)


(Also without the BTC)


(Also without the BTC)


(Other variation in link)

[canton]

Who is able to run a couple of tests to check if a regular home printer can align a print onto a specific area accurately enough to avoid problems with the "bleed" margins?

Oops, I meant to redact this concern! I don't think the bleed issues will impact end users at all. For them we just have to make sure that the final design accommodates at least 1/8" (maybe even 3/16") wiggle room for the placement of the QR codes and keys -- e.g. slightly generous whitespace around the QR code areas.

Here's a virtual test of how my own design would fare if the die cut perfs had a 1/32" vertical variance:



As you can see the left part of my design doesn't do so well, but it's not such a big deal with the right side because of the direction of the gradient.

Whether we have 5 horizontal perfs with no accounting for bleed or 8 horizontal perfs with some added bleed doesn't matter to me -- I can design accordingly -- I just want to confirm what kind of vertical shift on the cuts we should anticipate so that the design can gracefully accommodate any variation.

[Timbo925]

Isn't the back the only place you should be worried about this problem?
If you cutout the bill with a exacto knife you can make a perfect cut if you do it patiently. Then we just need to design a back of the notes with some extra space to cover the bleed. Sounds like the easiest solution.

[canton]

If you see a something wrong or have a good idea to change something, let me know

I like it a lot! A couple ideas:

1) For the alphanumeric full-length private key in the square, if we use a fixed width font and rely on the fact that (as I understand it) WIFs always have 51 characters, then we can always show the key in a tidy pretty square like this:



With the 7 "firstbits" larger for easier checking, and then exactly 4 rows of 11 characters.

2) Regarding bleed/perforation offsets, your design should do great even with a 5-horizontal perforation template -- except for the darker "ripples" in the upper left corner (which would ripple onto the bottom of the bills above it).

If you cutout the bill with a exacto knife you can make a perfect cut

I'm coming into this project pretty late so I might be confused on this matter, but I thought the plan was to have diecut perfs so you can just fold and tear -- no exacto knifes needed.

[Timbo925]

Thank for the tips. Faded the top out of the orange lines so if it is not perfect, it should look ok.
The format of the private key is indeed a good idea. Currently using Ubuntu Mono as the fixed width font. The normal Ubuntu font is used in the 'official' bitcoin logo so it should be the perfect fit.
Updated previous post to not spam the topic with pictures.

Must have missed the part about the use of diecut perfs. In that case it might be a good idea to only use horizontal gradients to fix most of the problems.

[proff]

Hi aantonop

I assume your kit is meant for "average" users? If so, isn't the software arguably the most delicate component of this? Since I do not see that it has been explicitly mentioned here, your kit should include something like a bootable DVD or USB stick with secure software pre-installed on it, together with easy-to-follow instructions on how to verify the checksum. Also do not forget that since you are booting into a more-or-less known state it is absolutely essential to introduce enough entropy before generating any random numbers, etc.