CryptoGraffiti - Block Chain Message Encoder & Decoder

[Hyena]

What exactly is the format used for encoding? ASCII encoded in base58 format (minus the version, and the checksum)?

Bitcoin address in its raw format contains 20 bytes of data. These 20 bytes are later converted into base58 and a checksum is appended, thus we get a typical bitcoin receiving address. To decode an address I restore the underlying 20 bytes from the bitcoin address and seek common ASCII characters from it. If more than 90% of the characters are visible ASCII characters then the address is assumed to be human constructed and thus is probably a human readable text.

[Giulio Prisco]

Very cool! Is this a Twitter on the blockchain? Are there possible ways to permit editing messages already sent?

You cannot edit anything that has been saved in the block chain unless you control more than 50% of the network's hashing power.

Of course, but perhaps we can replace old pointers to transactions in the blockchain with new pointers.

For example, I write a graffiti "I Love Lucy." Then I stop loving Lucy and start loving Linda. I make a new graffiti "I Love Linda" and replace the old "I Love..." graffiti with the new one (in CryptoGraffiti, not the blockchain).

[Hyena]

Of course, but perhaps we can replace old pointers to transactions in the blockchain with new pointers.

For example, I write a graffiti "I Love Lucy." Then I stop loving Lucy and start loving Linda. I make a new graffiti "I Love Linda" and replace the old "I Love..." graffiti with the new one (in CryptoGraffiti, not the blockchain).

That's a nice idea, thank you! I could implement a possibility for the message's creator to remove the message from my database. The message will still be in the block chain but my service would not display it any more. The process would require the user to prove ownership over one of the input addresses that were used to make the transaction. The latter is easily done with the signature functionality that comes with Bitcoin's private keys.

[molecular]

Is it similar to www.proofofexistence.com?

I don't know about that site, but when I did "proof of existence" for some casascius coins, the process was different:

proof of existence doesn't necessarily have to store the data in a readable form in the blockchain. In my case I just used the hash of an image as a private key and published a transaction to the respective address. This has the advantage that I can remove the money again from that address so I'm not polluting the unspent transaction output list with dust. It has the disadvantage of not actually having the data in publicly readable form in the blockchain (the private key (data) cannot be derived from the address).

example:

click for larger version

the proof can still be conducted, but only with an unpruned full blockchain available.

[m_yaw]

Feature request:
It is currently really difficult to find one's graffitis after a while. I propose a search function for btc address or transaction, or maybe even part of the text of the graffiti!

[Hyena]

Feature request:
It is currently really difficult to find one's graffitis after a while. I propose a search function for btc address or transaction, or maybe even part of the text of the graffiti!

Thank you for your idea. We have these features in our TODO list but at the moment we are working on a new framework which is a prerequisite for all these advanced use cases.

[HeroCat]

I think it must be implemented in the BTC wallets, as option 

[andrewbb]

The service is free?  Or 5500 Satoshi's per message block?

For locating the messages, do you search for your Bitcoin address?  Or is it a pre-pended identifier in the Blockchain?

Have you considered approaching Google to index the Blockchain?

[andrewbb]

Two questions:
1. The date/time stamp on the latest messages on your site shows 2014/09/24.  It is 10/24.  Is this a programming typo?

2. If instant is selected, how long does it take for it to show on your website (assuming Blockchain.info shows the transaction)?

[andrewbb]

Two questions:
1. The date/time stamp on the latest messages on your site shows 2014/09/24.  It is 10/24.  Is this a programming typo?

2. If instant is selected, how long does it take for it to show on your website (assuming Blockchain.info shows the transaction)?

Answer to #2 is after 1 confirmation.

Answer to #1 is it appears to be a programming typo.

New question:
I posted a message to the blockchain and it broke it into 4 separate messages.  Why?

[Hyena]

The service is free?  Or 5500 Satoshi's per message block?

For locating the messages, do you search for your Bitcoin address?  Or is it a pre-pended identifier in the Blockchain?

Have you considered approaching Google to index the Blockchain?

Service is free. It is currently hardcoded to the user interface to have 5500 satoshis per output address as it was the minimum amount some time ago. However, right now the Bitcoin-core wallet lets you send 1500 satoshis to a single output at minimum. The CryptoGraffiti decoder does not care about the number of satoshis sent to an address, all it cares about is whether the address contains human readable text or not.

edit 2:
I locate messages by attempting to decode every BTC transaction that has been included in a new block. If I find any human readable text then I will store the transaction hash as a pointer to an "interesting" transaction. I haven't approached Google.

Two questions:
1. The date/time stamp on the latest messages on your site shows 2014/09/24.  It is 10/24.  Is this a programming typo?

2. If instant is selected, how long does it take for it to show on your website (assuming Blockchain.info shows the transaction)?

1. This is a typo that will be fixed in the next release.
2. Under normal conditions the message should show on CryptoGraffiti.info instantly (3-10 seconds). Heavy load or network lag could cause it to take longer.

Answer to #2 is after 1 confirmation.

Wrong, when instant is selected it will appear on CryptoGraffiti.info in 3-10 seconds.
edit: However, due to the bug in the current UI you would need to refresh the page yourself though.

New question:
I posted a message to the blockchain and it broke it into 4 separate messages.  Why?

CryptoGraffiti.info considers 1 message to be encoded in a single bitcoin transaction. If you're using some other message writing service then it can happen that a single message gets written into multiple bitcoin transactions. Such messages cannot be intuitively extracted because there is no known standard to determine the order of those messages.

What might also be the case is that you failed to make a send-to-many bitcoin transaction.

[andrewbb]

What might also be the case is that you failed to make a send-to-many bitcoin transaction.

Thanks.  I am using MultiBit which doesn't support the send-to-many.

[Hyena]

What might also be the case is that you failed to make a send-to-many bitcoin transaction.

Thanks.  I am using MultiBit which doesn't support the send-to-many.

Thanks for pointing that out. I haven't actually used any other bitcoin wallet than Bitcoin-core and Electrum. However, now that I know some wallets haven't implemented multisend I must raise the priority of getting automated message writing implemented soon. The latter would allow you to write a message by making a transaction with a single output to our to-be-developed message encoder which would in turn make the real multi-output transaction for you. The drawback of such a payment proxy would be waiting for confirmations.

[Buffer Overflow]

Could this service be changed so it uses recommended OP_RETURN outputs for messages? This would then keep these outputs out of UXTO.

[andrewbb]

FYI, the site does not work on the iPhone.

[Grand_Voyageur]

FYI, the site does not work on the iPhone.

iPhone is not a meangiful Platform.  You must turn to Android ones for better. 

[Hyena]

Could this service be changed so it uses recommended OP_RETURN outputs for messages? This would then keep these outputs out of UXTO.

It can but that does not stop people from encoding messages into transaction outputs. The current approach allows free, manual and immediate message storing while OP_RETURN would require a payment proxy that would make the actual transaction. The GUI of a typical bitcoin wallet such as Bitcoin-core does not let you to play around with OP_RETURN.

FYI, the site does not work on the iPhone.

Thanks for the info, I created an issue for the UI developers about that.

[Giulio Prisco]

Shouldn't CryptoGraffiti be implemented as a sidechain, with the option to make as many changes to Bitcoin Core as the application needs? I started a related thread here. https://bitcointalk.org/index.php?topic=838813.0

[Hyena]

Shouldn't CryptoGraffiti be implemented as a sidechain, with the option to make as many changes to Bitcoin Core as the application needs? I started a related thread here. https://bitcointalk.org/index.php?topic=838813.0

Seems like a plausible idea. If the same computer network secures sidechains too then it will probably work out pretty well.

edit: But... that still does not stop people from encoding messages into Bitcoin's block chain.

[Hyena]

Sad news, both of the new Graphical User Interface developers quitted after failing to complete the new version in 2 months. They worked 5 hours per week and were offered 25% share (per face) of the donations/profit the site generated. One said "Right now it's difficult to work assuming that one day Bitcoin may become SOMETHING."