Base58 messages Updated +Created
Bitcoin addresses are by convention expressed in Base58, which is a human readable binary-to-text encoding invented by Bitcoin.
It is a bit like Base64, but obsessed with eliminating characters that look like one another in popular but stupid fonts like capital "I" and lower case ell "l". As such, any embedded text is rather obfuscated due to this limitations, and people often resort to leet-like replacements such as '1' to represent 'I'.
This seems to be one of the earliest strategies used to encode messages into the Bitcoin blockchain. The first known example appears in 2011. Then starting November 2011, a large number of messages were inscribed n short successsion, presumably by a single person or small group.
The interest in Base58 encoding might have initially arisen with people's desire to have "vanity addresses", that is Bitcoin addresses that have real words in them, much like vanity plates or vanity numbers. Such addresses with long words in them are hard to find while keeping the address spendable, because they have to correspond to a private key. An extreme notable example is:which contains the awkward 13 letter word:
embarrassable
in it. TODO: proof that it is pendable?
Perhaps inspired by this, some people also decided to use Base58 addresses as a way to create more general unspendable inscriptions, even even though the method is much more clumsy and complicated than P2FKHS. There is however a certain art to working under limitations.
Figure 1. . Although it is not solely focused on inscriptions and may also contain functional burn addresses, it is likely that the methods of Khatib/Legout capture the overall trend of base58 inscription counts.
These messages were originally found with: github.com/cirosantilli/bitcoin-inscription-indexer#payload-size-out-utxo-2vals which tracks the largest transactions with unspent outputs.
Bitcoin Burn Addresses: Unveiling the Permanent Losses and Their Underlying Causes later revealed many new ones.
Finding Base58 messages is intrinsically hard for a few reasons
  • the words may be garbled by Base58 leet
  • only very small ammounts of data can be encoded at a time, and all of it contains ASCII, so you can't just "find all long ASCII strings" as we started doing for other ASCII inscriptions a la strings -n20; you have to use some dictionary as a basis
  • the Base58 does not show up raw on the blockchain, as it is just a human representation for the actual binary data that does, so you can't just strings the blockchain, you have to parse it
The interesting following transactions contain base58 encoded messages on addresses, sorted chronologically, and heighlighted either due to their earliness or historical or artistic quality:
Related: