/mnemonic-encoding

Human readable encoding for small amounts of data

Primary LanguageJavaMIT LicenseMIT

Human Encoding

The goal of this project is to provide a variety of encoding (and decoding) schemes for reliably storing or transmitting small amounts of data in a human readable, writable, speakable, and listenable way. There is no single best encoding scheme for all purposes, so advice and best practices around scheme use cases will be provided.

The encoding is represented as an ordered list of words, typically called a mnemonic phrase. "Mnemonic" is a bit of a misnomer, as they are used much more often written down than memorized. The words come from a wordlist which typically has a length that is power of 2.

This project does not deal with the use of mnemonic phrases for the purpose of further seed generation, although such projects could benefit from the use of this library.

This project has no external dependencies.

Use cases

  • Paper backup for seed data
  • Paper backup for private keys
  • Relaying small amounts of data over a secure text channel
  • Relaying small amounts of data over a secure voice channel

Research TODOs

  • Wordlists that are built specifically to minimize phonetic ambiguities
  • Wordlists that have stronger guarantees (ability to automatically deal with mistyped characters)
  • Best practices around creating and storing paper backups (archival paper, pencil vs pen, polyester sleeves, etc)

References

Build status

  • Master: Build Status

Other TODOs

  • Licensing issues around encoding schemes
  • Better bit manipulation
  • Benchmarking with JMH
  • Mutation testing with PIT