Update 2 : Improved structures and explanations
In the current Duniter protocol all documents are managed in a textual format. While being practical as readable and verifiable by humans, it's less beneficial for programs usage since it needs parsing, fields content verifications and waste space with unused codepoints of text formats.
We propose a new protocol format which reuses ideas of the current protocol while providing a more concise and less error-prone structure while staying simple enough so it can be analysed with appropriate tools or even converted to text.