EventChain: A Distributed History Without a Distributed System
Abstract
Today each party to a process keeps its own records of it and reconciles them with the others’ after the fact. When the records disagree, the party that cannot prove what happened carries the loss. Blockchains promised one shared history, but a blockchain agrees on each entry before it writes it, so every member had to build and pay for its system before the first entry was worth writing, and several consortia closed before they reached that point.
This paper defines EventChain, a protocol and file format for that shared history. It states what each guarantee assumes, compares the guarantees and costs with those of public and permissioned blockchains on dated published figures, and takes five processes through it in worked examples. A reference implementation accompanies the specification.
EventChain allows different holders to know that every other holder checking the same chain of events holds the same history through the entry it names. That is the certainty a blockchain promises, and no party has to join a network to get it. An entry costs one signature and one write. One company starts alone, the work grows with the number of events, and adding a party adds a copy of the file.
What becomes worth recording is every reading, revision, approval and handoff that a blockchain made too expensive to record one at a time. A dispute then starts from evidence both sides already hold, and a payment or an approval can follow the evidence the moment it arrives. A valid chain shows which key signed what, and in what order. Whether a statement was true stays with its signer, and what to do about it stays with the application that acts on the evidence. Where strangers settle payments with one another, a blockchain’s round of votes remains the right tool.
Revision 91a4776 — October 2026