A Prehistory of the Ethereum Protocol

Key ideas

Buterin traces how Ethereum's protocol changed from a narrow Mastercoin proposal to launch, showing collaborators, failed designs and deliberate compromises shaping it along the way.

1. 1

   From a narrow proposal to general contracts

   Buterin writes that his first design, sent to the Mastercoin team after a visit in October 2013, covered only two-party contracts. The team liked it but would not drop their work for it, while he grew more convinced of the direction. After realizing on a San Francisco walk that smart contracts could potentially be fully generalized, he rebuilt the design so contracts became accounts that hold assets and keep permanent storage.

2. 2

   Collaborators reshaped the design

   Buterin says the protocol was his alone only until new people arrived around December 2013. Gavin Wood implemented contract calls synchronously without realizing the spec intended otherwise; both preferred that, so the spec changed. Buterin also credits a discussion with Wood for the sender-pays gas model, and largely credits Wood for the shift from programmable money toward general-purpose computing, while admitting one of his own choices, a register-based machine, had little argument behind it and was later reversed on others' suggestion.

3. 3

   A run of discarded mining algorithms

   Buterin recounts a run of mining algorithms. Dagger came first, adaptive proof of work collapsed under long-range attacks, the hardware experts consulted were fairly pessimistic about Random Circuit, and a SAT-solving idea was rejected. The work ended up back near the start with Dagger Hashimoto, which after rounds of tweaking became Ethash.

4. 4

   Modest versions won over ambitious ones

   Buterin describes logs winning over a Merkle tree of full execution traces as a compromise between simplicity and completeness. Native contracts with miner-voted gas prices were dropped for lack of a safe design, in favor of a few precompiles. Heaps were skipped for lack of time, and he says later audits and attacks showed they were harder to do safely than the team had thought.

5. 5

   Miner voting, then a code freeze

   When Andrew Miller showed that Buterin's dynamic gas limit could be exploited and no sturdier rule emerged, the team let miners vote, judging that any specific approach was likelier to fail than miners were to abuse the power. Buterin then describes the path to release: audits in early 2015, a proof of concept in March intended as the final one, a four-month testnet with bugs and consensus failures that stabilized around June, and a code freeze in July.

In Vitalik’s words

What’s available here

• Original source

  Linked. Opens on the publisher’s site.

• Player on this page

  Not applicable. Written piece; nothing to play.

• Full text or transcript

  Not reviewed. No license to show a text version here.

• Our summary

  Draft. Written by us with AI from the full public text and checked against it; quotes are word for word. Not yet rights-reviewed.

• Ask questions with AI

  Not reviewed. Off until there is licensed text to work from.