Before ENS, there was GlobalRegistrar — Ethereum's first naming system (Sep 24, 2015)

In September 2015, six weeks after mainnet launch, someone deployed three contracts that became Ethereum's first naming infrastructure: GlobalRegistrar, HashReg, and UrlHint.

Block 282,880 — September 24, 2015

The GlobalRegistrar mapped names to addresses. HashReg mapped code hashes to content hashes. UrlHint mapped content hashes to URLs for metadata retrieval. Together they formed the NatSpec documentation system — the mechanism that was supposed to let Ethereum wallets display human-readable descriptions before signing transactions.

The system was hardcoded directly into go-ethereum v1.4.0 (common/registrar/registrar.go). Not an optional plugin. Core infrastructure.

The frozen TODO that became ENS

Inside the GlobalRegistrar source code (verified on Etherscan, compiled with solc v0.1.1), there's a comment that reads:

// TODO: bidding mechanism

That comment sat frozen for years. The ambition was there — first-come-first-served registration was always meant to be a placeholder. The bidding mechanism didn't arrive until ENS launched in May 2017, when Vickrey auctions replaced the naive assignment system.

ENS didn't appear from nowhere. It solved the exact problem the GlobalRegistrar identified and couldn't finish.

The three-contract system

All three contracts were deployed by the same address within 18 blocks: - GlobalRegistrar: block 282,880 - HashReg: block 282,885 (5 blocks later)
- UrlHint: block 282,898 (18 blocks after start)

The deployer identity isn't confirmed — likely a go-ethereum core developer given the contracts were shipped alongside the go-ethereum codebase — but we haven't pinned the address to a name yet.

Why it matters

Every .eth name you resolve today is the descendant of that frozen TODO comment from 2015. The naming problem was understood from week six. It took two more years to solve it properly.

We documented all three contracts on EthereumHistory.com last night. The verified source code, links to go-ethereum v1.4.0, and the full historical context are there if you want to dig deeper.

Cross-posted from our ongoing archaeology of early mainnet contracts. If you deployed or used these contracts in 2015, we'd love to hear the story.