Curriculum

The double-spend and shared-truth problem

A group of strangers who do not trust each other, and have no shared clock and no referee, must agree on one history of who paid whom — permanently.

units: Unit I, Unit II · experiments: 1

Wallet → Sepolia → faucet → transaction → block → explorer

Vocabulary stays abstract until the learner produces one traceable transaction on a public testnet and reads it back.

units: Unit IV · experiments: 3, 9

The engineering problems Satoshi had to solve

Blockchain is usually taught as a finished artefact. It was actually a sequence of distributed-systems problems answered one at a time.

units: Unit I, Unit II · experiments: 1

Byzantine agreement, proof-of-work and chain selection

Nodes must converge on one history while messages are delayed, clocks differ and some participants lie.

units: Unit II · experiments: 1

Encode, encrypt, hash, sign — with the right words

Most blockchain misconceptions are vocabulary failures: hashing called encryption, signing called encrypting with a private key.

units: Unit III · experiments: 2, 7, 10

From leaves to root, and Merkle proofs to SPV

A light client cannot download every transaction, yet must still verify that one transaction is in a block.

units: Unit III · experiments: 8

UTXO, script, fees, subsidy and confirmations

Account-model intuition silently breaks when reading real Bitcoin transactions.

units: Unit II · experiments: 1

Accounts, state, gas, EVM, proof-of-stake

Ethereum is not 'Bitcoin plus smart contracts' — the state and consensus models differ fundamentally.

units: Unit IV · experiments: 3

Write, compile, reason about state

Contract code is judged by what state it can be forced into, not by whether it compiles.

units: Unit V · experiments: 4, 6

Local VM → tests → Sepolia → verification

Deploying before testing turns debugging into guesswork paid for in gas.

units: Unit V · experiments: 5

Web3 libraries, RPC, transfers and contract calls

A front end must handle pending, reverted and replaced transactions — not just the happy path.

units: Unit V · experiments: 11

Hyperledger Fabric: peers, orderers, channels, chaincode, MSP

Enterprise requirements — known identity, privacy, governance — change which consensus design is appropriate.

units: Unit VI · experiments: 12

Reentrancy, access control, randomness, oracles, key compromise

Security is a design property; it cannot be added after deployment to an immutable contract.

units: Unit V, Unit VI · experiments: 4, 5

End-to-end: problem → design → contract → tests → Sepolia → evidence

Implementer standard means a working system you can defend, not a set of completed exercises.

units: Unit V, Unit VI · experiments: 4, 5, 11