Smart contracts & decentralized applications

Step through a deposit-then-release script. Every action is a pure function of the current state — no hidden variables, no server making a judgment call. Anyone who replays these same 3 actions against the same starting state lands on the exact same numbers, every time.

1. A contract's state is just data

   This escrow's state is { phase, balance, buyer, seller } — nothing more than that. It starts "created" with balance 0.

2. Actions transition state deterministically

   $\text{applyAction}(\text{state}, \text{action}) \to \text{state}'$ is a pure function: deposit moves created → funded, then release or refund moves funded → released or funded → refunded. Same state, same action, same result — always.

3. The state hash is what nodes actually agree on

   $\text{stateHash}(\text{state}) = \text{toyHash}(\text{phase} \mathbin\Vert \text{balance} \mathbin\Vert \text{buyer} \mathbin\Vert \text{seller})$. Two nodes that replay the same actions independently compute the same hash without ever talking to each other — the determinism of the code is the consensus mechanism, not a vote about it.

4. This is what 'code is law' means operationally

   There's no customer-support line to appeal to. Whatever applyAction computes from the actions actually submitted is final — which is exactly why the code has to be gotten right before it runs, not fixed after.

• phase — created, funded, released, or refunded; the contract's only source of "what's happened so far."
• action — one of deposit, release, refund; the only way the state is allowed to change.

Final state hash: 400 — anyone replaying these same 3 actions computes this exact number, no coordination required.

Fund the contract, then pick which way it resolves. Both branches run through the exact same applyAction function — nothing about the code changes based on who's watching or what they'd prefer the outcome to be.

Alice deposits 50 into escrow with Bob as the seller.

1. Created

   $\text{stateHash}(\texttt{created}, 0, \texttt{alice}, \texttt{bob}) = 905$.

2. Funded

   Alice deposits 50: phase becomes funded, balance becomes 50. $\text{stateHash}(\texttt{funded}, 50, \texttt{alice}, \texttt{bob}) = 400$ — a completely different hash, because the state actually changed.

3. Branch: released

   If the condition holds, release fires: $\text{stateHash}(\texttt{released}, 50, \ldots) = 367$.

4. Branch: refunded

   If it doesn't, refund fires instead: $\text{stateHash}(\texttt{refunded}, 50, \ldots) = 485$. Same starting state, same deposit — a completely different final hash, because the action taken from there was different.