11. The Blockchain: Data Structure, Ledger, and Immutability

almost every important ledger in ordinary life is controlled by one trusted institution. The bank decides what your balance "really" is. The government's registry decides who "really" owns the house. We do not each keep our own authoritative copy of the national land registry; we trust the keeper to maintain the one true copy honestly.

building a ledger nobody controls is not mainly a storage problem; it is an agreement problem. The hard part is getting thousands of strangers' computers to agree on one history of events, in one order, when some of them are dishonest and the network itself is unreliable.

imagine a sealed, numbered diary. Every page is numbered in sequence, and at the top of each new page, before writing anything else, you copy out a fingerprint of the entire previous page — every word it contained. Page 50 therefore "knows" exactly what page 49 said; page 49 in turn knows what page 48 said; and so on back to page 1. The pages are not merely stacked in order; each one cryptographically commits to the one before it.

you cannot quietly change one page of the sealed diary. Because every later page's fingerprint depends on it, altering one old entry breaks the chain at that point and forces you to forge every single page from there to the end. If you trust only the most recent page's fingerprint, you can verify the integrity of the entire history behind it — one check vouches for everything before it.

Haber and Stornetta's 1991 scheme is the sealed, numbered diary, described seventeen years before Bitcoin. Each entry is stitched to the previous ones with hashes, so the order and integrity of the whole history is locked in. Their motivation was proving when something happened; the structure they invented to do it is, recognizably, a blockchain.

a blockchain is a kind of data structure — a particular way of arranging records, namely the hash-linked chain of blocks just described. A distributed ledger is a kind of system — a network of computers that each keep a copy of such a record and follow shared rules to agree on what it contains. The first is a thing on a hard drive; the second is a living arrangement of many machines and the people running them. The relationship is like that between a book's binding (a structure for ordering pages so none can be secretly inserted or removed) and a library system (the institution, staff, and procedures that keep many identical copies in sync across many branches).

the magic is not in the chain of blocks. It is in handing identical copies of that chain to a crowd of strangers and giving them a rule for agreeing on each addition, so that no one of them — and no small group — can quietly change the past. The blockchain is the ledger's format; the distributed network and its consensus rule are what let the ledger run without a trusted keeper.

a private, members-only ledger is easier to run because you already know and have vetted everyone at the table. A public ledger that lets total strangers in must work much harder — and spend much more energy or capital — to stop one person from pretending to be a million people. Much of the cost and complexity of Bitcoin exists to solve a problem that permissioned systems simply define away.

why would anyone want a database that is slower, more expensive, harder to change, and thousands of times more redundant than the alternatives? The honest answer is: you would not — unless you specifically need a shared record that no single party controls and that no participant can secretly rewrite. The blockchain trades efficiency for the elimination of the trusted keeper. Where a trusted keeper is available, cheap, and acceptable, an ordinary database is the better tool. The blockchain earns its keep only where the keeper itself is the problem.