A blockchain is a way for participants to maintain a shared record in linked groups called blocks. The important idea is not that computers stop copying data. They copy and check records so they can agree on an accepted history under the network's rules. That history can record transfers, contract actions or other data.
Begin with a small points ledger
Imagine a club with three record keepers and fictional points. Alice starts with 10 points, Bob with zero and Carol with zero. Alice authorizes a transfer of 7 points to Bob. Once accepted under the club's rules, the balances become Alice 3, Bob 7 and Carol 0.
| Record | Alice | Bob | Carol |
|---|---|---|---|
| Starting state | 10 | 0 | 0 |
| Accept Alice → Bob: 7 | 3 | 7 | 0 |
| Reject Alice → Carol: 5 | 3 | 7 | 0 |
The second request fails because Alice has only 3 points left. Copying the first record to all three record keepers does not give Bob 21 points. It gives three copies of the same accepted 7-point transfer. This is the difference between replicating information and creating additional valid value.
The club model is deliberately simple. Bitcoin uses unspent transaction outputs rather than this account-style table, and real networks have more detailed authorization, fee and consensus rules. The example explains the coordination problem without pretending to implement a real cryptocurrency.
Why link records into blocks?
A block can collect records and include a cryptographic reference to an earlier block. If someone changes old data in a local copy, the linked hashes no longer match the originally accepted chain. NIST describes blockchains in terms of tamper evidence and resistance. That is more accurate than calling every blockchain unbreakable.
Hash links help detect differences, but they do not choose between two competing histories by themselves. The participants also need agreement rules. Bitcoin's whitepaper presents one approach to agreeing on transaction order and addressing double spending through proof of work. Other networks use different mechanisms and assumptions.
What if the record keepers disagree?
Suppose one keeper sees Alice's 7-point request first while another sees the 5-point request first. Both may initially think their first request is affordable. A complete system must define how to resolve the competing order and which result becomes accepted. Simply having several copies is not enough.
This is why confirmation matters. A newly received request, an included transaction and a transaction considered final under a protocol can be different stages. The details depend on the network; a beginner should not assume one universal number of seconds or blocks.
A ledger cannot check every real-world claim
Imagine the club records that Bob delivered a bicycle to Carol. The record can show which account submitted the statement and whether later entries changed the history. It cannot look inside Carol's garage. A false statement can be recorded accurately, so evidence about physical events must come from outside the ledger.
The same distinction applies to digital ownership. Control of a token does not automatically establish copyright in an image, ownership of an off-chain object or an unconditional right to redeem money.
When a normal database may be enough
If club members already accept one treasurer to maintain balances and correct mistakes, an ordinary database with backups and an audit trail may meet their needs. A blockchain introduces costs: replicated validation, coordination, key management and governance. Those costs need a reason, such as allowing parties to verify a shared history under rules they accept.
To assess a real system, ask who may submit records, who validates them, who can read them and who can change the rules. These questions reveal more than a promise to remove all intermediaries, eliminate every fraud or make every process cheaper.
Sources and scope
- NIST: Blockchain Technology Overview
- Satoshi Nakamoto: Bitcoin whitepaper, transaction ordering and double spending
Continue reading: Five foundations to learn next · Hashes, signatures and encryption · Deciding whether blockchain is useful
Reviewed 19 October 2026. Editorial update: Replaced unbreakable-chain and no-copy claims with a shared-points ledger; explained double spending, disagreement and replication; added clear limits and a database comparison.