A decentralised application, or dApp, can look just like any other website. It may have buttons, forms, charts, and a familiar login screen. The difference is not mainly visual; it lies in where the important rules and records live.
A typical dApp has three layers. The frontend is the website or mobile interface. A wallet connects the user’s account and signs instructions. Smart contracts on a blockchain hold the shared state and enforce the core rules. The interface might say Swap, but the contract determines what assets move and under which conditions.
Reading from a dApp can be almost invisible. The website asks a node for contract data and displays balances or prices. Writing is different: changing blockchain state requires a transaction. Your wallet shows what is being requested, you sign it, the network executes it, and gas is paid even if the transaction later fails.
Decentralisation is not an all-or-nothing label. A contract may be immutable while its website is hosted by one company. Its data may depend on a centralised server, or an administrator may retain an upgrade key. A useful evaluation asks which parts can be censored, changed, or switched off—and by whom.
Wallet-based access removes the need to create a new username and password for every service, but it changes the security model. A signature can authorise a harmless login, a token approval, or a valuable transfer. Users must understand the request because there may be no support desk capable of reversing a mistaken signature.
Good dApps make these boundaries clear: they show contract addresses, transaction details, risks, and the source of external data. Open contracts can be inspected and reused, but open code does not mean bug-free code. Audits help; they never turn software into a guarantee.
One puzzle remains. A smart contract can read blockchain data, but how can a lending app learn the rupee price of ether or whether it rained in Delhi? Tomorrow we meet the bridge called an oracle.