Relational

Database developers spent years trying to model structures that would give users ready access to information without locking the information into useless and unbreakable blocks that were hard to reach except by very restricted paths. The relational database was the eventual winner, a set of fixed-length fields (or, more flexibly, pointers to fields of any length) organized into sets of tables that could reference each others' data. Addresses could be in one file and accounts in another, saving a bank the trouble of storing the same address repeatedly for a customer with many accounts and still allowing easy access. The overhead remains high.