The database revolution: how SQL changed everything

1970

Before relational databases, storing and retrieving data was a nightmare. Programs in the 1960s used hierarchical or network databases where data was organized in rigid tree structures. If you wanted to ask a new question about your data, one the original programmer hadn't anticipated, you often had to rewrite the entire program.

In June 1970, Edgar F. Codd, a mathematician at IBM's San Jose Research Laboratory, published "A Relational Model of Data for Large Shared Data Banks." The paper proposed something revolutionary: store data in simple tables (relations) with rows and columns. Relationships between data would be expressed through shared values, not physical pointers or file structures.

The beauty was in the separation of concerns. The programmer would describe what data they wanted (declaratively), and the database system would figure out how to get it efficiently. This was the opposite of existing systems where programmers had to navigate the physical data structure themselves.

IBM, despite employing Codd, was slow to act. Their existing database product (IMS) was hugely profitable, and a relational system would cannibalize it. So others moved first. A small startup called Relational Software Inc. built the first commercial relational database in 1979 and named it Oracle. Larry Ellison, its co-founder, had read Codd's paper and realized IBM was leaving money on the table.

IBM eventually released SQL/DS (1981) and DB2 (1983), but Oracle had a head start. The SQL language, originally called SEQUEL (Structured English Query Language), became the standard way to talk to relational databases.

SQL's power was in its simplicity. A query like SELECT name FROM students WHERE grade > 90 reads almost like English. Non-programmers could learn it. Business analysts could query data directly instead of filing requests with the IT department.

Today, relational databases are everywhere. PostgreSQL, MySQL, SQLite, SQL Server, and Oracle power virtually every application that stores structured data, from banking systems handling trillions of dollars to the app tracking your running routes. NoSQL databases emerged in the 2010s for specific use cases, but SQL remains the most widely used database language in the world.

Codd received the Turing Award in 1981. He spent his later years frustrated that SQL implementations didn't fully follow his relational model. He died in 2003, leaving behind an invention that quietly runs the world.