langwhy

SQL

Query· alive, actively evolving· since 1974 · Donald D. Chamberlin, Raymond F. Boyce · ISO/IEC JTC 1/SC 32 (WG3)

Declarative language for defining, changing and querying data in relational databases.

Draft from 2026-10-30, not yet checked claim by claim against its sources. Help verify it.

Born from

In 1970 Edgar Codd of the IBM Research Laboratory in San Jose argued that users of large data banks should not have to know how the data is organized, and he faulted the tree and network systems that made programs follow fixed access paths. In 1974 Donald Chamberlin and Raymond Boyce, also at IBM, published SEQUEL, a structured English query language aimed at occasional users as well as programmers.

What it trades away

no memory modelmemory
no type systemtypes
not a runtimeruns as
n/aconcurrency

Cost: Each database speaks its own dialect, and no relational database fully conforms to the SQL standard yet.

Where it fits

SlotFit
database●●●
data●●●DuckDB embeds SQL in a host process and integrates with Python and R.
server○○○SQL runs inside the database server; it is not a language for writing the server.

Why it exists

In 1970 Edgar Codd, a researcher at IBM’s laboratory in San Jose, published a relational model for large shared data banks. His central demand was that users should not have to know how the data is organized in the machine. He criticised the tree and network systems of the day, where programs had to follow fixed access paths through the data (Codd, 1970). Donald Chamberlin and Raymond Boyce, also at IBM, set out to give that model a query language. Their 1974 paper called it SEQUEL and aimed it at occasional database users as well as programmers (IBM Research). Chamberlin later said they wanted to put Codd’s ideas into plain English that anyone could follow (interview).

What it trades away

SQL is mostly declarative. A query says which rows it wants, and the database decides how to fetch them, so tuning means learning the optimizer rather than writing loops (Wikipedia). The standard is also too large for any product to implement completely. PostgreSQL states that as of version 18 it conforms to at least 170 of the 177 mandatory SQL:2023 Core features, and that no relational database fully meets the standard yet. Products differ in syntax and function, so a query written for one database often needs edits before it runs on another (PostgreSQL).

Where it won

SQLite is on every Android device and iPhone, every Mac and Windows 10/11 PC, and in Firefox, Chrome and Safari (SQLite). Oracle introduced the first commercially available SQL-based relational database in 1979 (Oracle). Google’s Spanner, a globally distributed system, moved to the SQL dialect shared across Google’s other systems (Google Research). DuckDB brings SQL to analytics inside Python and R sessions (DuckDB).

Named wins

Evolution

Reacts to: –

Weigh it against: Python, r, julia

Timeline

Links

website · wikipedia · docs · PLDB

Sources

  1. https://www.engineering.upenn.edu/~zives/03f/cis550/codd.pdf
  2. https://research.ibm.com/publications/sequel-a-struciured-english-query-language
  3. https://www.mappingthejourney.com/single-post/2017/10/12/episode-11-interview-with-don-chamberlin-designer-of-sql-database-language/
  4. https://docs.oracle.com/en/database/oracle/oracle-database/21/cncpt/database-concepts.pdf
  5. https://www.dbvis.com/thetable/standard-ansi-sql-what-it-is-and-why-it-matters/
  6. https://research.google/pubs/spanner-becoming-a-sql-system/
  7. https://sqlite.org/mostdeployed.html
  8. https://duckdb.org/why_duckdb
  9. https://www.postgresql.org/about/
  10. https://en.wikipedia.org/wiki/SQL

Unverified draft, 2026-10-30.Edit the facts · edit the story · JSON · Markdown