The author revisits an old draft on modern relational query language improvements, inspired by discussions around Acadia. They critique SQL's implementations, noting issues like syntax, query planners, and data types. Drawing from PL/I, Prolog, and functional languages like F# and Elixir, they propose syntax improvements, better parsers, and enhanced user-defined types.

The author highlights PostgreSQL's strengths, influenced by QUEL and [PERSON_NAME]'s relational model, and mentions the need for sum types, pattern matching, and lazy evaluation. They also discuss the limitations of stored procedures and the importance of declarative programming over procedural approaches. The article invites further discussion on refining SQL for modern programming needs.