yinwang1.substack.com/p/database

The claim says RDBMS rows are fixed-width and therefore cannot store arrays or other variable-length data. That is not true.

• MySQL's own documentation says its fixed-length row format is used only when a table has no variable-length columns such as VARCHAR, VARBINARY, BLOB, or TEXT. InnoDB also documents storage for long variable-length columns.
• PostgreSQL explicitly documents that table columns can be defined as variable-length multidimensional arrays.

So while some storage layouts may use fixed-length records in some cases, it is false to describe this as an inherent limitation of RDBMS tables in general.

This sentence compares the introduction dates of Prolog, Datalog, and SQL, but it gives 1986 as SQL's appearance date. That is inaccurate.

• IBM Research lists the paper "Sequel: A structured english query language" as a 1974 publication, showing that SQL's predecessor SEQUEL was already publicly described in the mid-1970s.
• Oracle's official history of SQL says SEQUEL was developed by IBM and that the first commercially available implementation of SQL appeared in 1979.
• IBM's overview of SQL likewise says SQL was developed in the 1970s, became commercially available in 1979, and was only later standardized by ANSI in 1986.

So 1986 is the standardization date, not the date SQL first appeared. That matters here because the article uses the wrong date to argue that SQL came after Prolog and Datalog.