The first multivalue database was the Pick operating system. Not only was it an operating system, but a database management system as well. Although the database management system was developed in the mid 1960’s it is still in use today. It may well be one of the best kept secrets in information technology history.
The basis of the database management system is hashed files storing variable length character delimited records. Each file (table) comes in the form of two operating system files. The data file and the dictionary file. By convention the dictionary stores the schema of the file and the data file holds the actual data records.
The hashed files provide excellent performance when the primary key of the record is known. The database records are stored as varaible length ASCII text strings. The variable length records are stored using ASCII character delimiters to delimit Attributes (columns), Values and Sub-values. This provides a method for modifying the database without rewriting the programs that access the data.
The strength of Multivalue databases were they were easy to use and did not require much knowledge of software development to create a working database. The weakness of Multivale databases is that industry experts with little knowledge of software development were able to create poorly designed databases.
Pick as been ported to many different hardware platforms and the database management system ported to many other operating systems including Linux and Windows. There is even an Open Source version of a Multivalue database management system called OpenQM. It is the Open Source version of the QM Multivalue database management system.