High level data management for the problem solver

Abstract

An APL version of an information retrieval system used for numerical taxonomy, is shown to have properties suitable for data management by APL users. The data model is that of a table whose columns are named by descriptors, and the data manipulation language is English-like. Statements in the language are strings of executable APL functions with intervening arguments, thus avoiding the limitation in extensibility of non-procedural query languages. Boolean selection predicates are generalized to include executable APL expressions. The system features are presented in detail and extensions are discussed.