An overview of the CAPITL software development environment

Abstract

The CAPITL programming environment is comprised of a shared, object-oriented, versioned database, an embedded logic-based data-manipulation language, and a graphical user interface. With each software object the database stores a rich set of attributes that describe its syntax, intended semantics, and relationship to other objects. CAPITL is implemented in POL, a data model and deductive query language with elements of persistent, object-oriented and logic-based programming languages. POL is implemented in and tightly coupled with C+ +. A request for a derived object consists of a partial description of its attributes. A planner written in POL searches the database for tools and sources that can be combined to create an object meeting the description. Since tools are stored in the database like other objects, plans that create tools as well as intermediate inputs are possible. A builders also written in POL, executes plans to materialize software products. The builder verifies that existing objects are current, minimally re-applying tools as sources, tools, or system descriptions change. After an overview of the database and the POL programming system, we outline CAPITL’s logic-based approach to system modelling, illustrating it with two examples. We conclude with a status report and an outline of future directions.