The source code is organized into modules and subroutines, each containing, where applicable, (1) a number and title for the section, (2) an abstract describing the section, (3) conditions that must be established before code in the section is invoked, (4) conditions that will prevail when the code of the section has completed execution, (5) a verbal description of the algorithm in a right-hand column, and (6) the code that accomplishes the algorithm in a left-hand column. The amount and style of the information presented is intended to supply that which is necessary and sufficient to validate the programs correctness, and which has been used in that way during development and testing of the program.
The code is a subset of the standard programming language C. The subset uses only those features necessary for efficient scientific programs. For example, indexed arrays rather than memory pointers are used to managing data within the program. This results in no loss of speed but affords an improvement in comprehensibility. The style of the coding is also meant to enhance readability and comprehensibility. For example, the local structure of the code is indicated by indentation, but with the block-defining symbols '{' and '}' omitted when unnecessary and placed unobtrusively when they are necessary, to reduce visual noise and clutter. Some keywords are shortened, such as 'double' becoming 'dec', for "decimal", via 'typedef' commands, again to reduce visual noise.