Coding Standards for Database Files › Coding Standards for COBOL Programs › Program Layout › New Topic › Use of GO

Use of GO

Only use a GO statement to branch to a point within the same subroutine. Never use a GO statement to branch from a subroutine to a point in the mainline code. Although this can be accomplished using COBOL, it should be regarded as being contrary to the basic tenets of structured programming.

Avoid nesting structured programming operations too deeply. For example, more than three or four levels of nesting in a given section level. If more are needed, use another section.

Programs should be as structured as possible. A program is not ‘structured’ just because it only uses structured operation codes. A structured program is one that is modularized in an efficient way and built up out of the structured programming constructs—SEQUENCE, CONDITION (CASE) or ITERATION, and structured combinations thereof. The constructs might even be implemented logically (for instance, with GO statements used in a structured manner) rather than with specific HLL structured operation codes.

The COBOL statements used to code the reading of a group of records from a file should be highly standardized.

A standard loop should be used because:

• It stresses the ‘device independence’ of the data. The file name, which is all that differs between different instances of the loop, appears at the beginning of the code.
• It serves as a standard construct that other programmers can instantly recognize as signifying the retrieval of a set of records. Although COBOL only has operation codes that will process one record at a time (such as READ), there is very often a requirement to process a whole set of records from a file (for example, all order records for a given customer). The loop construct emphasizes the ‘set’ nature of your processing, which is generally easier to understand.

READ loop - COBOL ‘85: