In Procedure Division rules, the term identifier means the complete specification of a data item. The term refers to all words required to make your reference to the item unique.

To reference a data item that is a function, a function-identifier is used. For information on functions, see Chapter 7.

The general formats for identifiers are as follows:

For more information on the methods of uniquely specifying data items, see the following:

• Section 6.2.1, Qualification
• Section 6.2.2, Subscripts and Indexes
• Section 6.2.3, Reference Modification
• Section 6.2.6, Scope of Names

If the name you use as a condition-name appears in more than one place in your program, it can be made unique through qualification, indexing, or subscripting. Your condition-name also is unique when the scope of names conventions by themselves ensure this as described in Section 6.2.6, Scope of Names.

The first qualifier for a condition-name can be the name of the item with which it is associated (the conditional variable). When qualifying condition-names, you must use the name of the conditional variable itself or the names of items that contain it.

References to a condition-name must have the same combination of subscripting or indexing that you use for the conditional variable.

The formats you use to ensure unique condition-names are the same as those used for an identifier, except condition-name replaces data-name.

In Procedure Division rules the term condition-name refers to a condition-name along with any qualification and subscripting or indexing needed to avoid ambiguity.

6.2.6 Scope of Names

A contained COBOL program can refer to a user-defined word in its containing program if the user-defined word has the global attribute. (See Section 1.2.1.1 in Section 1.2.1.) Some user-defined words always have the global attribute, some never have the attribute (that is, they are local), and some might or might not, depending on the use of the GLOBAL clause. The following rules explain how to use different kinds of user-defined words and what kinds of local and global name scoping to expect.

1. The following types of user-defined words are always local and can be referenced only by statements and entries in the program declaring them:
   • Paragraph-name
   • Section-name
2. These user-defined words are always local when you define them in the Report Section. Only those statements and entries in the program containing the entries can reference them:
   • Condition-name
   • Data-name
   • Record-name
3. The following user-defined word is always local when you define it in the Screen Section. Only those statements and entries in the program containing the entries can reference it:
   • Screen-name
4. Because you cannot specify a Configuration Section for a program contained within another program, the following types of user-defined words are always global when declared in the Configuration Section. You can reference them only by statements and entries either in the program that contains the Configuration Section or in any program contained within that program:
   • Alphabet-name
   • Condition-name (declared in the Special Names paragraph)
   • Mnemonic-name
   • Symbolic-character-name
   • Switch-name
   • Class-name
5. The following user-defined words are global if you specify the GLOBAL clause:
   • Condition-name (declared in the Data Division)
   • Data-name
   • File-name
   • Index-name
   • Record-name
   • Report-name
   • Segmented-key-name (if you specify the GLOBAL clause on the corresponding file-name)

Specific conventions for declarations and references apply to these types of user-defined words whenever the previous conditions do not apply.

Whenever duplicate names exist, a program always references the resource in its own program. If the resource is not in the referencing program, the following two conventions are used:

• Conventions for resolving program-name references
• Conventions for resolving other references

The next two sections describe these conventions.

Conventions for Resolving Program-Name References

The PROGRAM-ID paragraph of the Identification Division declares the program-name; a user-defined word to identify the program. Only the CALL and CANCEL statements and the END PROGRAM header can reference a program-name.

A run unit can contain multiple programs with duplicated program-names. However, when two programs have duplicate program-names, one of the two programs must directly or indirectly be contained within a separately compiled program that does not contain the program with the duplicated program-name.

The following rules regulate the scope of program-name:

1. Within a run unit, any separately compiled program can reference any other separately compiled program.
2. If a program-name does not have the COMMON attribute and it is contained directly within another program, the contained program can be referenced only by statements included in the directly containing program.
   For example, in the run unit consisting of the three separately compiled programs illustrated in Example 6-3, Example 6-4, and Example 6-5:
   • MAIN-PROGRAM (See (1) in Example 6-3) directly contains program PROG-NAME-A (3) and indirectly contains PROG-NAME-B (5), PROG-NAME-C (6), PROG-NAME-D (7), and PROG-NAME-F (9).
   • PROG-NAME-B (See (11) in Example 6-4.)
   • PROG-NAME-E (See (12) in Example 6-5.)
   
   The CALL "PROG-NAME-B" statement in PROG-NAME-A (See (4) in Example 6-3.) references PROG-NAME-B (5) in the same separately compiled program (MAIN-PROGRAM) because PROG-NAME-B (5) is directly contained in PROG-NAME-A. All other CALL "PROG-NAME-B" statements ((2) and (8) and (10) in Example 6-3 and (13) in Example 6-5) all reference PROG-NAME-B (11) in Example 6-4, the second separately compiled program.

Example 6-3 Separately Compiled Program 1

IDENTIFICATION DIVISION. PROGRAM-ID. MAIN-PROGRAM. (1) . . . CALL "PROG-NAME-B". (2) . . . IDENTIFICATION DIVISION. PROGRAM-ID. PROG-NAME-A. (3) . . . CALL "PROG-NAME-B". (4) . . . IDENTIFICATION DIVISION. PROGRAM-ID. PROG-NAME-B. (5) . . . IDENTIFICATION DIVISION. PROGRAM-ID. PROG-NAME-C. (6) . . . IDENTIFICATION DIVISION. PROGRAM-ID. PROG-NAME-D. (7) . . . CALL "PROG-NAME-B". (8) . . . END PROGRAM PROG-NAME-D. END PROGRAM PROG-NAME-C. END PROGRAM PROG-NAME-B. . . . IDENTIFICATION DIVISION. PROGRAM-ID. PROG-NAME-F. (9) . . . CALL "PROG-NAME-B". (10) . . . END PROGRAM PROG-NAME-F. END PROGRAM PROG-NAME-A. END PROGRAM MAIN-PROGRAM.

Example 6-4 Separately Compiled Program 2

IDENTIFICATION DIVISION. PROGRAM-ID. PROG-NAME-B. (11) . . .

Example 6-5 Separately Compiled Program 3

IDENTIFICATION DIVISION. PROGRAM-ID. PROG-NAME-E. (12) . . . CALL "PROG-NAME-B". (13) . . .

3. If a program-name has the COMMON attribute and it is contained directly within another program, the contained program can be referenced only by the following:
   • Statements in the directly containing program
   • Statements in any programs, directly or indirectly contained within the directly containing program, except statements in the program with the COMMON attribute and in any program it directly or indirectly contains.

For example in the run unit consisting of the three separately compiled programs illustrated in Example 6-6, Example 6-7, and Example 6-8:

• MAIN-PROGRAM (see (1) in Example 6-6), which directly contains PROG-NAME-A (3), and indirectly contains PROG-NAME-B (IS COMMON) (5), PROG-NAME-C (6), PROG-NAME-D (7), PROG-NAME-F (9), and PROG-NAME-G (11).
• PROG-NAME-B (See (13) in Example 6-7.)
• PROG-NAME-E (See (14) in Example 6-8.)

The CALL "PROG-NAME-B" statement in PROG-NAME-A (See (4) in Example 6-6) references PROG-NAME-B IS COMMON (5) because it is directly contained in PROG-NAME-A. The CALL "PROG-NAME-B" statement in PROG-NAME-F (See (10) in Example 6-6) references PROG-NAME-B IS COMMON (5) because PROG-NAME-F is directly contained in PROG-NAME-A. The CALL "PROG-NAME-B" statement in PROG-NAME-G (See (12) in Example 6-6) references PROG-NAME-B IS COMMON (5) because PROG-NAME-G is indirectly contained in PROG-NAME-A. The remaining CALL "PROG-NAME-B" statements ((2) and (8) in MAIN-PROGRAM and (15) in PROG-NAME-E) all reference the separately compiled program, PROG-NAME-B (13).

Example 6-6 Separately Compiled Program 1

IDENTIFICATION DIVISION. PROGRAM-ID. MAIN-PROGRAM. (1) . . . CALL "PROG-NAME-B". (2) . . . IDENTIFICATION DIVISION. PROGRAM-ID. PROG-NAME-A. (3) . . . CALL "PROG-NAME-B". (4) . . . IDENTIFICATION DIVISION. PROGRAM-ID. PROG-NAME-B IS COMMON. (5) . . . IDENTIFICATION DIVISION. PROGRAM-ID. PROG-NAME-C. (6) . . . IDENTIFICATION DIVISION. PROGRAM-ID. PROG-NAME-D. (7) . . . CALL "PROG-NAME-B". (8) . . . END PROGRAM PROG-NAME-D. END PROGRAM PROG-NAME-C. END PROGRAM PROG-NAME-B. . . . IDENTIFICATION DIVISION. PROGRAM-ID. PROG-NAME-F. (9) . . . CALL "PROG-NAME-B". (10) . . . IDENTIFICATION DIVISION. PROGRAM-ID. PROG-NAME-G. (11) . . . CALL "PROG-NAME-B". (12) . . . END PROGRAM PROG-NAME-G. END PROGRAM PROG-NAME-F. END PROGRAM PROG-NAME-A. END PROGRAM MAIN-PROGRAM.

Example 6-7 Separately Compiled Program 2

IDENTIFICATION DIVISION. PROGRAM-ID. PROG-NAME-B. (13) . . .

Example 6-8 Separately Compiled Program 3

IDENTIFICATION DIVISION. PROGRAM-ID. PROG-NAME-E. (14) . . . CALL "PROG-NAME-B". (15) . . .

6.2.6.1 Conventions for Resolving Other References

When a source program declares condition-names, data-names, file-names, record-names, report-names, and segmented-key-names, only the declaring source program can reference these names. The only exception is when names have the GLOBAL attribute and the program contains other programs.

For example, when a program such as PROG-NAME-A (See (1) in Example 6-9) contains other programs (PROG-NAME-B (10) and PROG-NAME-C (25)), each program can define the same user-defined word. When such duplicated names are referenced, the rules for qualification of names (see Section 6.2.1) apply; and, if necessary, the following three hierarchical rules resolve any ambiguity:

1. References in a program to names defined in that program are resolved within the program. For example:
   • The following names: (2), (3), (4), and (5) are both defined and referenced within PROG-NAME-A.
   • The following names: (11), (12), (13), (14), and (15) are both defined and referenced within PROG-NAME-B.
   • The following names: (26) and (27) are both defined and referenced within PROG-NAME-C.
2. A program cannot reference any condition-name, data-name, file-name, record-name, or report-name defined in a program it contains. For example, statements in PROG-NAME-A (See (6), (7), (8), and (9)) cannot reference items in either PROG-NAME-B or PROG-NAME-C. statements in PROG-NAME-B (See (16) through (24)) cannot reference items in PROG-NAME-C.
3. If a program contains another program, any GLOBAL names defined in the containing program can be referenced by the following:
   • Statements in a directly contained program, provided that the directly containing program does not declare the same user-defined word, in which case, rule 1 applies. For example, compare the Procedure Division statement MOVE EXAMPLE1 ... (21) with MOVE EXAMPLE2 ...(22) in the same contained program.
   • Statements in an indirectly contained program, provided that neither the indirectly containing program nor any program in between declare the same name as a GLOBAL name. For example, compare the Procedure Division statement MOVE EXAMPLE3 ...(29) with MOVE EXAMPLE2 ...(28) in the same contained program.

Example 6-9 Resolving References to Miscellaneous Names

IDENTIFICATION DIVISION. PROGRAM-ID. PROG-NAME-A. (1) DATA DIVISION. FILE SECTION. FD FILE-NAME ...(2) 01 RECORD-NAME ...(3) FD GLOBAL-FILE-NAME... IS GLOBAL... WORKING-STORAGE SECTION. 01 EXAMPLE1 ... IS GLOBAL... 01 EXAMPLE2 ... IS GLOBAL... 01 EXAMPLE3 ... IS GLOBAL... 01 SWITCH-STATUS. (4) 88 ON VALUE IS "1". 88 OFF VALUE IS "0". 01 DATA-NAME ...(5) PROCEDURE DIVISION. MOVE DATA-NAME TO ...(6) IF SWITCH-STATUS IS ON ...(7) MOVE RECORD-NAME TO ...(8) OPEN INPUT FILE-NAME ...(9) IDENTIFICATION DIVISION. PROGRAM-ID. PROG-NAME-B. (10) DATA DIVISION. FILE SECTION. FD FILE-NAME ...(11) 01 RECORD-NAME ...(12) WORKING-STORAGE SECTION. 01 SWITCH-STATUS. (13) 88 ON VALUE IS "1". 88 OFF VALUE IS "0". 01 DATA-NAME ...(14) 01 EXAMPLE2 ...(15) 01 EXAMPLE3 ... IS GLOBAL... 01 EXAMPLE4 ... IS GLOBAL... PROCEDURE DIVISION. MOVE DATA-NAME TO ...(16) IF SWITCH-STATUS IS ON ...(17) MOVE RECORD-NAME TO ...(18) OPEN INPUT FILE-NAME ...(19) OPEN OUTPUT GLOBAL-FILE-NAME. (20) MOVE EXAMPLE1 ...(21) MOVE EXAMPLE2 ...(22) MOVE EXAMPLE3 ...(23) MOVE EXAMPLE4 ...(24) IDENTIFICATION DIVISION. PROGRAM-ID. PROG-NAME-C. (25) WORKING-STORAGE SECTION. 01 EXAMPLE2 ...(26) 01 EXAMPLE4 ...(27) PROCEDURE DIVISION. OPEN OUTPUT GLOBAL-FILE-NAME. MOVE EXAMPLE1 ... MOVE EXAMPLE2 ...(28) MOVE EXAMPLE3 ...(29) MOVE EXAMPLE4 ... END PROGRAM PROG-NAME-C. END PROGRAM PROG-NAME-B. END PROGRAM PROG-NAME-A.

If a data item possesses either or both the EXTERNAL or GLOBAL attributes and includes a table defining an index-name, that index-name also possesses either or both attributes.

If the file associated with a segmented key possesses either or both the EXTERNAL or GLOBAL attributes, that segmented key also possesses either or both attributes.

6.2.7 External and Internal Data

External data is associated with a run unit. Any program in the run unit describing the external data can reference that data. (See the Section 5.3.21 clause in Chapter 5.) There is only one representation of an external data object.

Internal data is associated with a specific program.

External and internal data can have global names. (See the Section 5.3.25 clause in Chapter 5.)

6.3 Explicit and Implicit Specifications

The four types of explicit and implicit specifications follow:

• Procedure Division references
• Control transfers
• Attributes
• Scope terminators

A source program can refer to data items explicitly or implicitly in Procedure Division statements.

An explicit reference occurs when the name of the item is in a Procedure Division statement or copied into the Procedure Division by a COPY statement.

An implicit reference occurs under the following conditions:

• When a Procedure Division statement refers to an item whose name does not appear in the statement.
• During PERFORM statement execution. The PERFORM statement's control mechanism can initialize, change, and evaluate an index or data item referred to in the VARYING, AFTER, and UNTIL phrases. These implicit references occur only if the PERFORM statement execution involves the data item.

The mechanism that controls program flow implicitly transfers control from one statement to another in the order in which the statements appear in the source program. The transfer occurs in this sequence unless an explicit control transfer overrides it, or there is no next executable statement.

A program can contain both explicit and implicit changes to the control transfer mechanism.

Implicit control transfer can also occur when normal program flow changes without executing a procedure-branching statement. For example:

• A paragraph can execute under the control of another COBOL statement (such as PERFORM, USE, SORT, and MERGE). If the paragraph is the last in the controlling statement's range, an implied control transfer occurs from the last statement in the paragraph to the controlling statement's control mechanism.
  An implicit control transfer occurs between the control mechanism of a PERFORM statement that causes iterative execution and the first paragraph in its range. The transfer occurs for each iterative execution of the paragraph.
• When a SORT or MERGE statement executes, an implicit control transfer occurs to associated input or output procedures.
• When the execution of a statement causes the execution of a Declaratives Section, the control transfer is implicit. Another implicit control transfer occurs after execution of the Declaratives Section.
• If the Procedure Division does not have any Declaratives Sections, the program's first executable statement is the first executable statement in the Procedure Division. Otherwise, the program's first executable statement is the first executable statement after the declaratives part of the Procedure Division.

An explicit control transfer is a change to the implicit control transfer mechanism caused only by execution of either:

• A procedure-branching statement
• A conditional statement

The EXIT procedure-branching statement causes an explicit control transfer only when it has the PROGRAM phrase.

The Procedure Branching statement ALTER does not cause an explicit control transfer. However, it affects the explicit control transfer of the associated GO TO statement.

The term next executable statement refers to the next COBOL statement to which control transfers according to these rules and those associated with each language element.

There is no next executable statement when the program has no Procedure Division. This is also the case after:

• The last statement in a Declaratives Section, when the paragraph in which it appears is not executing under the control of another COBOL statement
• The last statement in a program, when the paragraph in which it appears is not executing under the control of another COBOL statement
• A STOP RUN or EXIT PROGRAM statement, when execution control transfers outside of the COBOL program containing the statement
• An END PROGRAM header

When there is no next executable statement and control does not transfer out of the program, program control flow is undefined. However, an EXIT PROGRAM statement implicitly executes when the program is under the control of a CALL statement.

6.3.3 Explicit and Implicit Attributes

An explicit attribute is an attribute the program explicitly specifies. If the program does not explicitly specify an attribute, the attribute assumes a default; it is then an implicit attribute.

For example, a program need not specify USAGE for a data item. If it does not, the data item's implicit usage is DISPLAY.