Named Data

A named data object, such as a variable, named constant, or function result, is given the properties of an intrinsic or user-defined data type, either implicitly (based on the first letter of the name) or through a type declaration statement. Additional information about a named data object, known as the data object's attributes, can also be specified, either in a type declaration statement or in separate statements specific to the attributes that apply.

Once a data object has a name, it can be accessed in its entirety by referring to that name. For some data objects, such as character strings, arrays, and derived types, portions of the data object can also be accessed directly. In addition, aliases for a data object or a portion of a data object, known as pointers, can be established and referred to.

In the absence of a type declaration statement, a named data object's type is determined by the first letter of its name. The letters I through N begin INTEGER data objects and the other letters begin REAL data objects. These implicit typing rules can be customized or disabled using the IMPLICIT statement. IMPLICIT NONE can be used to disable all implicit typing for a scoping unit.

A type declaration statement specifies the type, type parameters, and attributes of a named data object or function. A type declaration statement is available for each intrinsic type, INTEGER, REAL (and DOUBLE PRECISION), COMPLEX, LOGICAL, or CHARACTER, as well as for derived types (see Derived Types).

Attributes

Besides type and type parameters, a data object or function can have one or more of the following attributes, which can be specified in a type declaration statement or in a separate statement particular to the attribute:

DIMENSION -- the data object is an array (see DIMENSION Statement).

PARAMETER -- the data object is a named constant (see PARAMETER Statement).

POINTER -- the data object is to be used as an alias for another data object of the same type, kind, and rank (see POINTER Statement).

TARGET -- the data object that is to be aliased by a POINTER data object (see TARGET Statement).

EXTERNAL -- the name is that of an external procedure (see EXTERNAL Statement).

ALLOCATABLE -- the data object is an array that is not of fixed size, but is to have memory allocated for it as specified during execution of the program (see ALLOCATABLE Statement).

INTENT(IN) -- the dummy argument will not change in the subprogram

INTENT(OUT) -- the dummy argument is undefined on entry to the subprogram

INTENT(IN OUT) -- the dummy argument has an initial value on entry and may be redefined within the subprogram (see INTENT Statement).

PUBLIC -- the named data object or procedure in a MODULE program unit is accessible in a program unit that uses that module (see PUBLIC Statement).

PRIVATE -- the named data object or procedure in a MODULE program unit is accessible only in the current module (see PRIVATE Statement).

INTRINSIC -- the name is that of an intrinsic function (see INTRINSIC Statement).

OPTIONAL -- the dummy argument need not have a corresponding actual argument in a reference to the procedure in which the dummy argument appears (see OPTIONAL Statement).

SAVE -- the data object retains its value, association status, and allocation status after a RETURN or END statement (see SAVE Statement).

SEQUENCE -- the order of the component definitions in a derived-type definition is the storage sequence for objects of that type (see SEQUENCE Statement).

VOLATILE -- the data object may be referenced, become redefined or undefined by means not specified in the Fortran standard (see VOLATILE Statement).

DLL_EXPORT -- the name is an external procedure, or a common block name, that to be a DLL (see DLL_EXPORT Statement).

DLL_IMPORT -- the name is an external procedure, or a common block name, that uses a DLL (see DLL_IMPORT Statement).

ML_EXTERNAL -- the name is an external procedure, or a common block name, that is available for calling from a mixed language procedure (see ML_EXTERNAL Statement).