Digital PDFs

AA-DO34A-TE

August 1978

226 pages

Original

9.4MB

view download

OCR Version

8.8MB

view download

Document:	VAX-11 Fortran IV-Plus Language Reference Manual
Order Number:	AA-DO34A-TE
Revision:	0
Pages:	226
Original Filename:

OCR Text

August 1978

This document describes the FORTRAN language elements supported by
VAX-11 FORTRAN IV-PLUS. Itis intended to be used as a reference

manual in preparing FORTRAN source programs. It is not a tutorial
document, nor does it present information on the FORTRAN user’s interface to the VAX/VMS operating system,

VAX-11 FORTRAN IV-PLUS
Language Reference Manual
Order No. AA-DO34A-TE

SUPERSESSION/UPDATE INFORMATION:

This is a new document for this release.

OPERATING SYSTEM AND VERSION:

VAX/VMS V01

SOFTWARE VERSION:

VAX-11 FORTRAN IV-PLUS 01

To order additional copies of this document, contact the Software Distribution
Center, Digital Equipment Corporation, Maynard, Massachusetts 01754

digital equipment corporation - maynard, massachusetts

First Printing,

August

1978

The information in this document is subject to change without notice
and should not be construed as a commitment by Digital Equipment
Corporation.
Digital Equipment Corporation assumes no responsibility
for any errors that may appear in this document.
The software described in this document is furnished under a license
and may only be used or copied in accordance with the terms of such
license.

’

No responsibility is assumed for the use or reliability of software on
equipment that is not supplied by DIGITAL or its affiliated companies.

1978 by Digital Equipment Corporation

The postage-prepaid READER'S COMMENTS form on the last page of this
document requests the user's critical evaluation to assist us in preparing future documentation.

The following are trademarks of Digital Equipment Corporation:

DIGITAL

DECsystem-10

DEC

DECtape

MASSBUS
OMNIBUS

PDP

DIBOL

0s/8

DECUS

EDUSYSTEM

PHA

UNIBUS
COMPUTER LABS

FLIP CHIP
FOCAL

RSTS
RSX

COMTEX
DDT
DECCOMM
ASSIST-11

INDAC
LAB-8
DECSYSTEM-20
RTS-8

TYPESET~-8
TYPESET-11
TMS-11
ITPS-10

VAX

VMS

SBI

DECnet

IAS

CONTENTS

Page

PREFACE
CHAPTER

ix
1

INTRODUCTION

1.1
1.2
1.2.1
1.2.2
1.2.3

LANGUAGE OVERVIEW
ELEMENTS OF FORTRAN PROGRAMS
Statements
Comments
FORTRAN Character Set

1-1
1-3
1-3
1-3
1-4

1.3
1.3.1
1.3.2

FORMATTING A FORTRAN LINE
Character-per-Column Formatting
Tab-Character Formatting

1-5
1-5
1-6

Statement Label Field
Comment Indicator
Debugging Statement Indicator
Continuation Field
Statement Field
Sequence Number Field

1-7
1-7
1-7
1-7
1-8
1-8

1.3.3
1.3.3.1
1.3.3.2
1.3.4
1.3.5
1.3.6

CHAPTER

VAX-1ll

FORTRAN

IV-PLUS

1.4

PROGRAM UNIT

1.5

INCLUDE

STATEMENT

1-9

FORTRAN

STATEMENT COMPONENTS

2-1

2.1
2.2

SYMBOLIC NAMES
DATA TYPES

2-2
2-3

2.3

CONSTANTS

2-4

2.3.1
2.3.2
2.3.3
2.3.4
2.3.5
2.3.6
2.3.7
2.3.8
2.4
2.4.1
2.4.2

Integer Constants
Real Constants
Double Precision Constants
Complex Constants
Octal and Hexadecimal Constants
Logical Constants
Character Constants
Hollerith Constants
VARIABLES
Data Type Specification
Data Type by Implication

2-5
2-6
2-7
2-7
2-8
2-10
2-10
2-11
2-12
2-12
2-13

2.5

ARRAYS

2-13

2.5.1

Array

2.5.2

Subscripts

2.5.3

Array

2.5.4

Data

2.5.5
2.5.6
2.6
2.7
2.7.1

STRUCTURE

1-1

Declarators

2-14

2-15

Storage

Type

1-8

2-15

an Array

2-16

Array References Without Subscripts
Adjustable Arrays
CHARACTER SUBSTRINGS
EXPRESSIONS
Arithmetic Expressions

2-16
2-17
2-17
2-18
2-18

2.7.1.1

Use of Parentheses

2-20

2.7.1.2
2.7.2
2.7.3
2.7.4

Data Type of an Arithmetic Expression
Character Expressions
Relational Expressions
Logical Expressions

2-21
2-22
2-23
2-24

iii

CONTENTS

(Cont.)

Page

CHAPTER

ASSIGNMENT

3.1

ARITHMETIC ASSIGNMENT

3.2

LOGICAL ASSIGNMENT

3.3

CHARACTER ASSIGNMENT

3.4

ASSIGN

CONTROL STATEMENTS

4-1

4.1
4.1.1
4,1.2

4-2
4-2
4-2

4.1.3
4.2

TO STATEMENTS
Unconditional GO TO Statement
Computed GO TO Statement
Assigned GO TO Statement
STATEMENTS

Block

4.2.3.3

Nested

IF Statement
Statement

Block

4-5
4-8

4-8

Constructs

STATEMENT

4.3.3

Control

4,.3.4

Extended

4-13

Loops

Transfers

4-10
4-12

DO Iteration Control
DO

4-3
4-4
4-5

IF Examples

Nested

3-3
3-4

4-4

4,.2.3.2

4,3.2

3-1

3-5

Block IF Statements
Statement Blocks

4-14
in

Loops

4-15

Range

4-15

STATEMENT

4-17

4.4

CONTINUE

4.5

CALL

4.6

RETURN

4.7

PAUSE

4.8

STOP

4.9

END STATEMENT

SPECIFICATION

5.1
5.2

TYPE

5.2.1
5.2.2

CHAPTER

STATEMENT

4.2.3
4.2.3.1

STATEMENT
STATEMENT

Arithmetic
Logical IF

4,.3.1

3-1

STATEMENT

4,2.1
4.2.2

4.3

CHAPTER

STATEMENTS

STATEMENT

4-18

STATEMENT

4-19

STATEMENT

4-21

STATEMENT

4-22

4-23
STATEMENTS

5-1

IMPLICIT STATEMENT
DECLARATION STATEMENTS

Numeric Type Declaration Statements
Character Type Declaration Statements

5-2
5-3

5-3
5-4

5.3
5.4

DIMENSION STATEMENT
COMMON STATEMENT

5-5
5-6

5.5
5.5.1
5.5.2

EQUIVALENCE STATEMENT
Making Arrays Equivalent
Making Substrings Equivalent

5-8
5-8
5-10

5.5.3
5.6
5.7
5.8
5.9

EQUIVALENCE and COMMON
EXTERNAL STATEMENT
DATA STATEMENT
PARAMETER STATEMENT
PROGRAM STATEMENT

5-13
5-14
5-16
5-18
5~-19

5.10

BLOCK

SUBPROGRAMS

6.1
6.1.1
,

SUBPROGRAM ARGUMENTS
Actual Argument and
Association

6.1.1.1
6.1.1.2

Interaction

DATA STATEMENT

5-20

6-1
6~1
Dummy Argument

Adjustable Arrays
Passed Length Character Arguments

6-1
6-2
6~3

CONTENTS

(Cont.)

FUNCTIONS

Function

References

Generic Function References
Processor-Defined and Generic

W N

00 00 0o

[00]

N =

NN
N NSNS NN NN
e ©®
&
o
L]
e
&
e
&
s
@
L]
L]
NOOAONOUTOTUTd
R WWWNNDN
¢
o
L]
[]
L)
L)
L[]
.
L]

e e

ALTUIUT
D W N H

[

'_.l

I/0

STATEMENT COMPONENTS
Logical Unit Numbers
Direct Access Record Numbers
Format Specifiers
Input/Output Records
Input/Output Lists
Simple Lists

INPUT/OUTPUT STATEMENTS

CHAPTER

Character Library Functions

Usage

CHAPTER

Implied DO Lists
Transferring Control on End-of-File or
Error Conditions
FORMATTED SEQUENTIAL INPUT/OUTPUT
Formatted Sequential Input Statements
Formatted Sequential Output Statements
LIST-DIRECTED SEQUENTIAL INPUT/OUTPUT

List-Directed Input Statements
List~Directed Output Statements
UNFORMATTED SEQUENTIAL INPUT/OUTPUT
Unformatted

Sequential

Input

Unformatted

Sequential

Output

Statement

Statement
FORMATTED DIRECT ACCESS INPUT/OUTPUT
Formatted Direct Access Input Statement
Formatted Direct Access Output Statement
UNFORMATTED DIRECT ACCESS INPUT/OUTPUT

Unformatted Direct Access
Unformatted Direct Access
ENCODE

AND

DECODE

FORMAT

STATEMENTS

FIELD

DESCRIPTORS

STATEMENTS

I
O

Field Descriptor
Field Descriptor

Field

Descriptor

>
N O o o

6~17

Function

B_WWwwdhDNDNDN

LIBRARY

Processor-Defined

Input Statement
Output Statement

HHEONNSO

FORTRAN

6-15
O

Statement

ENTRY in Function Subprograms
ENTRY in Subroutine Subprograms

Functions
Character Functions
Function Reference
Subroutine Subprograms

Functions

1
N
I T
U
i
s
o HEHEOONNOUTOO

Statement

Subprograms

1
M

SUBPROGRAMS

A OY
O

Functions

DN
=0

w N

Arguments

o)W )]

Return

Numeric

ENTRY
=

$LOC Built-In Function

Function

Hollerith Constants

Built-In Functions
Argument List Built-In

Arithmetic

Alternate

USER-WRITTEN

WWWWRNRNNNNNNNNNN
.
.
[}
[3
°
whH
R B W NN NN
NN

AN
AN
O O
.
*
L[]
.
L]

and

Arguments

Character
Actual

NNNNNNNN

o)l

[

i.—l

Page

CONTENTS

(Cont.)

AUXILIARY INPUT/OUTPUT STATEMENTS

9-1

9.1

OPEN

Keyword

BLOCKSIZE

BUFFERCOUNT

Keyword

CARRIAGECONTROL Keyword
DISPOSE Keyword
ERR Keyword

INITIALSIZE Keyword
MAXREC Keyword
NAME Keyword

9.1.13
9.1.14
9.1.15
9.1.16
9.1.17
9.1.18
9.1.19
9.1.20
9.1.21
9.2

NOSPANBLOCKS Keyword
ORGANIZATION Keyword
READONLY Keyword
RECORDSIZE Keyword
RECORDTYPE Keyword
SHARED Keyword
TYPE Keyword
UNIT Keyword
USEROPEN Keyword
CLOSE STATEMENT

FORM Keyword

]

[]

[

EXTENDSIZE Keyword

L[]

9.1
9.1
9.1
9.1
9.1
9.1.10
9.1.11
9.1.12

9.3
9.4
9.5
9.6
9.7

REWIND STATEMENT
BACKSPACE STATEMENT
FIND STATEMENT
ENDFILE STATEMENT
DEFINE FILE STATEMENT
vi

8-22
8-23
8-23

8-24
8-24

FHWOWVWVOOOLIIIJN
0N

L[]
.

ASSOCIATEVARIABLE

9.1
9.1
9.1

STATEMENT

ACCESS

NNFRHHREHRERFRHEBOO0ONO0S
WN

9.1

8.1.4
8.1.5
8.1.6
8.1.7
8.1.8
8.1.9
8.1.10
8.1.10.1
8.1.11
8.1.12
8.1.13
8.1.14
8.1.15
8.1.16
8.1.17
8.1.18
8.1.19
8.1.20
8.2
8.3
8.4
8.5
8.6

CHAPTER

8-5
8-6
8-7
8-8
8-9
8-10
8-11
8-11
8-12
8-12
8-13
8-13
8-14
8-14
8-14
8-16
8-17
8-18
8-18
8-19
8-20
8-21

8.7
8.7.1
8.7.2
8.7.3

F Field Descriptor
E Field Descriptor
D Field Descriptor
G Field Descriptor
L Field Descriptor
A Field Descriptor
H Field Descriptor
Character Constants
X Field Descriptor
T Field Descriptor
Q Field Descriptor
Dollar Sign Descriptor
Colon Descriptor
Complex Data Editing
Scale Factor
Repeat Counts and Group Repeat Counts
Variable Format Expressions
Default Field Descriptors
CARRIAGE CONTROL
FORMAT SPECIFICATION SEPARATORS
EXTERNAL FIELD SEPARATORS
RUN--TIME FORMAT
FORMAT CONTROL INTERACTION WITH INPUT/OUTPUT
LISTS
SUMMARY OF RULES FOR FORMAT STATEMENTS
General Rules
Input Rules
Output Rules

PPOVVVOOVVOOVYOY

Page

-10
-10
-10
9-10
9-11
9-12
9-12
9-12
9-12
9-13

9-14
9-15
9-16
9-17
9-18

CONTENTS

(Cont.)

Page

APPENDIX A
~

A.l
A.2
A.3

APPENDIX B

CHARACTER SETS

A-1

FORTRAN CHARACTER SET
CHARACTER SET
RADIX~-50 CONSTANTS AND

A-2

FORTRAN

B.1
B.2
B.3

A-1

ASCII

LANGUAGE

CHARACTER

SET

A-2

SUMMARY

EXPRESSION OPERATORS

B-1

STATEMENTS

B-1
B-2

LIBRARY FUNCTIONS

Bfl7

INDEX

Index-1

FIGURES

FIGURE

1-1

1-2

1-3
2-1

4-1
4-2
4-3

FORTRAN

Coding Form
Formatting Example
Required Order of Statements
Array Storage

1-5

Line

Examples of Block
Nested DO Loops

and

ILines

1-6
1-8
2-16

IF Constructs

4-7

5-1
5-2

Control Transfers and Extended Range
Equivalence of Array Storage
Equivalence of Arrays with Nonunity Lower

5-3
5-4
6-1
6-2
8-1

Equivalence of Substrings
Equivalence of Character Arrays
Multiple Functions in a Function Subprogram
Multiple Function Name Usage
Variable Format Expression Example

Bounds

4-14
4-16
5-9
5-10

5-11
5-12
6-15
6-20
8-17

TABLES

2-1

2-2

2-3
3-1
6-1

Entities

Identified

by Symbolic Names
Requirements
Allowed Combinations for Exponentiation
(
Conversion Rules for Assignment Statements

Data

Type

Storage

Argument List Built-In Functions
Defaults

6-2
6-3

Default
Carriage
Summary

Subprograms
Summary

OPEN

ASCII

Field

Descriptor Values
Control Characters

FORMAT

Codes

Statement Keyword

Character

Generic

2-4

2-19
3-2
6-6
6-7
6-18

7-14
8
8
8
8-25

Values

Set

and Processor-Defined

vii

2-3

and

List-Directed Output Formats
of Data Magnitude on G Format
Conversions

Effect

wns?cnaam

7-1

Types of User-Written
Generic Function Name

8-1
s
W N

TABLE

Functions

PREFACE

MANUAL

OBJECTIVES

This manual describes the elements of
VAX-11
manual is designed for reference, rather than

The user interface to VAX-11

companion

manual

this

FORTRAN

IV-PLUS

document,

Guide.

FORTRAN
as

the

1is

VAX-11

IV-PLUS.
The
tutorial document.

described

FORTRAN

IV-PLUS

the

User's

INTENDED AUDIENCE

This

manual

intended

readers

should

have

order

derive

maximum

for use
basic

reference

understanding

benefit

from

the

the
VAX/VMS
operating
system
is
information concerning VAX/VMS refer
under

"ASSOCIATED

STRUCTURE

This

manual
®

THIS

document.
the

manual.

helpful,

the

DOCUMENTS."

Therefore,

FORTRAN
Some

language

knowledge

in
of

but not necessary.
For
documents
listed
below

DOCUMENT

contains

nine

chapters

and

Chapter

1 consists of general
introduces
basic
facts

and

two

appendixes.

information
needed

concerning
FORTRAN,
prior to writing FORTRAN

programs.

Chapter
as

Chapter

define
e

describes

Chapter

control

from

Chapter

used

deals

define

data

the

FORTRAN

statements,

the

which

are

program

used

discusses

covers

Chapter

describes

formatted

subprograms;

VAX-11
the

I/O

FORTRAN
topic
the

both

user-written

IV-PLUS.

FORMAT

statements.

FORTRAN

transfer

another.

describes

with

such

etc.

program.

control

point

Chapter

with

assignment

with

one

components
variables,

specification statements, which
are
characteristics of symbols used in the program,
type, array dimensions, etc.

Chapter

the

constants,

values

supplied
e

describes

symbols,

and

used

such

those

input/output.

statements

used

conjunction

Chapter 9 contains information on
CLOSE,

such as OPEN,

auxiliary

and DEFINE FILE.

I/0

statements,

Appendix A summarizes the character sets supported
FORTRAN

VAX-11

IV-PLUS.

Appendix B summarizes the language elements of VAX-11

FORTRAN

IV-PLUS.

ASSOCIATED DOCUMENTS

FORTRAN

IV-PLUS

including
complete 1list of all VAX-1ll documents,
For
a
descriptions of each, see the VAX-1ll Information Directory.

brief

The following documents are of

interest

VAX-11

programmers:

VAX/VMS Primer

VAX-11 FORTRAN IV-PLUS User's Guide
VAX/VMS Command Language User's Guide

CONVENTIONS USED IN THIS DOCUMENT

The following syntactic conventions are used in this manual:
Uppercase words and letters used in examples indicate that you
should type the word or letter as shown

Lowercase words and letters used in format examples indicate
that you are to substitute a word or value of your choice
Brackets

indicate optional elements

([])

Braces ({}) are used to enclose lists from which one element

item(s)

can

denote

special

chosen

Ellipses

indicate that

(...)

repeated one or more times

In addition, the

nonprinting

following

characters:

Tab character

notations

the

are

preceding

used

TAB

Space character A
or

(in terminal dialog examples)

CHAPTER

INTRODUCTION

1.1

LANGUAGE

TO VAX-11

FORTRAN

IV-PLUS

OVERVIEW

VAX-11 FORTRAN IV—PLUSl is based on American National
FORTRAN

X3.9-1966,

and

includes

the

following

type

provides

Standard

enhancements

(ANS)

ANS

FORTRAN:

The

CHARACTER data

character
functions,
and

means

manipulating

data.
Character
constants,
variables,
and expressions using
the
concatenation

substring

references

are

arrays,
operator

provided.

Block IF statements provide a way to conditionally
execute
a
block
of
statements.
The
block IF statements are IF THEN,
ELSE IF THEN, ELSE, and END IF.

Any
If

arithmetic expression can be used as an
the
expression
1is not of type integer

integer
®

type.

Mixed-mode

including
e

The

array
subscript.
it is converted to

expressions
complex.

following

data

can

types

contain

have

been

elements

any

data

type,

added:

LOGICAL*1
LOGICAL*2
INTEGER*2
CHARACTER

The

IMPLICIT

symbolic

statement

redefines

the

implied

data

type

names.

e The following input/output (4}0) statements have been added:
ACCEPT

TYPE

Device-oriented

I/0O

1
the

READ
WRITE
FIND

(u'r)

READ
WRITE

(u'r,fmt)
(u'r,fmt)

(u'r)
(u'r)

Unformatted

Formatted

direct-access

VvAX-11 FORTRAN IV-PLUS is referred to simply as FORTRAN
rest

this

manual.

I1/0

I/0

throughout

INTRODUCTION TO VAX-11

OPEN

FORTRAN

IV-PLUS

CLOSE
DEFINE FILE

File control and attribute
specification

ENCODE
DECODE

Formatted data conversion
in memory

The specifications END=s and/or ERR=s can be included
in
any
READ
or
WRITE statement to transfer control to the specified
statement when an end-of-file or error condition occurs.

List-directed I/0 can be used to perform formatted I/O without
a

format

specification.

I/0O

1lists

PARAMETER statements can be used to assign symbolic

names

Constants and expressions are permitted in the

WRITE,

TYPE,

constant

and

statements.

values.

Generic-function selection by argument data type
for many FORTRAN-supplied

functions.

provided

A DO statement control
variable
can
be
a
real
or
double
precision
variable.
Any arithmetic expression can be used as
the initial value, increment, or limit
parameter
1in
the
DO
as the control parameter in the computed GO TO
or
statement,
statement.

The DO statement

increment

parameter

you

optionally

can

have

negative

value.

For

readability,

statements.

For

I=1,10

can

example:

Lower bounds for array dimensions
array

declarators.

can

use

commas

specified

1in

all

ENTRY statements
can
be
used
in
SUBROUTINE
and
subprograms
to
define
multiple , entry
points
1in
program

FUNCTION
a single

unit.

A PROGRAM statement can be used

in a main program.

The INCLUDE statement incorporates FORTRAN source text from
file

separate

You can
FORTRAN

into

a FORTRAN program.

include an explanatory comment on the same line as any
statement.
Begin
comments
with exclamation points

(1.

You can include debugging statements in a program
by
placing
the
letter D in column 1.
These statements are compiled only
when you specify the associated
compiler
command
qualifier;
otherwise,

they are

treated

as comments.

The statement label list in an assigned
optional.

statement
8,

INTRODUCTION

Octal and hexadecimal
numeric constants.

Symbolic
letters,

TO VAX-1ll

constants

FORTRAN

can be

IV-PLUS

used

names can be up to 15 characters
digits,
dollar
signs ($), and

place

any

long and consist
of
underline characters

().
VAX-11 FORTRAN IV-PLUS is also a compatible superset of PDP-11 FORTRAN
IV-PLUS.
This
means
you can compile existing PDP-11 FORTRAN source
programs, as well as new programs that incorporate features
available
in VAX-11l FORTRAN IV-PLUS.

1.2

ELEMENTS

FORTRAN

PROGRAMS

FORTRAN

programs consist of FORTRAN statements and optional
comments.
The
statements are organized into program units.
A program unit is a
sequence of statements
that
define
a
computing
procedure
and
is
terminated
by
an END statement.
A program unit can be either a main
program or a subprogram.
An executable program consists of
one
main
program and, optionally, one or more subprograms.

1.2.1

Statements

Statements are
grouped
1into
two
general
classes:
executable
and
nonexecutable.
Executable
statements
describe
the
action
of the
program.
Nonexecutable
statements
describe
data
arrangement
and
characteristics, and provide editing and data-conversion information.

Statements are divided into physical sections called lines.
A line is
a string of up to 80 characters.
If a statement is too long to fit on
one line, you can continue it on one or more additional lines,
called
continuation
lines.
A
continuation
1line
1is
identified
by
a
continuation character in the sixth column of that line.
(For further
information on continuation characters, see Section 1.3.4.)
You can identify a statement with a
statement
1label
so
that
other
statements
can refer to it, either for the information it contains or
to transfer control to it.
A statement label takes
the
form
of
an
integer
number
1in
the
first
five columns of a statement's initial
line.

1.2.2

Comments

not

affect

program processing

any

way.

They

are

merely

a
documentation
aid
to
the programmer.
You can use them freely to
describe the actions of the program, to identify program sections
and
processes,
and
to provide greater ease in reading the source program
listing.
The letter C in the first column of a source line identifies
that
line
as
a
comment.
In addition, if you place an ex¢lamation
point (!) in column 1 or in the statement portion of
a
source
1line,
the rest of that line is treated as a comment.

INTRODUCTION TO VAX-11

1.2.3

The

FORTRAN Character

FORTRAN character

FORTRAN

IV-PLUS

Set

set consists of:

All uppercase and lowercase letters

(A through Z,

through

z)
2.

The

numerals

through

The special characters listed below
Character

Name

A or (me)

Space or

tab

Equal

Plus sign

Minus sign

Asterisk

Slash

(

Left parenthesis

)

Right parenthesis

’

Comma

Period

Apostrophe

Quotation mark

Dollar

sign

Underline
Exclamation point
:

Colon

Left

Right angle bracket

Percent

Amper sand

angle bracket

sign

Other printable ASCII characters can appear
in
a
only
as part of a character or Hollerith constant
a list of printable characters).
in

FORTRAN
statement
(see Appendix A for

Any printable character

can

appear

comment.

Except

in character and Hollerith constants,

distinction between uppercase and

the

compiler

lowercase letters.

makes

INTRODUCTION

1.3

FORMATTING A FORTRAN

Each

FORTRAN

line

has

Statement

label

Continuation

®©

Statement

field

Sequence

number

fields,

indicator

form,
use

can

field.

entering

tab-character

terminal

1.3.1

or
using

lines

shown

Figure

text

1-1,

labels,

sequence

1.3.3

from

terminal
when

using

you

are

text

editor.

entering

lines

Formatting

FORTRAN

Each

through

only

editor.

line

continuation

numbers.

Sections

a
FORTRAN
line:
1) on
a
using the tab character.
You can
when
punching
cards,
using
a

formatting

Character-per-Column

statement

follows:

field

to
format
basis or 2) by
character-per-column
formatting

You

IV-PLUS

field

There
are
two
ways
character-per-column
use
coding

FORTRAN

LINE

four

TO VAX-11

column

1.3.6

1is

divided

indicators,
represents

describe

the

CODER

CODING FORM

PROBLEM

fields

for

text,

and

single

use of

FORTRAN

1into

statement

each

character.

field.

DATE

PAGE

C Convmer

STATEMENT
N UM

FORTRAN STATEMENT

[

IDENTIFICATION

1234

506§7891001213141516171819202122232425262728293031323334
3536373839404 424344454647 484950 515253 54555657 5859 6061626364 6556676869707 72|7376 757677
7879 80

THI1S, PROGRAM CALCULATES PRIME NUMBERS FROM 11 TO 50
(210 TR RS NN P IR
J.=1

.. @

@ @ ..,

ettt

A N
T

.......

+———+

—t

ettt

A R TN

A= 1/A

L=1/4

D
———

B=A-L

BF (B

LF (J. LT SGRT, ((FLOAT, (l)))

CONTINUE

—

1,05

FORMAT

PRlME")

MND

TYPE

5,
105

yes

—

GQ TO. ‘4...........................

—

bttt

AAAAAAAAAAAAAAAAAAAAAAAAAA

(I4 'IS

e
L

10,

LANRLANAGHE

VST

SIS

SR
BN

——s

PN
SN
SN
o

WU W

Attt

oottt

bbbt

Attt

4ttt

A+

AT W

+7T

'
T

bbbt

-ttt

gy T e W i

1234 35016[789101012131415161718192021222324252627282970 31323334 133637307940 41 4243444540 47484950 315233 343356 57 39390061 626164 836687680970 1N 71 717473
7877 787980

PG-3

DIGITAL

EQUIPMENT

Figure

CORPORATION

1-1

MAYNARD,

FORTRAN

MASSACHUSETTS

Coding

Form

INTRODUCTION TO VAX-11] FORTRAN IV-PLUS

the

To enter an item in a field, enter it in the column(s) in
form,

<coding

as listed below:

Field

Column (s)

Statement label

1 through 5

Continuation indicator

Statement

7 through 72

Sequence number

73 through 80

Tab-Character Formatting

1.3.2

statement

the

‘You can use tab-character formatting to specify

field, the continuation indicator field, and the statement field.

label

You

cannot specify a sequence number field with tab-character formatting.
Figure 1-2 illustrates FORTRAN lines with tab-character formatting and
the equivalent lines with character-per-column formatting.
Character-per-Column Format

Format Using TAB Character

112

51617

10|11

C @a® FIRST VALUE

FIIIRISIT

10@B | = J + 5xK +

110

Figure 1-2

before the first tab character.
5

characters.

After you type the first

tab

51+

I {+1]2

Line Formatting Example.

The statement label field consists of the
than

VAL JUIE

I {V[AlL

@B IVAL = 1+2

12 13 14 15|16 17 18 19 20

L]|+IM

1 L+M

characters

that

you

type

either

the

The statement label field cannot have

character,

you

can

continuation indicator field or the statement field.

type

the
To enter the continuation indicator field, type any digit after
the
field,
indicator
continuation
the
enter
you
If
first tab.
statement field consists of all the characters after the digit to the
end of

the

line.

To enter the statement field without a continuation indicatorK field,
Note that no
the first tab.
type the statement immediately after
FORTRAN statement starts with a digit.
1-6

INTRODUCTION

TO VAX-11l

FORTRAN

IV-PLUS

Many

text editors and terminals advance the terminal print carriage to
a
predefined print position when you type the TAB key.
However, this
action is not related to the FORTRAN compiler's interpretation of
the
tab character described above.

You

can use the space character to improve the legibility of a FORTRAN
statement.
The
compiler
ignores
all
spaces
1in a statement field
except those within a character or Hollerith constant.
For
example,

and

GOTO

statement

are

equivalent.

field

the

the compiler produces,
printed

1.3.3

the

Statement

same

The

compiler

space.

1In

treats

the

the tab causes the character

tab

Label

stop

(located

columns

the

tab

source

character

listing

that

that follows to be
17,

25,

33,

etc.).

Field

A statement label or statement number consists of one to five
decimal
digits
in
the
statement
1label field of a statement's initial line.
Spaces and leading zeros are ignored.
An all-zero statement label
is
illegal.
Any

statement

two

statements

that

another statement
within a program unit

refers to must have a label.
can have the same label.

You can use two special indicators -- the comment
debugging
statement
indicator
-in
the first
field.
These indicators are described below.
The

statement

1.3.3.1
indicate
the

label

field

continuation

Comment Indicator - You can
that the line is a comment.

source

program

listing,

then

line

indicator
column of

must

and
the
the label

blank.

use the letter C in column
1
The compiler prints that line

ignores

the

to
in

line.

1.3.3.2
Debugging Statement Indicator - You can use the letter
D
in
column
1
to designate debugging statements.
The initial line of the
debugging statement can contain a statement
label
in
the
remaining
columns
of
the
1label
onto more than one line,
column
The
be

compiler
compiled

compiler

statements

not

well

1.3.4

are

continuation

the debugging statement
a comment, depending on
qualifier.

compiled

LINES,

Continuation

indicator.

treats
or as

command

specify

field.
1If a debugging statement is continued
every continuation line must contain a
D
in

you

either as source
text
to
the setting of the D_LINES

specify

as a part of the

debugging

statements

LINES,

source program;

are

treated

debugging

if you do

comments.

Field

A continuation indicator is any character, except zero
or
space,
in
column
6 of a FORTRAN line or any digit, except zero, after the first
tab.
A statement can be divided into distinct
lines
at
any
point.
The compiler considers the characters after the continuation character
as the characters following the last character of the
previous
1line,

zero,

break occurred

then

FORTRAN

the

compiler

statement.

that point.

considers

the

a continuation

1line

1initial

indicator
1line

is
a

INTRODUCTION TO VAX-11l FORTRAN IV-PLUS

Comment lines can occur between a
Comment lines cannot be continued.
statement's initial line and its continuation line(s), or between
successive continuation lines.

1.3.5

Statement Field

placed

The text of a FORTRAN statement is
2 A

aWa

OIRD

Q-Nno

the

A N d

statement

E A

field.
-‘

character and Hollerith constants), you can space the text in any

way

desired for maximum legibility.

NOTE

1line

position 72,

1
72 is ignored
printed.
70

1.3.6

extends

character

beyond

the text following position
and

no warning message

1
1s

Sequence Number Field

A sequence number or other identifying information can appear in
columns 73 through 80 of any line in a FORTRAN program. The compiler
ignores the characters in this field.

1.4

PROGRAM UNIT STRUCTURE

Figure 1-3 shows the order of statements in a FORTRAN program unit.
In this figure, vertical lines separate statement types that can be
interspersed. For example, DATA statements can be interspersed with
executable statements. Horizontal lines indicate statement types that
For example, type declaration statements
cannot be interspersed.
cannot be interspersed with executable statements.

PROGRAM, FUNCTION, SUBROUTINE, or BLOCK DATA Statements
IMPLICIT
Statements

.
PARAMETER

Statements

Other

Comment
Li

a'r':ss

INCLUDE

Statements

Specification
Statements

FORMAT

and

ENTRY

Statements

DATA

Statements

Statement Function

Definitions

Executable
Statements

END Line

Figure 1-3

Required Order of Statements and Lines
1-8

INTRODUCTION TO VAX-11 FORTRAN IV-PLUS
1.5

INCLUDE STATEMENT

The INCLUDE statement specifies that the contents of a designated file
are to be incorporated in the FORTRAN compilation directly following
the INCLUDE statement. The INCLUDE statement is described in this
has
chapter rather than with the other FORTRAN statements because it read
no effect on program execution, except to direct the compiler to
FORTRAN statements from a file.

The INCLUDE statement has the form:

INCLUDE 'file specification[/[NO]JLIST]'
file specification

Is a character constant that specifies the file to be included in
This file specification must be acceptable to
the compilation.
(See the VAX-11 FORTRAN IV-PLUS User's
the operating system.

Guide for the form of a file specification.)

The /LIST qualifier indicates that the statements in the specified
file are to be listed in the compilation source listing. An asterisk
(*) precedes each statement listed. The /NOLIST gqualifier indicates
that the included statements are not to be listed in the compilation
source listing. The default is /LIST; that is, the compiler assumes
/LIST if you do not specify a qualifier.

When the compiler encounters an INCLUDE statement, it stops reading
statements from the current file and reads the statements in the
included file. When it reaches the end of the included file, the
compiler resumes compilation with the next statement after the INCLUDE
statement.

An INCLUDE statement can be contained in an included file.
An included file cannot begin with a continuation line. Each
statement must be completely contained within a single file.

FORTRAN

The INCLUDE statement can appear anywhere that a comment line can
appear, as shown in Figure 1-3. Any FORTRAN statement can appear in
an included file. However, the included statements, when combined
with the other statements in the compilation, must satisfy the
statement ordering restrictions described in Section 1.4.

In the following example, the file COMMON.FOR defines the size of

blank COMMON block and the size of the arrays X, Y, and Z.
Main Program File

File COMMON.FOR

INCLUDE 'COMMON.FOR'
DIMENSION Z (M)

PARAMETER M = 100
,Y (M)
COMMON X (M)

CALL

CUBE

1I=1,M

Z(I)

X(I)+SQRT(Y(I))

SUBROUTINE

INCLUDE

DO 10,

X(I)

I=1,M

= Y(I)**3

RETURN
END

CUBE

'COMMON.FOR'

the

CHAPTER 2

FORTRAN STATEMENT COMPONENTS

The basic components of FORTRAN statements are:

Constants —- fixed, self-describing values.

Variables -- symbolic names that represent stored values.
contiguously and
Arrays -- groups of values that are storedtively
. Individual
collec
or
can be referred to individually
values are called array elements.

ts, or
Expressions -- single constants, variables, array elemen
or, combinations of these components
function references;
that specify
plus certain other elements, called operators,these
components
of
values
the
on
med
perfor
be
to
ations
comput
to obtain a single result.

optionally followed
Function references —-- names of functions,
nce performs the
refere
on
functi
A
nts.
argume
by 1lists of
The
ition.
defin
ion
funct
computation indicated by the
resulting value is used in place of the function reference.

ic name
have symbolic names. A symbol
variables, arrays, and functionsidenti
fy entities in the program.
is a string of characters that

Constants, variables, arrays, expressions, and functions can have

the

following data types:
Logical
Integer

Real

Double precision
Complex
Character

components of FORTRAN, with
The following sections detail the basic which
are described in Chapter
the exception of function references,
6.

FORTRAN

2.1

SYMBOLIC

Symbolic

unit.
A

symbolic

(S$)

and

are

used

entities

name

underline

symbolic

name

maximum of
Examples

must

are

identify

listed

string

(_)
be

valid

and

symbolic

names

with

(begins

B.4

(contains
other

SFREQ

for

wuse

conflicts,
that

Symbolic
cannot

you

the

same
consisting
following

must

the

unique

within

symbolic

unit.

Function

Subroutine

Common

Main

Block

Function

Subroutine

That

is,

the

name

any

dollar

sign

character
can

contain

a
a

unique

numeral)
or

dollar

the

sign

components.

symbolic

program

within

character

§)

identify

Furthermore,
in
more program units,

unit.
two

($)-

are

avoid

names

vyour

That

1s,

you

entities

executable

the symbolic names
entire program:

program
of

the

cannot

use

functions

subprograms
subprograms

blocks

programs
data

your

subprograms

entries
entries

program

symbolic

program,

unit.
entity with
Sections
2.4.1

contains
name

even

"yes" under
and
2.4.2

name.

Within a
subprogram,
arguments.
A
dummy
element,

the

name

software

define
sign.

of
two
or
entities must

BTU

not

dollar

Processor-defined

Each

program

are:

containing

same

program

names

FORTRAN

first

special

with

DIGITAL-supplied

should

contain

names
use

than

(begins

symbolic

program

and

The

symbolic

convention,

block

digits,

NUMBER

reserved

name

The

Invalid

FIND_IT
R

within

2-1.

characters.

letter.

Table

letters,

special

invalid

entities

characters.

Valid

-«

COMPONENTS

NAMES

names

These

STATEMENT

constant,

any

the

function
other

name

"Typed"
discuss

named

BTU,

subprogram,

appears

you

entry,

different

in Table 2-1
has
how to specify the

common

program

a
data
type.
data type of a

you
can
also
use
symbolic
names
as
argument can represent a variable, array,
expression, or subprogram.

dummy
array

FORTRAN STATEMENT COMPONENTS

2-1

Table

Entities Identified by Symbolic Names
Typed

Entity
Variables

yes

Arrays

yes

Arithmetic statement functions
Processor-defined functions

yes
yes

Function subprograms
Subroutine subprograms
Common blocks
Main programs
Block data subprograms

yes
no
no
no
no

yes
no

Function entries
Subroutine entries
Parameter

2.2

yes

constants

DATA TYPES

The
types.
several
of
Each basic component represents data of one
data type of a component can be inherent in its construction, implied
in
available
The data types
by convention, or explicitly declared.
FORTRAN,

and their definitions,

are:

-- a whole number

Integer

Real -- a decimal number;

Double precision -- similar to real, but with more than

that
values
Complex -- a pair of real
represents
value
first
the
number ;

fraction,

that is, a whole number, a

or a combination of the two

the degree of accuracy in its representation

number,

the second represents the

twice

complex
a
represent
the real part of that

imaginary part

true or false

Logical -- the logical value,

Character -- a sequence of characters

An important attribute of each data

decimal

type

1is

the

amount

memory

required to represent a value of that type. Variations on the basic
the
represented value or
the
types affect either the accuracy of

allowed

range of values.

of
amount
ANS FORTRAN specifies that a "numeric storage unit" is the
Double
storage needed to represent a real, integer, or logical value.
In
units.
storage
precision and complex values occupy two numeric
bytes
4
to
corresponds
unit
storage
numeric
a
IV-PLUS,
FORTRAN
VAX-11
of

memory.

ANS FORTRAN specifies that a "character storage unit" is the amount of
1In VAX-11l FORTRAN
represent one character value.
to
needed
storage
IV-PLUS, a character storage unit corresponds to 1 byte of memory.

VAX-11 FORTRAN IV-PLUS provides additional data types for optimum
Table 2-2 lists the
selection of performance and memory requirements.
and
data types available, the names associated with each data type,

The form *n appended to a
amount of storage required (in bytes).
the
specifier.
length
type
data
a
called
is
name
type
data

FORTRAN

Data

Type

STATEMENT

COMPONENTS

Table 2-2
Storage Requirements

Data Type

Storage

Requirements

(in

BYTE

LOGICAL

LOGICAL*1

bytes)

5
or

LOGICAL*2

LOGICAL*4

INTEGER

2 or 4P

INTEGER*2
INTEGER*4

2
4

REAL

REAL¥*4

REAL*8
DOUBLE

8
8

PRECISION

COMPLEX

COMPLEX*8

CHARACTER*1len

len®

CHARACTER*
(*)

fThe

l-byte storage area can contain

false,

single

character,

the

logical

integers

values
the

true

range

-128

to +127.

gither 2 or 4 bytes are allocated depending on
command

qualifier

specified.

The

default

the

compiler

allocation

bytes.
€

The value.of len is the number of characters specified.
The
value
of
1len can be in the range 1 to 32767.
Passed length
format -- *(*) -- applies
only
to
dummy
arguments
or
character
functions,
and
actual arqument or function

indicates
that the length
is used (see Chapter 6).

The following sections contain additional
types and their representations.

2.3

these

the

data

CONSTANTS

A constant represents a fixed
logical value, or a character

Octal,
assume
2.3.5

descriptions

hexadecimal, and
the data type of
and

2.3.8).

value and
string.

can

numeric

Hollerith constants have no data
the context in which they appear

value,

type.
They
(see Sections

FORTRAN STATEMENT COMPONENTS

2.3.1

Integer

Constants

An integer constant is a whole number with no decimal point.
have a leading sign and is interpreted as a decimal number.
An

constant has the

integer

It can

form:

snn

An optional

sign.

A string of numeric

Leading

zeros,

if any,

characters.

are

ignored.

A plus
A minus sign must appear before a negative integer constant.
1is optional before a positive constant (an unsigned constant is
sign
assumed to be positive).

constant
integer
an
sign,
algebraic
1leading
Except for a
contain any character other than the numerals 0 through 9.

The absolute value of an
2147483647.

Examples

of valid

and

integer

invalid

constant

integer

cannot

greater

than

constants are:

Invalid

Valid
0
-127
+32123

99999999999
3.14
32,767

(too large)
(decimal point and
comma not allowed)

it
If the value of the constant is within the range -32768 to +32767,
represents
a
2-byte signed quantity and is treated as INTEGER*2 data
type.
If the value is outside that
range,
it
represents
a
4-byte
signed quantity and

Integer

data

type.

constants can also be specified in octal

form.

The octal

treated

INTEGER*4

form of an integer constant is:

llnn

A string of digits
Examples of valid and

in the

range

invalid octal

integer

constants are:

Valid

Invalid

"107

"108

(contains a digit outside the allowed

"177777

"1377.
"17777"

(contains a decimal point)
(contains a trailing quotation mark)

range)

FORTRAN

2.3.2

Real

constant

real

STATEMENT COMPONENTS

Constants
can

any

one

basic

real

constant

basic

real

constant

integer

constant

A basic real constant
following forms:

the

following:

followed

followed
string

by
a

decimal

exponent

decimal

digits,

(decimal

digits).

any

the

S.nn
snn.nn
snn.

An optional

sign

numeric

string

characters

The decimal point can appear anywhere in the string.
The
number
of
digits
1is
not
1limited, but typically only the leftmost 7 digits are
significant.
Leading zeros (zeros to the left of
the
first
nonzero
digit)
are
ignored
in counting the leftmost 7 digits.
Thus, in the
constant

zeroes,
A

0.00001234567,

are

decimal

all

the

nonzero

digits,

and

none

the

significant.

exponent

has

the

form:

Esnn

optional

integer

sign.

constant.

The

exponent represents a
integer
constant
1is
to
the

value

1.0

power

of 10 by which the preceding
real
or
multiplied (for example, 1.0E6 represents

6),.

real constant occupies 4 bytes
as
a
real number with a degree
digits.

VAX-1l

storage and is
interpreted
precision of, typically, 7 decimal

A minus sign must appear between the letter E and
A plus sign is optional for a positive exponent.
Except
used),

for
a

numerals

algebraic signs, a decimal point,
real
constant
cannot contain any
through

negative

and
the
character

exponent.

1letter
E
other than

(if
the

If the letter E appears
1in
exponent field must follow.
it can be zero.
The magnitude
approximately

a
real
constant,
The exponent field

of a non-zero
real
0.29E-38 or greater

an
integer
constant
cannot be omitted, but

constant
cannot
than approximately

be
smaller
1.7E38.

than

FORTRAN STATEMENT COMPONENTS

Examples of valid and invalid real constants are:
vValid

Invalid

3.14159
621712.
-.00127
+5.0E3
2E-3

1,234,567
325E-45

(commas not allowed)
(too small)

(too large)
(decimal point missing)
(special character

-47.E47
100
$25.00

not allowed)

Double Precision Constants

2.3.3

A double precision constant is a basic real

constant

constant followed by a decimal exponent of the form:

integer

Dsnn

An optional

sign.

integer

constant.

The number of digits that precede the exponent

not

typically only the leftmost 16 digits are significant.

1limited,

but

A double precision constant occupies 8 bytes of VAX-11l storage and is
interpreted as a real number with a degree of precision that is
typically 16 significant digits.

A minus sign must appear before a negative double precision constant.
Similarly, a
A plus sign 1is optional before a positive constant.
exponent;
minus sign must appear between the letter D and a negative
a plus sign is optional for a positive exponent.

The exponent field following the letter D cannot be
can be

omitted,

but

zero.

The magnitude of a non-zero double precision constant cannot be less
than approximately 0.29D-38 or greater than approximately 1.7D38.
Examples of valid and invalid double precision constants are:
Valid

1234567890D+5

+2.7182815.846182D00

-72.5D-15

1DO

Invalid

1234567890D45

(too large)

+2.7182812846182

(no Dsnn present;

1234567890.0D-89

(too small)

this is a valid

single precision

constant)

2.3.4

Complex Constants

A complex constant is a pair of real constants separated by a comma
The first real constant represents the
in parentheses.
and enclosed
real part of that number and the second real constant represents the
imaginary part.

FORTRAN

complex

constant

has

the

STATEMENT

COMPONENTS

form:

(rc,rc)
rc

real

constant.

The

parentheses

See

Section

complex

and

2.3.2

comma

for

constant

interpreted

Examples

valid

are

the

part

rules

occupies

complex
and

for
8

the

constant

forming
bytes

real

and

are

required.

constants.

VAX-1ll

storage

and

number.

invalid

complex

constants

Valid

Invalid

(1.70391,-1.70391)

(1,2)
(1.23,)

(+12739E3,0.)

are:

(integers are not allowed)
(second real constant is

missing)
(1.0,1.000)

(double

constants

2.3.5

Octal

and

numeric

and

Hexadecimal

hexadecimal

They

not

allowed)

Constants

constants

constants.

precision
are

are

can

alternative

used

wherever

ways

numeric

represent

constants

are

allowed.
An

octal

and

constant

followed

the

string

octal

digits

alphabetic

character

range

enclosed

octal

apostrophes

constant

has

the

form:
1

clczc3...cn t 0]

A
A

digit

hexadecimal

enclosed

hexadecimal

the

constant

a string of hexadecimal digits
and
letters
and followed by the alphabetic character X.
A
has the form:

apostrophes

constant

'e.c.c....c
123

A digit
to f.
Leading

zeros

the

are

constants

must

constants

must

Examples

valid

range

ignored
in
in
and

the
the

octal

range
range

invalid

'0'0
'0'X
octal

Valid

Invalid

'07737'0
'1'0

'7782'0
7772'0
'0737"

and
to
to

letter

the

hexadecimal

range

constants.

'37777777777'0,
'FFFFFFFF'X.

constants

and

Octal

hexadecimal

are:

(invalid character)
(no initial apostrophe)
(no O after second apostrophe)

FORTRAN STATEMENT COMPONENTS

Examples of valid and invalid hexadecimal constants are:
Invalid

Valid

(invalid character)

'999.'X

'AF9730'X

(no apostrophe before

'FI9X

'FFABC'X

the

They
Octal and hexadecimal constants are typeless numeric constants.
assume data types based on the way they are used, and thus are not
converted

)

before

use.

including
When the constant is used with a binary operator,
the assignment operator, the data type of the constant is the
data type of the other operand.

For example:
Data

Type
Length of
of Constant Constant

Statement
INTEGER*2 ICOUNT
REAL*8 DOUBLE

)

When a specific data type

1is

required,

that type is assumed for the constant.

generally

For example:

INTEGER*4

Y(IX)=Y('15'0)+3.

When the constant is used as an actual argument, no data type
is assumed.

For

example:

Data Type

When the constant is used in
is assumed.

I =
J =

1,2,3

any

For example:

Statement

IF('AF77'X)

'"7777'0 - 'A39'X
.NOT.'73777'0

none

CALL APAC('34BC2'X)
data type

Length

of Constant Constant

Statement

)

integer,

Length of
Data Type
of Constant Constant

Statement

4
2
8
4

REAL*4
INTEGER*2
REAL*8
INTEGER*4

RAPHA = '99AF2'X
JCOUNT = ICOUNT + '777'0
DOUBLE = 'FFF99A'X
IF(N.EQ.'123'0) GO TO 10

other

INTEGER%*4

context,

Data Type

Length

INTEGER* 4

of Constant Constant

INTEGER*4
INTEGER*4

4
4

An octal or hexadecimal constant actually specifies 4 bytes of data.
When the data type implies that the length of the constant is more
When the data
than 4 bytes, the leftmost digits have a value of 0.
4 bytes, the
than
less
is
constant
the
of
length
the
that
implies
type
lists
Table 2-2 (in Section 2.2)
constant is truncated on the left.
the number of bytes that each data type requires.

FORTRAN

2.3.6

Logical

STATEMENT

COMPONENTS

Constants

A logical constant
only the following

specifies a
two logical

logical value, true
or
constants are possible:

false.

Thus,

.TRUE.

.FALSE,

The

delimiting

2.3.7

periods

Character

are

constant
1is
apostrophes.

constant

Both

has

A printable

character.

delimiting

apostrophes

value

the

delimiting

apostrophes,

string

the

The

required

part

each

constant.

Constants

A character
enclosed by
character

character
does

printable

ASCII

must

constant

The

include

present.

the

value
all

string

does

not

spaces

characters

include
or

apostrophes.

represent

must

Examples

the character
apostrophes,

single

the

apostrophe.

range

valid

constant
is
except
that

and

The

the
two

length

within

invalid

character

constants

the

is
represented
other character

number
of
consecutive
a

between

delimiting

characters
apostrophes

character

255.

Valid

constant

are:

Invalid

'WHAT?'

'HEADINGS

'TODAY''S

DATE

'HE

"HELLO"'

SAID,

IS:

(no

trailing

apostrophe)

(character

"NOW OR NEVER"

(quotation marks

constant
must contain at least
1 character)
,
cannot

place
If a
the

the

tabs

Within a character constant, the apostrophe character
by
two
consecutive
apostrophes
(with
no
space or
between them).
The length of
between
the

characters

form:

apostrophes.
but

character

used in
apostrophes)

constant appears in a numeric context (for example,
as
expression
on
the
right
side
of
an
arithmetic
assignment
statement), it is considered a Hollerith constant.
See Section 2.3.8.

FORTRAN

2.3.8

STATEMENT

COMPONENTS

Hollerith Constants

A Hollerith constant is a string of printable characters preceded by a
character

count and the

letter H.

A Hollerith constanfi has the form:
nHc.
n

¢c.Cc....C
n
1273

An unsigned, nonzero
integer
constant
stating
the
characters in the string (including spaces and tabs).

number

A printable

character.

The maximum number

of characters

255.

Hollerith constants are stored as byte strings,

1 character per

byte.

Hollerith constants have no data type.
They
assume
a
numeric
data
Hollerith
in which they are used.
context
the
to
according
type
constants cannot assume a character data type;
they
cannot
be
used

where

a character

value

Examples of valid and

is expected.

invalid Hollerith constants are:

Valid

Invalid

16HTODAY'S DATE 1IS:

3HABCD

(wrong number of characters)

1HB

When Hollerith constants are used in numeric expressions,
data types according to the following rules.

they

assume

including
operator,
When the constant is used with a binary
the assignment operator, the data type of the constant is the
data type of the other operand.

For example:

Length of
Data Type
of Constant Constant

Statement
INTEGER*2 ICOUNT
REAL*8 DOUBLE

RALPHA = 4HABCD
JCOUNT = ICOUNT + 2HXY
DOUBLE = 8HABCDEFGH
IF(N.EQ.1HZ) GO TO 10

)

When a specific data type

1is

required,

that type is assumed for the constant.

Y (IX)=Y(1HA)+3.
When the constant
is

assumed.

For

generally

For example:

integer,

INTEGER*4

is used as an actual

argument,

no data type

example:
Data

Statement

CALL APAC

4
2
8
4

Length of
Data Type
of Constant Constant

Statement

)

REAL*4
INTEGER*2
REAL*8
INTEGER*4

(9HABCDEFGHI)

Type

Length

of Constant Constant

none

FORTRAN

When
data

When

the

data

type,

STATEMENT

the constant is used in
any
other
type is assumed.
For example:

INTEGER*4

Data Type
of Constant

(2HAB)

INTEGER*4

1HC-1HA

INTEGER*4

.NOT.

INTEGER*4

spaces

the

are

1,2,3
1HB
constant

appended

less

the

the constant is greater than
constant is truncated on the

Table 2-2
each data

type.

Section
Each

2.2)

lists

character

the

than

the

constant

length of
type, the

2.4

context,

Statement

length

(in

COMPONENTS

Length of
Constant

length

the length
right.

the

implied

right.

implied

the

When

the

data

number of characters required
1 byte of storage.

for

occupies

VARIABLES

A variable is a symbolic name associated with a storage
value
of the variable is the value currently stored in
that value can be changed by assigning a new value
to
(See Section 2.1 for the form of a symbolic name.)
Variables are
data
type of

classified
a variable

by data type,
indicates the

precision, and itc¢ storage
assigned
to
a
variable,
type of the variable.
You
by
type
declaration
typing rules.

location.
The
that location;
the
variable.

as
constants
are.
The
of data it represents, its
requirements.
When data
of
any
type
|is
it is converted, if necessary, to the data
can establish the data type of
a
variable

statements,

just
type

IMPLICIT

statements,

predefined

Two or more variables are associated with
each
other
when
each
is
associated
with
the
same
storage
location.
They
are
partially
associated, when part (but not all) of the storage associated with one
variable
is
the
same
as part or all of the storage associated with
another variable.
Association and partial association occur when
you
use the COMMON statements, EQUIVALENCE statements, or actual arguments
and dummy arguments in subprogram references.
A variable is
contains data

considered defined if the
storage
associated
with
it
of the same type as the name.
A variable can be defined

before

program execution by a DATA
assignment or input statement.

1f variables
associated)

of different data
types
with
the
same
storage

variable is defined (for example,
other variable becomes undefined.

2.4.1

Data

Type

declaration
variables are to
COMPLEX
DOUBLE

statement

during

execution

are
associated
(or
partially
location,
and the value of one
assignment),

the

value

the

Specification

statements
(see
Section
5.2)
represent specified data types.

VARI1
PRECISION

VAR2

specify
that
For example:

given

FORTRAN

STATEMENT

COMPONENTS

These statements indicate that the variable VARl is to
be
associated
with
an
8-byte storage location that is to contain complex data, and
that the variable VARZ2 is to
be
associated
with
an
8-byte
double
precision storage cation.
The IMPLICIT statement (see Section 5.1)
has
a
broader
scope.
states
that,
1in
the
absence
of
an explicit type declaration,
variable with a name that begins
with
a
'specified
letter,
or

letter

within

specified

range,

1is

represent

specified

It
any
any
data

only once.
over
the

An
type

type.

You can explicitly specify the data
explicit
data
type
specification
implied by an IMPLICIT statement.
Character type
specify
that
the

length

type of
takes

a variable
precedence

declaration statements (see
Sections
5.1
given
variables are to represent character

specified.

CHARACTER*72

For

example:

INLINE

CHARACTER NAME*12,

NUMBER*9

These statements indicate that the variables
are
to be associated with storage locations
of lengths 72, 12, and 9, respectively.
Passed

length character arguments are used
process character strings of different

character

and
5.2.2)
values with

argument

has

length

INLINE, NAME, and
NUMBER
containing character data

within

a
lengths.

specification

single

The

(*).

subprogram
passed length

For

example:

CHARACTER*
(*) CHARDUMMY
The

passed

argument

2.4.2

length

(see

Data

Type

In
the
absence
statements,
all
N are assumed to
with

any

Real

2.5

character

Chapter

other

argument

assumes

the

length

the

actual

6).

Implication

of
either
IMPLICIT
statements
or
explicit
type
variables with names beginning with I, J, K, L, M, or
be integer variables.
Variables with names beginning

letter

are

assumed

Variables

real

Integer

variables.

For

example:

Variables

ALPHA

JCOUNT

BETA

ITEM

TOTAL

NTOTAL

ARRAYS

An array
single

1is a group of contiguous storage locations associated with
a
symbolic
name,
the
array
name.
The
1individual
storage
locations, called array elements,
are
referred
to
by
a
subscript
appended

An
of

array can
figures

column

the

array

name.

Section

have from one to seven
1is
a
one-dimensional

figures

2.5.2

discusses

dimensions.
array.
A

two-dimensional

array.

subscripts.

For example,
table of more

refer

a column
than one

specific

FORTRAN

STATEMENT COMPONENTS

value
in
this
array,
you
must specify both its row number and its
a
1is
pages
several
covers
A table of figures that
column number.
locate a value in this array, you must
To
array.
three-dimensional
specify the row number, column number, and a page number.
The following FORTRAN statements establish arrays:
e

Type declaration statements

The DIMENSION statement

The COMMON statement

(see Section 5.2)

(see Section 5.3)

(see Section 5.4)

These statements contain array declarators (see
Section
define
the
name of the array, the number of dimensions
and the number of elements in each dimension.

2.5.1)
that
in the array,

An element of an array is considered defined if the storage associated
of the same data type as the array name (see
data
contains
it
with
defined
be
can
An array element or an entire array
Section 2.5.4).
An array element can be
before program execution by a DATA statement.
statement;
defined during program execution by an assignment or input
be defined during program execution by an
can
array
entire
an
and
input

statement.

2.5.1

Array Declarators

An array declarator specifies the symbolic
name
that
identifies
an
array
within
a
program
wunit
and
indicates the properties of that
array.

An array declarator
a
a

(d[,d]l

has the

...)

The symbolic
name
of
(Section 2.1 gives the
A dimension declarator;
upper

form:

bound

the
array,
that
1is,
the
form of a symbolic name.)

4 can

specify both a lower

array

name.

bound and

follows:

[d1l:]du
dl

The lower

bound of

the dimension.

The upper

bound of

the dimension.

The number of dimension declarators indicates the number of dimensions
in the array.

The number of dimensions can range from one to seven.

The value of the lower bound dimension

declarator

can

negative,

zero,
or positive.
The value of the upper bound dimension declarator
must be greater
than
or
equal
to
the
corresponding
1lower
bound
dimension
declarator.
The
number
of
elements in the dimension 1is
du-dl+1l.
If a lower bound is not specified, it is assumed
to
be
1,

and
the
value of the upper bound specifies the number of elements in
that dimension.
For example, a dimension declarator of
50
indicates
that the dimension contains

elements.

FORTRAN STATEMENT COMPONENTS

Each dimension bound

is an

integer

arithmetic expression in which:

Each
operand
1is
an
integer
constant,
an
integer
argument, or an integer variable in a COMMON block

Each

operator

dummy

** operator

Note that array references and function references are not allowed

dimension bounds expressions.

Dimension bounds that are not constant expressions can be
used
in
a
subprogram to define adjustable arrays.
You can use adjustable arrays
within a single subprogram to process arrays with different
dimension
bounds
by
specifying
the
bounds
as
well
as
the
array
name as
subprogram arguments.
See Section 6.1.1.1
for
more
information
on
adjustable arrays.
Dimension bounds that are not constant expressions
are

not permitted

in a main program.

The number of elements
number

elements

in an array is equal

in each

An array name can appear
unit.

2.5.2

the

product

the

in only one array declarator within a program

Subscripts

A subscript qualifies
expressions,
called
that determine which
subscript is appended
A

dimension.

subscript

has

the

an
array
name.
A
subscript
1is
a
1list
of
subscript
expressions, enclosed in parentheses,
element
in
the
array
1is
referred
to.
The
to the array name it qualifies.

form:

(s[,s]...)
A

subscript

expression.

A subscripted array reference must contain
one
subscript
expression
for
each
dimension
defined
for
that array (one for each dimension
declarator).

Each subscript can be any valid arithmetic expression.
If
the
value
of
a
subscript is not of type integer, it is converted to an integer
value by truncation of any fractional part before use.

2.5.3

Array Storage

As discussed earlier in this section, you can think of the
dimensions
of
an array as rows, columns, and levels or planes.
However, FORTRAN
always stores an array in memory as a linear sequence
of
values.
A

one-dimensional
array
is
stored with its first
storage location and its last element in the last
the sequence.
A multidimensional array is stored
subscripts vary most rapidly.
This is called the
progression." For example, Figure 2-1 shows array
and

three dimensions.

element in the first
storage location
of
so that the leftmost
"order of
subscript
storage in one, two,

FORTRAN

STATEMENT

COMPONENTS

1-Dimensional Array BRC (6)

F[BRC(H l 2|BRC(2)[ BIBRC(3)I 4IB'RC(4)| sl BRC (5) | GlBRC(G)J
Memory Positions

BAN (1,1)

BAN (1,2)] 7 { BAN (1,3) |10 | BAN (1,4)

BAN (2,1)

|BAN (2,2)|

—l
W N

2-Dimensional Array BAN (3,4)

BAN (2,4)

BAN (3,1){ 6 |BAN (3,2}| 9| BAN (3,3)| 12|

BAN (3,4)

BAN (2,3}

Memory Positions

3-Dimensional Array BOS (3,3,3)

19 | BOS (1,1,3) | 22

BOS (1,2,3) | 26 | BOS (1,3,3)

20 | BOS (2,1,3) | 23|B0OS (2,2,3) | 26 | BOS (2,3,3)
10 | BOS (1,1,2) { 13|B0S (1,2,2) | 16 | BOS (1,3,2)

BOS (3,3,3)

11 | BOS(2,1,2) | 14|B0S (2,2,2) | 17 | BOS (2,3,2)
1]180S(1,1,1)|

4 |BOS (1,2,1)| 7 | BOS {1,3,1)

18 | BOS (3,3,2)

,BOS (2,1,1) | 5 | BOS (2,2,1) | 8 | BOS (2,3,1)
BOS (3,1,1)

BOS (3,2,1)

BOS (3,3,1)

Memory Positions

Figure

2.5.4

Data Type of

2-1

Array

Storage

an Array

The

data type of an array is specified the same way as the
data
type
of
a
wvariable.
That
1is,
the
data
type of an array is specified
implicitly by the initial letter of the name, or explicitly by a
type
declaration statement.
All the values in an
array
have
the
same
data
type.
Any
value
assigned
to
an
array
element
1is converted to the data type of the
array.
If an array is named in
a
DOUBLE
PRECISION
statement,
for

example,
the
compiler
allocates an 8-byte storage location
element of the array.
When a value of any type
1s
assigned
element of that array, it is converted to double precision.

2.5.5

Array

for
to

each
any

statement, you can
specify
an
array
indicate that the entire array is to be

name
used

References Without

In the following types
without
a
subscript,
(or defined):
e

Type

COMMON

DATA

of
to

declaration
statement

statement

Subscripts

statements

FORTRAN

EQUIVALENCE

FUNCTION

SUBROUTINE

ENTRY

Input/output

STATEMENT

COMPONENTS

statement

statement
statement

statement

statements

You can also use unsubscripted array
names
as
actual
arguments
1in
references
to
external
procedures.
The use of unsubscripted array
names in all other types of statements is not permitted.

2.5.6

Adjustable Arrays

Adjustable

arrays

dimensions.

the

allow

wuse

subprograms

adjustable

array bounds, as well as its
6 for more information.

CHARACTER

manipulate arrays
of
variable
array in a subprogram, you specify

name,

Chapter

2.6

character

subprogram

arguments.

See

SUBSTRINGS

A character substring is a contiguous
or character array element.
A

substring

reference

has

segment

one

the

character

following

variable

forms:

v(lel]:[e2])

a(s[,sl...)

(lel]:[e2])

character

variable

character

array

subscript

expression.

name.

el
A
numeric
position of

expression
that
the substring.

specifies

the

leftmost

character

rightmost

character

e2
A

numeric
expression
that
position of the substring.

Character

positions

within

numbered

from

left

right,

specifies

character
beginning

variable
at

specifies the substring
and
ending
with
the

beginning with the
seventh
<character

variable

LABEL.

CHARACTER*8

XVERSUSY,

then

the

substring

the

has

array

element

example,

second
character
position
of the

variable

LABEL(2:7)

or
For

LABEL

value

has

are

LABEL(2:7)

position
character
value

VERSUS.

FORTRAN

If the value of the numeric
integer,
it
1is
converted
fractional part before use.

The value of
1

STATEMENT COMPONENTS

expression
el
or
e2
1is
not
of
type
to
an integer value by truncation of any

the numeric expressions,

LE.

.LE.

el and e2, must be such that:
1len

where len is the length of the character variable or

array element.

If el

If e2

is omitted,

FORTRAN

assumes

FORTRAN assumes that el equals 1.

that

equals

len.

is omitted,

For example, NAMES(1,3)
(:7) specifies the substring starting with
the
first
character
position
and
ending
with
the
seventh
character
position of

2.7

the character

array element NAMES(1,3).

EXPRESSIONS

An expression represents a single value.
It can
be
a
single
basic
component,
such
as a constant or variable, or a combination of basic
components with one or more operators.
Operators specify computations
to be performed, using the values of the basic components, to obtain a
single

value.

Expressions are classified as arithmetic,
character,
relational,
or
logical.
Arithmetic
expressions
produce numeric values;
character
expressions produce character
values;
and
relational
and
1logical
expressions produce logical values.

2.7.1

Arithmetic Expressions

Arithmetic
arithmetic
single

expressions
operators.

numeric

are
formed
with
arithmetic
elements
and
The evaluation of such an expression yields a

value.

An arithmetic element can be any of

the

following:

)

A numeric

constant

A numeric variable

)

A numeric

)

An arithmetic

expression enclosed

An arithmetic

function

array

element

in parentheses

reference

The term "numeric," as used above, can also be interpreted to
include
logical
data, 'since logical data is treated as integer data when used
in

arithmetic

context.

FORTRAN

Arithmetic
values
of
result.

STATEMENT

COMPONENTS

operators specify a computation to be performed
arithmetic
elements.
They
produce a numeric

The

operators

and

their

meanings

Operator

These
two

the
as a

are:

Function

Exponentiation

Multiplication

Division

Addition

Subtraction

operators

are

elements.

called

The

binary

plus

(+)

operators,
and

minus

Any arithmetic operator can be
except as noted in Table 2-3.

used

with

or array element must have
arithmetic expression.

Table 2-3 shows the allowed
and exponent data types for

Allowed

and

unary

and

because

unary
each

arithmetic

any valid

defined

2-3
for

minus

symbols

are

used
also

value

with
unary

element

arithmetic

before

combinations and result data
the exponentiation operator.

Table
Combinations

plus

(-)

operators when written immediately preceding
denote a positive or negative value.

A variable
used in an

using
value

element,

types

can

base

Exponentiation

Exponent
Base

Integer

Real

Double

Integer

Real

Double

Double | Double

Complex

Note:

only

negative
an integer

zero value
zero-value

cannot

element
element;
be

Complex

can
and

be
an

exponentiated
element with a

exponentiated

another

element.

In any valid exponentiation, the result has the same data type as
the
base element, except in the case of a real base and a double precision
exponent.
The result in this case is double precision.

FORTRAN STATEMENT COMPONENTS

in an order
Arithmetic expressions are evaluated
The
precedence associated with each operator.

by a
determined
precedence of the

is:

operators

Operator

Precedence

First

* and /

Second

+ and -

Third

(such as + and -)
When two or more operators of egual precedence
long as the order of
as
order,
any
in
evaluated
be
can
they
appear,
evaluation is algebraically equivalent to a left-to-right order of
Exponentiation, however, is evaluated from right to left.
evaluation.
is evaluated
B**C
For example, A**B**C is evaluated as A** (B**C);
then A is raised to the resulting power.

first,

Use of Parentheses - You

2.7.1.1

particular

evaluation.

order

can

use

When

parentheses

part

force

expression is

and the
that part 1is evaluated first,
enclosed in parentheses,
resulting value is wused in the evaluation of the remainder of the
the
the numbers below
In the following examples,
expression.
operators indicate the order of the evaluations.
6

N>
+

(4+3)
4

((4+3)

As shown in the third and fourth examples

above,

expressions

within

expression,

parentheses are evaluated according to the normal order of precedence,
unless you override the order by using parentheses within parentheses.

Nonessential parentheses, as

affect expression evaluation.
4 +

(3*2)

the

following

not

(6/2)

1is often
The use of parentheses to specify the evaluation order
In such
computations.
numerical
y
high-accurac
in
important
computations, evaluation orders that are algebraically equivalent

might not be computationally eguivalent when processed by a computer.

FORTRAN

2.7.1.2

an
by

Data

Type

arithmetic
the

an Arithmetic

expression

STATEMENT

also

that

associated
with
is as follows:
Data

each

data

Expression

the

data
types
are
combined
in
expression and the data type of

COMPONENTS

same

data

an
the

type.

data

every

type.

the

The

rank

element

value

elements

expression, the
resulting value

Type

different

evaluation
depend on

assigned

produced
of
a

that
rank

each data

type

Rank

Logical

Integer*2

Integer*4

Real

Double

type,

precision

(Low)

Complex

(High)

The

data type of the value produced by an operation on two
arithmetic
elements
of
different
data
types
1is
the
data
type
of
the
highest-ranked element in the operation.
For example, the
data
type
of
the
wvalue
resulting
from
an operation on an integer and a real
element is real.

The

data

last

type

operation

determined
)

expres:tion

that

the

exrression.

data

The

type

data

the

type

Integer

operations

Integer

fraction

that

can

rounded.

For

example:

1/3

Real

value

1/3

the

operations

are

elements.
Any
data
type
by

expression
however,

division

is
that

only

from

division

not

operations

are

performed

only

integer

and

logical

expression

Real

combinations

integer
giving

then
in

operation

multiplication

result

performed

in an arithmetic
arithmetic,
any
is

truncated,

not

1/3

this

operations

elements

)

expression

follows:

on integer elements.
(Logical entities used
context are treated as integers.) In integer

The

result

real,

elements present
a fractional

each

evaluated
the

using

real

statement

performed

are
part

that

converted
equal to

to
0.

arithmetic.

(I/J)*X,

and

result

and

real
real
The

Note,
integer

real

Double precision operations -- Any real or integer element in
a double precision operation is converted to double precision

data type by making the existing element the most
portion
of
a double precision datum.
The least
portion is 0.
The expression is
then
evaluated
precision

arithmetic.

significant
significant
in
double

FORTRAN STATEMENT COMPONENTS

does
Converting a real element to a double precision element
For example, the real number
not increase its accuracy.

0.3333333

converted

0.3333333000000000D0
not

0.3333333333333333D0

Complex

)

operations -- In

operation

that

contains

any

complex
element, integer elements are converted to real data
Double precision elements are
type, as previously described.

converted to real data type by rounding the least significant
The real element thus obtained is designated as the
portion.
The
the imaginary part is 0.
real part of a complex number;
expression is then evaluated using complex arithmetic and the
resulting value

is of complex data type.

Character Expressions

2.7.2

elements and character
Character expressions consist of character
The evaluation of a character expression yields a single
operators.
value of character data type.

A character element can be any one of the following:
constant

A character

A character variable

A character

array element

A character

substring

A character expression enclosed in parentheses

)

A character

function reference

The only character operator

is the concatenation operator

(//).

A character expression is a sequence of one or more character elements
separated by the concatenation operator.

A character expression has the

character element

form:

[//character element]...

The value of a character expression is a character

string

formed

left-to-right concatenations of the values of the elements

successive

of the character expression. The length of a character expression is
For example, the
the sum of the lengths of the character elements.
value of the character expression 'AB'//'CDE' is 'ABCDE', which has a
length

expression.
character
Parentheses do not affect the value of a
example, the following character expressions are equivalent.
('ABC'//'DE') //'F!
"ABC'///('DE'//"F')

For

FORTRAN

Each

these

character

STATEMENT

expressions

has

COMPONENTS

the

value

'ABCDEF'.

a character element in a character expression contains spaces,
spaces
are
included
in
the value of the character expression.
example, 'ABCA'//'DAE'//'FA' has a value of 'ABCADAEFA'.

The
are

order in which
evaluated
1is

the
For

the character elements of
a
character
expression
determined by the compiler even if parentheses are

present.

2.7.3

Relational

relational

Expressions

expression

consists

two

arithmetic

expressions

two

character
expressions, separated by a relational operator.
The value
of the expression is either true or false, depending on whether or not
the

stated

relationship

A relational
expressions

exists.

operator tests for a relationship between
or
between
two
character expressions.

two
arithmetic
These operators

are:

Operator

The

Relationship

.LT.

Less

than

.LE.

Less

than

.EQ.

Equal

.NE.

Not

.GT.

Greater

than

.GE.

Greater

than

equal

required

part

each

delimiting

periods

Complex expressions
operators.
Complex
imaginary parts are
In

arithmetic

first

evaluated

both

determine

For

example:;

expression

related
are equal

equal

only
by
if their

operator.

the
.EQ.
and
.NE.
corresponding real and

equal.

relational expression, the arithmetic expressions are
to
obtain
their
values.
These
values
are then

compared

This

can
be
entities

exists.

APPLE+PEACH

are

equal

.GT.

whether

the

relationship

stated

the

operator

PEAR+ORANGE

states

the

relationship,

"The

sum

the

real

variables
APPLE
and
PEACH
1is
greater
than
the
sum
of the real
variables PEAR and ORANGE." If tiat relationship exists, the value
of
the expression is true;
if not, the value of the expression is false.

FORTRAN STATEMENT COMPONENTS

expressions are
In a character relational expression, the character
These values are then
first evaluated to obtain their values.
compared to determine whether the relationship stated by the operator
In character relational expressions "less than" means
exists.

"precedes in the ASCII collating seqguence," and "greater
"follows in the ASCII collating sequence." For example:
'AB'//'22Z"'

.LT.

than"

means

'CCCCC!

That
'CCCCC'.
is less than
'ABZZZ'
This expression states that
1If
relationship does exist, so the value of the expression is true.
1is
expression
the
of
value
the
exist,
not
the relationship stated does
false.

If the two character expressions in a relational expression are not
the shorter one is padded on the right with spaces
the same length,

until the lengths are equal.
'ABC'

.EQ.
.LT.

'AB'

For example:

'ABCAAA'
'C!

The first relational expression has a value of true even though the
lengths of the expressions are not equal, and the second has a value
of true even though

'AB'

is longer than

'C'.

All relational operators have the same precedence.
a higher precedence
have
operators
character

Arithmetic and
than relational

operators.

to
You can use parentheses, as in any other arithmetic expression,
a
in
expressions
arithmetic
the
of
evaluation
of
alter the order
However, arithmetic and character operators
relational expression.

are

evaluated before relational operators so you need not enclose the

entire arithmetic or character expression in parentheses.

Two numeric expressions of different data types can be compared by a
relational expression. 1In this case, the value of the expression with
the lower-ranked data type is converted to the higher-ranked data type
.

before the comparison is made.

Logical Expressions

2.7.4

A logical expression can be a single logical element, or a combination
A logical expression
of logical elements and 1logical operators.
yields a single logical value,

true or false.

A logical element can be any of the following:
®

An integer or logical constant

An integer or logical variable

An integer or logical array element

A relational expression

)

A logical expression enclosed in parentheses

)

An integer or logical function reference

FORTRAN

The

logical

operators

Operator

STATEMENT

COMPONENTS

are:

Example

.AND.

.OR.

Meaning

Logical
if, and

conjunction:
only if, both

Logical

disjunction

expression
both,
. XOR.

.XOR.

.EQV.

(inclusive

OR):

1if

either

Logical

but

the

have

the
B is

The
When

.NOT.

delimiting

expression

the

same

equivalence:

periods

Logical

negation:

if,

only

and

logical

the

expression

false,

false

if,

the

operators

is
are

or
if

is
vice
both

value.
expression

if,
and only if, both A and B
logical value, whether true or
.NOT.

true

true.

elements
A

true

Logical exclusive
OR:
true
if
A
1is true and
versa;

.EQV.

the expression is
A and B are true.

true

have the
false.

same

expression

1is

true

false.
required.

logical operator operates on logical
elements,
the
resulting
type
1is
logical.
When a logical operator operates on integer
elements, the logical operation
is
carried
out
bit-by-bit
on
the

data

corresponding

integer
operator

bits
of
the
internal
(binary) representation of the
elements.
The resulting data type is integer.
When a logical
combines

first converted to
as for two integer

integer

and

an integer
elements.

1logical

values,

the

logical

value,
The

value

then the operation is carried
resulting data type is integer.

out

logical expression is evaluated according to an order of
precedence
assigned
to its operators.
Some logical expressions can be evaluated
before all their subexpressions are evaluated.
For example, if
A
is
.FALSE.,

the

-FALSE..

The

without

expression

value

the

.AND.

expression

(F(X,Y)
can

.GT.

2.0)

determined

evaluating

.AND.

testing

F(X,Y).
Thus, the function subprogram F may not be
called, and side-effects resulting from the call, for example changing
variables in COMMON, cannot occur.

The following list summarizes all the operators that can appear
logical expression, in the order in which they are evaluated:
Operator

Precedence

* %

First

*,/

Second

ty=4//

Third

(Highest)

Relational
Operators

Fourth

.NOT.

Fifth

.AND.

Sixth

.OR.

Seventh

.XOR. , .EQV.

Eighth

FORTRAN STATEMENT COMPONENTS

Operators of equal

rank

are

evaluated

left

from

For

right.

example:

A*B+C*ABC

.EQ. X*Y+DM/ZZ

.AND.

.NOT. K*B .GT. TT

The sequence in which this logical expression is evaluated is:
(((A*B)+(C*ABC)).EQ.((X*Y)+(DM/ZZ))).AND.(.NOT.((K*B).GT.TT))

As in arithmetic expressions, you can use
normal sequence of evaluation.

parentheses
.

Two logical operators cannot appear consecutively, unless
operator

.NOT..

alter

the

second

CHAPTER
ASSIGNMENT

Assignment
or

statements

character

assigning

the

character

substring.

The

define

substring.

resulting

value

STATEMENTS

value

this

evaluating

the

assignment

statements

are:

)

Arithmetic

assignment

statement

)

Logical

Character

]

ASSIGN

assignment

arithmetic

the
on

arithmetic
v

numeric

and

element,

array

statement

STATEMENT

assignment

element

element,

expression

statement

The
array

array
an

statement

assignment

ARITHMETIC ASSIGNMENT

expression

variable,

four

3.1

The

the

They

statement

assigns

right

the

equal

sign

left

the

equal

sign.

the

assignment

statement

has

the

value

the

mathematics.

numeric

variable

form:

variable

array

element.

expression.

The equal
means "is

sign does not mean "is equal
replaced by." For example:

KOUNT

the

undefined,
references

symbolic

values
in

The

expression

the

wvariable

example,

the

real

"replace
the
sum

name

must

have

expression

must

This statement means,
variable
KOUNT
with
constant 1."
Although

to,"

yield

the
current
value
of
the
of that current value and the

the

left

been

previously

the

value

array

element

expression

that

right

that

the

conforms

to which

produces

equal

assigned

equal
to

value

the

sign

all

integer
integer

can

symbolic

sign.
requirements

assigned.

For

greater

than

32767

ASSIGNMENT STATEMENTS

1if
invalid
variable.

the

entity on the left of the equal sign is an INTEGER*2

the
If the variable or array element on the left of the equal sign has
assigns
t
statemen
the
right,
the
on
on
expressi
the
as
type
data
same
If the data types are different, the value of the
the value directly.
is converted to the data type of the entity on the left of
expression
Table 3-1 summarizes
the equal sign before it is assigned.
conversion rules for assignment statements.
Table

the

3-1

Conversion Rules for Assignment Statements

Expression

Variable or

(E)

Array
Element

Integer,
Logical,

or
Octal Constant

Real

Double
Precision

Complex

Truncate E to
integer and assignto 'V

Truncate E to
integer and assign toV

Truncate real part
ofE to integer
and assign to V;

TM)

Integer
or
Logical

Assign Eto V

imaginary part

of E is not used

Real

Append fraction

(.0) to E and

AssignEtoV

assign to'V

Double
Precision

Append fraction
(.0) to E and as-

sign to MS* porV; LS*
tion of
portion of Vis 0

Assign MS* por-

Assign real part

LS* portion of

inary part of E

AssignEtoV

Assign real part
of E to MS* por-

tion of E to V;
E is rounded

Assign E to MS*
portion of V;

of E to V;imag-

is not used

tion of V; LS*
portion ofV is
zero, imaginary

LS* portion of
Vis0

part of E is not
used

Complex

Append fraction

(.0) to E and assign to real part
of V; imaginary
part of Vis 0.0

Assign E to

real part of V;
imaginary part
of Vis 0.0

Assign MS* portion of E to real
part of V; LS*
portion of E is
rounded; imaginary part of V
is 0.0

* MS = most significant (high order); LS = least significant (low order)

Assign EtoV

ASSIGNMENT

Examples

valid

and

invalid

STATEMENTS

assignment

statements

are:

valid
BETA

3.14159

SUM

=1./(2.*X)+A*A/(4.%*
(X*X))

SUM+1.

Invalid
3.14

A-B

(entity

variable
-J

I*%*4

(entity

the

left

array

the

must

element)

left

must

not

signed)
ALPHA

((X+6)*B*B/(X-Y)

(left

and

right

parentheses

not-

balance)
ICOUNT

'A'//'B'

(expression

character

data

the

3.2

LOGICAL ASSIGNMENT

left

right

not

must

type,

not

the

character

entity

data

on
type)

STATEMENT

The

logical assignment statement assigns
the
value
of
the
logical
expression
on
the
right
of the equal sign to the variable or array
element on the left of the equal sign.
See Table 3-1
for
conversion
rules.

The

logical

assignment

statement

logical

variable

logical

expression.

has

the

form:

The variable
logical data

array

or array element
type.
1Its value

element.

on the
can be

left of the
undefined.

Values, either numeric or logical, must have
to
all
symbolic
references
that
appear
expression must yield a logical value.
of

PAGEND

logical

assignment

.FALSE.

PRNTOK

]

Examples

LINE

.LE.

ABIG

.GT.

132

.AND.

statements

.AND.
A

.GT.

.NOT.
C

equal

are:

must

been previously
assigned
1in
the
expression.
The

PAGEND

.AND.

sign

.GT.

ASSIGNMENT STATEMENTS

3.3

CHARACTER ASSIGNMENT STATEMENT

The character assignment statement assigns the value of the

character

the right of the equal sign to the character variable,
expression on
array element, or substring on the left of the equal sign.

The character assignment statement has the form:
vV

A character variable,
A character

array element,

or substring.

expression.

1length
the
If the length of the character expression is greater than
the character variable, array element, or substring, the character
of
expression is truncated on the right.

of
1length
If the length of the character expression is less than the
element, or substring, the character
array
variable,
character
the
is filled on the right with spaces.

expression

be
can
sign
egqual
the
of
1left
Although the symbolic name on the
values must have been previously assigned to all symbolic
undefined,
in the expression.

references

assign
You cannot
The expression must be of character data type.
numeric value to a character variable, array element, or substring.

affect
not
Note that assigning a value to a character substring does
the character variable or array element that
in
positions
character
If a character position outside of
are not included in the substring.
has a value previously assigned, it remains unchanged,
substring
the
and if the character position is undefined, it remains undefined.
Examples of valid and invalid character assignment statements
follow.
variables and arrays in the examples are of character
all
that
Note
data

type.

valid
FILE

'PROG2'

REVOL (1)

'MAR'//'CIA'

LOCA(3:8)

'PLANTS'

TEXT(I,J+1) (2:N-1)

= NAME//X

Invalid

'ABC'

= CHARS

CHARS = 25

(element on left must be a character variable,
array element,

substring

reference)

(expression on right must be of character data
type)

ASSIGNMENT STATEMENTS

ASSIGN
3.4

ASSIGN STATEMENT

integer
The ASSIGN statement assigns a statement label value to ana transfer
specify
to
used
be
then
can
variable
The
variable.
destination in a subsequent assigned GO TO statement (see Section
4.1.3) L4

The ASSIGN statement has the form:
ASSIGN

s TO v

The label of an executable statement in the same program unit
the ASSIGN

statement.

An integer variable.

The ASSIGN statement assigns the statement number to the variable.

is similar to an arithmetic assignment statement, with one exception:
the variable becomes defined for use as a statement label reference
and becomes undefined as an integer variable.

The ASSIGN statement must be executed before the assigned GO TO
(s) in which the assigned variable is to be used. The ASSIGN
statement
t
statemen and the assigned GO TO statement(s) must occur in the same
program unit.
For

example:

ASSIGN

TO NUMBER

100

This statement associates the variable NUMBER with the statement labelg
Arithmetic operations on the variable, as in the followin
100.
statement then become invalid, since a statement label cannot be
altered.

NUMBER = NUMBER+1

The next statement dissociates NUMBER from statement 100, assigns it
an integer value 10, and returns it to its status as an integer
variable.

NUMBER =

The variable NUMBER can no

assigned

(ERROR must have been

defined

longer

used

statement.

Examples of ASSIGN
ASSIGN 10

statements are:

TO NSTART

ASSIGN 99999

TO KSTOP

ASSIGN 250 TO ERROR

integer

variable)

CHAPTER 4
CONTROL STATEMENTS

Statements are normally executed in the order in which they are
However, you may interrupt normal program flow to transfer
written.
control to another section of the program or to a subprogram.

of control from a given point in the program may occur every

Transfer

time that point is reached in the program flow;
a decision made at that point.

or it may be based on

You use the FORTRAN control statements to transfer control to a point
within the same program unit or to another program unit. These
statements also govern iterative processing, suspension of program
execution, and program termination.

The

control

statements

are:

GO TO statements —-- transfer control within a program unit
control,

END
and
ELSE,
THEN,
IF
ELSE
THEN,
IF
statements
of
blocks
execute
ly
conditional
-statements

DO statement -- specifies iterative processing

CONTINUE statement -- transfers control to the next executable

CALL statement —-- invokes a subroutine subprogram

RETURN statement —-- returns control from a subprogram

PAUSE statement -- temporarily suspends program execution

STOP statement —-- terminates program execution

END statement -- marks the end of a program unit

statements —-- conditionally

conditionally execute a statement

transfer

statement

The following sections describe these statements, giving
and

the

calling program unit

examples of

use.

their

forms

CONTROL

STATEMENTS

control

within

GO TO
4.1
GO

GO
TO

STATEMENTS

statements

transferred
or

The

one

set

types

the

statements,

GO TO

same

based

statement

Computed

statement

Assigned

GO TO

statement

Unconditional

program

every
on

the

unit.

time

value

Control

executed,

expression.

are:

statement

Unconditional GO TO Statement

unconditional

statement

The

4.1.1
The

three

transfer

either

every

unconditional
GO

time

statement

executable

label

the

statement.

The unconditional GO
identified
by
the

Examples

control

the

has

the

same

form:

The

executable

transfers

executed.

the

same

TO statement transfers control
specified
1label.
The
1label

statement

the

same

statements

program

unit

program

unit

to
the
statement
must
identify an

the

statement.

are:

GO TO 7734
GO

4.1.2

99999

Computed

The

computed

the

value

list

Statement

statement transfers control to
expression within the statement.

(slist)[,]

statement

based

slist
by

commas.

one

The

more
list

labels
of

labels

executable
is

called

An arithmetic expression in the range 1 to
number of statement labels in the transfer
The

computed

necessary,

converts

statement
the

evaluates

resulting

value

the
to

statements

the

transfer

separated
list.

n
(where
list).
expression

integer

data

is transferred to the statement label in position e
in
list.
For example, if the list contains (30,20,30,40),
of e is 2, control is transferred to statement 20.

type.

the
and

the

and,

Control

transfer
the value

CONTROL STATEMENTS

If the value of e is less than 1, or greater than the number of labels
in the transfer list, control is transferred to the first executable
statement after the computed GO TO.

Examples of computed GO TO statements are:
GO TO

(12,24,36),INDEX

GO TO

(320,330,340,350,360), SITU(J,K)+1

4.1.3

Assigned GO TO Statement

The assigned GO TO statement transfers control to

statement label

The relationship between the
1is represented by a variable.
that
be -established by an
variable and a specific statement label must
the transfer destination can be changed,
Thus,
ASSIGN statement.
depending on the most recently executed ASSIGN statement.

The assigned GO TO statement has the form:
GO TO v[[,](slist)]

An
slist

integer

variable.

statements' separated
A list of one or more labels of executable
slist does not affect statement execution and can be
by commas;
omitted.

The assigned GO TO statement transfers control to the statement
label was most recently assigned to the variable v.

whose

The variable v must be integer data type and must have been assigned a

statement

value

label

ASSIGN

statement

(not an arithmetic

assignment statement) before the GO TO statement is executed.

(s)
The assigned GO TO statement and its associated ASSIGN statement
must exist in the same program unit. Statements to which control is
transferred must be executable statements in the same program unit.
Examples of

assigned GO TO statements are:

ASSIGN 200
GO TO IGO

450

IGO

Equivalent to GO TO 200.

ASSIGN

IBEG

GO TO IBEG, (300,450,1000,25)
Equivalent to GO TO 450.

CONTROL

STATEMENTS

IF
4.2

1IF

STATEMENTS

statements

statement

conditionally
or

block

transfer

control,

statements.

The

three

ELSE

conditionally
types

execute

statements

are:

For

Arithmetic

Logical

statement

Block

each

type,

statement

statements

the

decision

block

expression within

the

4,2.1

The

Arithmetic

arithmetic

statements,

The

statement

arithmetic

(e)

arithmetic

sl,

THEN,

transfer

statements

ELSE,

END

IF)

control

based

the

evaluation

execute

statement.

statement

the

value

statement

s2,

THEN,

the
of

Statement

based

(IF

transfers
of

has

the

control

arithmetic

one

three

expression.

form:

expression.

sl,s2,s3
Labels

executable

statements

the

same

program

All three labels (sl,s2,s3) are
required;
however,
refer to three different statements.
You can use one
refer

The

the

statement

arithmetic

parentheses.

labels

the

Equal

Greater

This
to

statement

100

they
or two

need
not
labels to

statement.

the
one

expression

the

(e)

three

statement

real

variable

follows:

0
IF

evaluates

Control

statements

passes

Label

to:

follow.

50,50,100

transfers
than

the

control

arithmetic

less

list,

after

(THETA-CHI)

first

transfers

is:

than

statement

THETA

statement
then

value

than

Examples

IF
It

transfer

Less

immediately

unit.

only

control

equal
if

THETA

(NUMBER/Z*Z—NUMBER)

to
the

statement
real

greater

variable
than

the

CHI.

Control

passes

CHI.

20,40,20

This statement transfers control
integer variable NUMBER is even;
if the value is odd.

to statement 40 if the value
of
the
it transfers control to statement 20

CONTROL STATEMENTS

Logical

4.2.2

IF Statement

1IF

logical

The

statement.

statement

decision

conditionally

execute

executes

single

FORTRAN

the statement is based on the

value of a logical expression within the logical IF statement.
The logical IF statement has the form:
st

(e)

A logical expression.
st

any
be
The statement can
A complete FORTRAN statement.
,
statement
IF
block
a
,
statement
DO
a
except
statement
e
executabl
an END statement, or another logical IF statement.

expression (e).
(st) is
statement
the
true,
is
n
expressio
the
of
If the value
executed. If the value of the expression is false, control transfers

The logical IF statement first evaluates the logical
to

(st)

is not executed.

Examples of logical

(REF(J,K)

(ENDRUN)

IF statements are:

.OR.

.GT.

.LT.

.NE. HOLD)

GO TO 250

REF(J,K)

= REF (J,K) *(-1.5D0)

CALL EXIT

Block IF Statements

4.2.3

statements

Block IF

statements.

The

The statement

the next executable statement after the logical IF.

four block
IF

THEN

ELSE

END

THEN

conditionally

statements are:

execute

blocks

(or

groups)

CONTROL

These

statements

construct
IF

has

the

are used
form:

STATEMENTS

block

1IF

constructs.

The

block

1IF

(e) THEN
block

ELSE

(e)

THEN

block

ELSE

block
END

logical

sequence

expression.

block
sequence

Figure

4-1

constructs.

zero

called

describes

the

complete

statement

flow

FORTRAN

statements.

This

block.

control

for

four

examples

block

STATEMENTS

CONTROL

Flow of Control.,

Construct

False

IF (e) THEN

block

True

END IF
Ex\ecute
block

Test

False

IF (e) THEN
block

True

ELSE

blocks
END IF

Execute

block 4

block,

IF (eq) THEN
block4

ELSE IF (e5) THEN
blocksy
END IF
'

Execute

block 4

blockoy

IF (eq) THEN

False _

block4

ELSE iF (ep) THEN
block,
ELSE IF (e3) THEN

b|OCk3
ELSE
blockg

Execute

block4

block,

blocks

blockg

END IF

\
Figure

4-1

‘
Examples

of
4-7

Block

Constructs

CONTROL

Each block
statement

IF statement
block.
The

STATEMENTS

except the END IF statement
statement block consists of

following

the

block

statement

block

statement

the

construct.

1IF

block

conditionally executed

based

the

preceding

block

statements.

The

statement

THEN

ELSE

statement

THEN

statement

block

the

begins

following it is executed if
IF THEN statement is true.
The

the

block

the

optional

not

values

value

executed

(but

the

has
all

associated
statements

including)
The

the

statement

logical

statement

The

block

expression

that

the

block

expressions

construct.

logical

value

an
the

the

specifies

logical

expression

in the ELSE IF THEN statement is true and no preceding statement block
in
the
block
IF
construct
was executed.
A block IF construct can
contain any number of ELSE IF THEN statements.
The ELSE statement specifies
preceding
statement
block
No

block

statement

statement.
The

END

statement
After

The ELSE

the

passes

last

construct

terminates

statement

the

may

ELSE

THEN

END

statement

transferred

Section

and

4.2.3.1

block.

4.2.3.3

ELSE

statement

block

except

construct.

the

END

The

END

block

END

1IF

block

executed.

time

the

from within

describes

THEN

cannot

restrictions

Blocks

statement

nested

statement

You

statement

have

statement

labels.

but

control

block

construct.

the
on

label,

statements

examples

block

statement.

control

the

each

label

executed,

following

statement

one

have

describes

statement block
Note
that
you

statement

only

4.2.3.2

FORTRAN

statements

can

its

Statement

executable

ELSE

describes

Section

4.2.3.1

executed

The

Section

the

executable

Consequently,

The

the

required.

statement.

statement is optional.

statement

can

a statement block to be
executed
if
no
in the block IF
construct
was executed.

1in

block

a
IF

can

only

statement
constructs.

constructs.

block

<can

can

transfer

contain

control

out

any

but control cannot
cannot
transfer

be transferred back into the block.
control from one statement block %o

overlap

blocks.

another.
DO

loops

cannot

contains

statement,
each

loop

DO
and

statement

statement,
vice

versa.

must

wholly

4.2.3.2

Block

Examples

THEN

one

statement

the

executes

Form
IF

and

must

you

contained

END

The

also
use

within

simplest

statements;

loops
one

(e)

THEN

with

this

THEN

TOTERR=TOTERR+ABS (ADJU)
END

block

loop's

terminal

statement

blocks,

block.

construct

block.

QUEST=ADJU/FNDVAL
IF

statement

block

(ABS(ADJU).GE.1.0E-6)

the

Example

block
END

When

contain

consists

conditionally

CONTROL STATEMENTS

The statement block consists of all the statements between the IF THEN
the END

and

statements.

(e),
The IF THEN statement first evaluates the 1logical expression
If the value of e is true, the statement block
ABS (ADJU) .GE.1.0E-6.
is executed. If the value of e is false, control transfers to the

next
not

executable

executed.

statement

the

after

the block is

statement;

END

The following example contains a block IF construct with
THEN

ELSE

statement.

Example

Form

(el)

(A.

THEN

GT.

THEN

blockl

F=A-2B

ELSE IF (e2)
block2

ELSE IF (A .GT. B/2.)
D = B/2.

THEN

F=A -

B/2.

END

THEN

Blockl consists of all the statements between the IF THEN and the ELSE
block2 consists of all the statements between the
IF THEN statements;
ELSE

IF THEN and

the END IF

blockl

than B but

If A is greater than B,
If A is not greater

statements.

executed.

greater

than

block2

B/2,

Iis

executed.

neither
B/2,
than
greater
not
If A is not greater than B and A is
next
the
to
directly
transfers
control
executed;
is
block2
nor
blockl
executable statement after the END IF statement.

The following example contains a

construct

block

ELSE

with

statement.

Example

Form

(e) THEN
blockl

(NAME .LT. 'N') THEN
IFRONT = IFRONT + 1

T)
=NAME (1:2)
FRLET (IFRON

ELSE

block2
IF

END

IBACK=IBACK + 1
IF

Blockl consists of all the statements between the IF THEN and ELSE
block2 consists of all the statements between the ELSE
statements;
and

the

END

statements.

If the value of the character variable NAME is less than
is

'N',

blockl

executed.

If the value of NAME is greater

than

egual

'N',

block2

executed.

The following example contains a block IF construct with several
IF THEN statements and an ELSE

statement.

ELSE

CONTROL

Form

STATEMENTS

Example

(el)

THEN

blockl

.GT.

=2B

THEN

F=A-B

ELSE

(e2)

THEN

ELSE

block?2

.GT.

THEN

D =2¢C
F=A-2¢C

ELSE

(e3)

THEN

ELSE

block3

F=A

ELSE

0.0

END

There

the

-1

ELSE

block4
END

.GT.

=12

are four
statements

statement blocks in this example.
Each consists
between the block IF statements listed below.

Block

Delimiting

blockl

block2

First

THEN

Block

and

ELSE

first

and

THEN

second

block3

Second

block4

ELSE

and

END

greater

If A

not

greater

than

but

If A

not

greater

than

greater

than

THEN

and

ELSE

blockl

executed.

greater

but

not

4.2.3.3

Nested

included
nested

1in

block

Block

following

example

block

must

must
not

contains

Form

Constructs

construct

statement ‘-block;

The

statement

block4

than

block

executed.

2%,

block3

construct

can

construct.

But

the

contained

within

blocks.

construct.

Example

(e)

THEN

.LT.

blockl

(e)

THEN
blocka

ELSE

100)

END

ELSE

block?2
IF

THEN
+

(ABS (A-AVG)
.LE.
INAVG = INAVG + 1

5.)

ELSE

blockb

OUTAVG

END

OUTAVG

ELSE

OUTRAN = OUTRAN +
END

completely
statement

nested

INRAN=INRAN

END

block2

executed.

block

another

overlap

than

greater

executed.

ELSE

THEN

If A

all

Statements

ELSE

THEN

CONTROL

STATEMENTS

If A is less than 100, blockl is executed.
Blockl contains
a
nested
block IF construct.
If the absolute value of A minus AVG is less than
or equal to 5, blocka is executed.
If the absolute value of
A
minus
AVG is greater than 5, blockb is executed.

If A is greater than or equal to 100, block2 is executed;
IF construct is not executed because it is not in block2.

the

nested

CONTROL

STATEMENTS

DO
4.3

STATEMENT

The

DO
statement
statements
1in its

specifies
1iterative
processing.
That
range are executed repeatedly a specified

1is,
the
number of

times.
The

statement

s|,]

has

the

form:

v=el,e2[,e3]

The
label
physically

of
an
executable
statement.
follow in the same program unit.

The

statement

must

integer,

el,e2,e3
Arithmetic

real,

double

precision

variable.

expressions.

The variable v is the control variable;
and el, e2, and
e3
are
the
initial,
terminal,
and
increment
parameters, respectively.
If you
omit the increment parameter, a default increment value of 1 is used.
The

terminal statement of a DO
appears
in
the DO statement.
of the following statements:
e

Arithmetic

Any

block

®¢

END

statement

RETURN

IF
IF

statement
statement,

ENDIF

statement

range

statement,

The

statement

first

determine

the

values

initial,
to

except

statement

the

use,

is identified by
the
label
that
terminal statement must not be one

The

statement

The

loop

statement

for

data

includes

including

evaluates
the

terminal,

the

and

the

initial,

and

type

the

statements

terminal

expressions

terminal,

increment
of

all

the

parameters

control

el,

and

that

e2,

increment

are

variable,

statement.
and

converted,
if

parameters.

before

necessary.

The

value of the initial parameter is assigned to
the
control
variable.
The
executable
statements
in
the
range
of
the
DO loop are then
executed

the

(e2)

repeatedly.

increment

must

Conversely,

The

be
if

increment

parameter

greater
e3

(e3)

than

negative,

parameter

(e3)

positive,

equal

must

cannot

to
be

the
less

zero.

the

terminal

initial
than

parameter

parameter
equal

(el).
e2.

CONTROL

STATEMENTS

The number of executions of the DO range,

called the iteration

count,

by:

is given

where
whose
the

[X] (the value of the expression) represents the largest integer
magnitude does not exceed the magnitude of X and whose sign is
as

same

the

sign of X.

If the iteration count is zero or negative, the loop is executed once.

4.3.1

DO Iteration Control

iteration

After each

the

range,

the

following

steps

are

executed:

The value of the increment parameter is

The

If
the
iteration
transfers
to
the

the control

iteration count

statement

If the

for

algebraically

added

variable.

is decremented.

<count
first

another

iteration count

1is
greater
than
zero,
control
executable
statement after the DO

iteration of the

zero,

range.

execution of

the DO statement

terminates.

double
or
Note that if the data type of the control variable is real
the number of iterations of the DO range might not be what
precision,
is expected,

because of rounding errors.

a
using
by
statement
terminate execution of a DO
You can also
range that. transfers control outside the loop.
the
statement within
its
remains defined with
statement
The control variable of the DO
current

value.

if other DO

When execution of a DO loop terminates,

1loops

its

most
next
the
to
outward
transfers
control
statement,
terminal
enclosing DO loop in the DO nesting structure (Section 4.3.2).
If
no
other
DO
loop shares the terminal statement, or if this DO statement
is outermost, control transfers
to
the
first
executable
statement
after

the

terminal

statement.

You cannot alter the value of the control variable within the range of
use the control variable for
can
you
However,
statement.
DO
the
reference as a variable within the range.

increment
and
terminating,
initial,
You can modify the
within the loop without affecting the iteration count.

parameters

long
as
The range of a DO statement can contain other DO statements,
Section 4.3.2
1loops meet certain requirements.
DO
nested
these
as
describes

these

requirements.

from
loop
a
You can transfer control out of a DO loop, but not into
Exceptions to this rule are described in
the program.
1in
elsewhere
Sections 4.3.3

and

4.3.4.

)

CONTROL

Examples of
DO

This

100

iteration control

STATEMENTS

follow.

K=1,50,2

statement

specifies

iterations;

K=49

during

the

final

iterations;

J=-2

during
)

the

final

IVAR=5

during

the

final

iteration.
DO

350

J=50,-2,-2

This
statement
iteration.
DO

IVAR=1,5

This statement
iteration.

This

specifies

NUMBER=5,40,4

invalid

M=2.10

iterations;

statement;

the

statement

label

is missing.

This is an invalid DO statement;
it contains a decimal point
instead
of
a
comma.
This example illustrates a common typing error.
It is
the valid arithmetic assignment statement:
DO40M

4.3.2

Nested

2.10

Loops

A DO loop can contain one or more complete DO loops.
The range of
an
inner
nested DO loop must lie completely within the range of the next
outer loop.
Nested loops can share a terminal statement.
Figure

4-2

illustrates

nested

loops.

Correctly Nested
DO

Incorrectly Nested

Loops

DO 45 K=1,10

Dé 35 L=2,50,2

L 45

CONTINUE

DO 15

| 35 CONTINUE

"15

CONTINUE
DO

4-2

Nested

K=1,10

Dé 25 L=1,20

M=1,20

Figure

Loops

M=1,15

CONTINUE

| 30

CONTINUE

Loops

CONTROL

4.3.3

Control

Within

outer

loop

two

nested
not

the

innermost
transfer

4.3.4
A

that
or

into

loop.
part

the

transfer

has

an
of

and

transfer

control

from

outer

the

same

terminal

statement

only

from

other

transfer

loop

range

that

inner

innermost

contains

inner
to

loop

inner

statement,

within

the

statement

loop,

the

an
loop

that

extended

Thus,

the

statement

govern

transfer

into

the

statement.

The

extended

the

loop

control

range

the
the

the

that

transfer

control

between

rules

range

out

another

statements

outer
the

can

loops

control

loop.

only

Any

statements,

following
l.

you

Loops

you

range

constitutes

because

the

shared

loop.

Range

transfers

executable

The

nested
control

Extended

However,

from

locp

loop,

permitted.

transfer

statement

loop.

can

Transfers

STATEMENTS

and

statement

the

loop

use

range

range of a
variable of the

made

DO
a

control

to
of

the

statement
DO

from

statement

statement.

extended

must

include
the

one
back

all

first

loop.

statement
the

control

extended
to

statement

execution

statement

control

after

returns

destination

returns

if,

not

range:
permitted

range

that

change

the

CONTROL STATEMENTS

Figure 4-3
transfers.

illustrates

valid

and

invalid

Control Transfers

GO TO 20

DO 35 K=1,10

DO 15 L=2,20

DO 35 L=2,20

CONTINUE

A=B+C

D=E/F

DO 35 M=1,15

CONTINUE

GO TO 50

GO TO 40

X=A*D

DO 45 M=1,15
.

40 X=A*D

CONTINUE

L \_35 CONTINUE

Range

A=B+C

Extended

DO 50 K=1,10

GO TO 20

Loop

range

Invalid

valid

extended

D=E/F

GO TO 30

Figure 4-3

GO TO 30

Control Transfers and Extended Range

control

CONTROL

STATEMENTS

CONTINUE
4.4

CONTINUE

The CONTINUE
statement.
loop

when

arithmetic
The

STATEMENT

statement
It

that

IF,

CONTINUE
CONTINUE

transfers

control

1is

wused

primarily

loop

would

otherwise

other

statement

prohibited
has

the

end

control

form:

the

terminal
illegally

statement.

executable

statement
with

of
GO

DO
TO,

CONTROL STATEMENTS

CALL
4.5

STATEMENT

CALL

The CALL statement executes a SUBROUTINE subprogram or other external
It can also specify an argument list for the subroutine.
procedure.
(See Chapter 6 for greater detail on the definition and use of
subroutines.)

form:

The CALL statement has the

CALL s[([all,[al]l...)]
s

procedure;
The name of a subroutine subprogram or other external
a dummy argument associated with a subroutine subprogram or
or
external procedure.

other

An actual argument.

(Section 6.1 describes actual arguments.)

If you specify an argument list, the CALL statement associates the
It
values in the 1list with the dummy arguments in the subroutine.
first executable statement of the
then transfers control to the
subroutine.

and
order,
The arguments in the CALL statement must agree in number,
They can be
the dummy arguments in the subroutine.
type with
data
constants,
references,
substring
variables, arrays, array elements,
specifiers, or
return
alternate
Hollerith constants,
expressions,
An unsubscripted array name in the argument list
subprogram names.
refers to the entire array.

Examples

of CALL

are:

(BASE,3.14159+X,Y,LIMIT,R(LT+2))

CALL CURVE
CALL PNTOUT
CALL

statements

(A,N,'ABCD')

EXIT

CALL MULT

(A,B,*10,*20,C)

The

example

last

illustrates

the

use

statement

label

The asterisk
lists.
argument
statement
CALL
in
identifiers
identifiers.
label
statement
are
*20
that *10 and
indicates
prefixed by asterisks (or ampersands (&)) are
identifiers
Label
See Section 4.6.
called alternate return specifiers.

CONTROL

STATEMENTS

RETURN
4.6

RETURN

The
the

program

RETURN

STATEMENT

statement is
that called

RETURN

used to transfer
the subprogram.

control from a subprogram
It has the form:

[1i]

The optional argument is used to indicate an alternate return from the
subprogram.
It can be specified only in subroutine subprograms.
When
specified, the value of i indicates that the ith alternate
return
in
the
actual
argument
1list
1is
to be taken.
(See the second example
below.) The value of i can be any integer constant or expression,
for
example,

I+J.

When

a RETURN statement is
to
the
calling
program,
reference (see Chapter 6).

subroutine,
statement

the

return
You
use

can
the

RETURN

control

following

is
the

CALL

in a function, control is returned
at the statement that contains the function
When a RETURN statement is executed
in
a

returned
CALL

statement

the

executed

1label

argument

either

statement

that

was

specified

SUBROUTINE

CONVRT
DATA (N),

.GE.

10)

subroutine,

the

ith

alternate

units.

You

cannot

(N,ALPH,DATA,PRNT,K)
PRNT(N)

THEN

DATA (K+2)

executable

the

examples:

DIMENSION

first

list.

use RETURN statements only in subprogram
RETURN i form in function subprograms.
statement

the

initiated

N-(N/10)*N

N/10

DATA

(K+1l)

PRNT

(K+2)

ALPH(DATA (K+2)+1)

PRNT

(K+1)

ALPH(DATA (K+1)+1)

ELSE

PRNT (K+2)
END

ALPH(N+1)

RETURN

END

In this example,
first executable
SUBROUTINE

50
60

IF(Z) 60,
RETURN

RETURN

control is returned
statement following
CHECK

70,

to the
calling
the CALL CONVRT

program
at
statement.

the

(X,Y,*,*,C)

END

This example
subroutine.
return

executable

the first
than
0,

shows
how
alternate
returns
If
the
value
computed
for Z

can
be
is less

included
than 0, a

taken, and
the
<calling
program
continues
at
statement
following
CALL CHECK.
If Z equals
alternate return (RETURN 1) is taken;
and if Z

the

second

alternate

return

(RETURN

taken.

in
a
normal

the
first
however,
is
greater

Control

CONTROL STATEMENTS

returned to the statement specified as the first or second alternate
return argument in the CALL statement argument list. For example,
CALL CHECK(A,B,&10,&20,C)

Thus, RETURN 1 transfers control to statement label 10, and RETURN 2
transfers control to statement label 20. Note that if a subroutine
includes an alternate return that specifies a value either 1less than
or greater than the number of alternate return arguments, control is
returned to the next executable statement after
That is, the alternate returns are ignored.
ensure that the value of i is within the range
arguments.

the CALL statement.
Therefore, you should
of alternate return

CONTROL

STATEMENTS

PAUSE
4.7

PAUSE

STATEMENT

The
PAUSE
statement
displays a message on
The

PAUSE

statement

PAUSE

disp

temporarily
the terminal

has

the

suspends
to permit

program
execution
and
you to take some action.

form:

[disp]

character

constant

decimal

digit

string

digits.

The disp argument is
optional.
The
effect
of
a
PAUSE
statement
depends
on how your program is being executed.
If it is running as a
batch job or detached process, the contents of disp are written to the
system output file but the program is not suspended.

If the program is running in interactive mode, the
contents
of
disp
are
displayed
at
your
terminal,
followed
by the prompt sequence,
indicating that the program is
suspended,
and
you
should
enter
a
command.
For
example,
if
the
following
statement is executed in
interactive mode:
PAUSE

You will

'ERRONEOUS

see

the

ERRONEOUS

RESULT

following

RESULT

DETECTED'

display

the

terminal:

DETECTED

$
If you do not

specify a value

for

disp,

you will

receive

the

following

message:

FORTRAN

You

can

PAUSE

respond

typing

one

CONTINUE

execution

resumes

STOP

eXxecution

DEBUG

execution
program.

the

following
at

the

commands:
executable

statement

terminated

resumes

under

control

the

DEBUG

CONTROL STATEMENTS

STOP
4.8

STOP STATEMENT

The STOP statement terminates program execution.
The STOP statement has the
[disp]

STOP
disp

form:

A character constant or a decimal digit string of 1 to 5 digits.

statement
1If you specify it, the STOP
The disp argument is optional.
displays the contents of disp at your terminal, terminates program
execution, and returns control to the operating system.
Examples of

STOP

statements

STOP

'END OF RUN'

are:

CONTROL STATEMENTS

END
4.9

END STATEMENT

The END statement marks the end of a program unit.
last source line of every program unit.

must

the

The END statement has the form:
END

In a main program,

execution

terminates.

implicitly executed.

control

reaches

the

subprogram,

END

statement,

RETURN

program

statement

1is

CHAPTER

SPECIFICATION STATEMENTS

you
1let
that
statements
Specification statements are nonexecutable
other
define
and
arrays,
and
variables
initialize
and
allocate
characteristics of the symbolic names used in the program.
The

specification

statements are:

statement —-- overrides

the

implied

type

data

IMPLICIT

Type declaration statement -- explicitly defines the data type

DIMENSION statement -- defines the number of dimensions in
array and the number of elements in each dimension

COMMON statement -- defines one or more

areas

symbolic

names

specified

symbolic names

contiguous

storage

EQUIVALENCE statement -- associates two or more entities

EXTERNAL statement -- defines the specified symbolic names

DATA statement -- assigns initial values to variables,
and array elements before program execution

the

same

external

storage

with

location

procedure

names

PARAMETER statement -- assigns a symbolic name to

arrays,
constant

value

PROGRAM statement -- assigns a symbolic name to a main program

BLOCK DATA statement -- establishes and defines common
blocks
and
assigns
initial
values
to
entities contained in those

unit

common

blocks

The following sections detail these statements,
and

examples of

use.

giving

their

forms,

SPECIFICATION

IMPLICIT
5.1

STATEMENTS

IMPLICIT

STATEMENT

By default,

all names beginning with
the
1letters
I
through
N
are
assumed
to
be
integer
data, and all names beginning with any other
letter are assumed to
be
real.
The
IMPLICIT
statement
overrides
implied data typing of symbolic names.

The

IMPLICIT

statement

IMPLICIT

typ

One

has

the

form:

typ(al,al...){[,typ(al,al...)]...

the

data

type

specifiers.

(See

Chapter

Table

2-2.)

An alphabetic specification in either of the general forms:
cl-c2,
where
¢
1is
an
alphabetic
<character.
The latter
specifies a range of letters, from
¢l
through
<¢2,
which

occur

alphabetical

¢ or
form
must

order.

When you specify typ as CHARACTER*len, len specifies
the
1length
for
character
data
type.
Len
1is
an
unsigned
integer constant or an
integer constant expression enclosed in parentheses, and
must
be
in
the range 1 through 32767.
The

IMPLICIT statement assigns the specified data type to all symbolic
names
that
begin
with
any
specified
1letter,
or
any letter in a
specified range, and which have no
explicit
data
type
declaration.
For

example:
IMPLICIT

INTEGER

IMPLICIT

REAL

(I,J,K,L,M,N)

(A-H,

0O-1%)

These statements
specifications.

represent

You

IMPLICIT

cannot

Examples

label

IMPLICIT

the default

the

statements.

statements

are:

PRECISION

(D)

IMPLICIT

DOUBLE

IMPLICIT
IMPLICIT

COMPLEX (S,Y), LOGICAL*1
CHARACTER*32 (T-V)

IMPLICIT

CHARACTER*2

(W)

(L,A-C)

absence of

any data

type

SPECIFICATION STATEMENTS

Type Declaration
5.2

TYPE DECLARATION STATEMENTS

Type declaration statements explicitly define the data type of
There are two forms of type declaration
specified symbolic names.
type
character
and
declarations,
type
numeric
statements:
declarations.

The following rules apply to type declaration statements:

5.2.1

declaration

statements

must

precede

executable

all

Type

You can declare the data type of a symbolic name only once.

You cannot label type declaration statements.

statements

Numeric Type Declaration Statements

Numeric type declaration statements have the form:
typ vi,vil...

typ

Any data type specifier except CHARACTER.

The symbolic name of a variable,
function subprogram,

array,

statement

function

an array declarator.

You can use a numeric data type declaration statement to define arrays
by including array declarators (see Section 2.5.1) in the 1list.

A symbolic name can be followed by a data type length specifier of the
form

*s,

where

s is one of the acceptable lengths for the data type

being declared. Such a specification overrides the length attribute
that the statement implies, and assigns a new length to the specified
If you specify both a data type length specifier and
item.
declarator, the data type length specifier goes first.
Examples of numeric type declaration statements are:
INTEGER COUNT,
REAL

MATRIX(4,4),

SUM

MAN,IABS

LOGICAL

SWITCH

INTEGER*2 I, J, K, M12*4, Q, IVEC*4(10)
REAL*8 WX1, WXZ, WX3*4, WX5, WX6*8

array

SPECIFICATION

5.2.2

Character

type

Type

declaration

CHARACTER([*1len]

The

Declaration

symbolic

array

unsigned

STATEMENTS

Statements

statements

have

the

form:

v([*len][,v[*1len]]...

name

variable,

array,

function

subprogram,

declarator.

len
integer
constant,
1in parentheses, or an
of len specifies
the

enclosed
The value

an
integer
constant
expression
asterisk enclosed in parentheses.
1length

the

character

data

elements.
If you specify CHARACTER*len, len is
for
that
1list.
If
an
item
in

the default length
that
1list does not

specification
have a length

specification, the item's length is len.
But if an item does
have
length
specification,
it
overrides
the default length specified

CHARACTER*1len.

length

specifies
have this

specification

passed

length

If you do not
specification

asterisk

Only

specification

specify a
must
be

specification

length.

dummy

(see

invalid.

You

example,

argument

Chapter

length, a length
in
the range 1

1is

(for

of
to

a
in

CHARACTER*
(*))

function

name

can

6).
1 is assumed.
The
32767.
Note that a

can

wuse

character

length
length

type

declaration
statement to define arrays by including array declarators
(see Section 2.5.1) in
the
1list.
If
you
specify
both
an
array
declarator and a length, the array declarator goes first.
Examples

character

CHARACTER*32

type

declaration

NAMES(100),

statements

SOCSEC(100)*9,

follow.

NAMETY

This statement specifies an array NAMES
comprising
100
32-character
elements,
an
array SOCSEC comprising 100 9-character elements, and a
variable NAMETY, 32 characters long.
PARAMETER

LENGTH=4

CHARACTER* (4+LENGTH)

This
(The

statement specifies
PARAMETER statement
SUBROUTINE

LAST,

two 8-~character variables,
LAST
is described in Section 5.8.)

statement

elements and a
defined by the

and

FIRST.

S1 (BUBBLE)

. CHARACTER LETTER(26),

This

FIRST

specifies

BUBBLE

* (*)

array LETTER comprising

dummy argument BUBBLE,
calling program).

CHARACTER*16

which

has

passed

one-character
length

(it

BIGCHR* (30000%*2) ,QUEST* (5*INT (A))

This statement is invalid;
the value
large
and
the
1length specifier for
expression.

specified
for
QUEST is not an

BIGCHR
integer

is
too
constant

SPECIFICATION STATEMENTS

DIMENSION
5.3

DIMENSION STATEMENT

The DIMENSION statement defines the number of dimensions in
and the number of elements in each dimension.

array

The DIMENSION statement has the form:
DIMENSION a(d)[,a(d)]...
a(d)

An array declarator

(see Section 2.5.1).

The symbolic name of an array.

A dimension declarator.

number of storage elements to each
The DIMENSION statement allocates a storag
e element is assigned to each
One
array named in the statement.
of each storage element
length
The
ion.
dimens
each
in
t
array elemen
The total number of
array.
the
of
is determined by the data type
product of all
the
to
equal
is
array
an
to
ed
assign
storage elements
array. For
dimension
example:

declarators

the

array declarator for that

DIMENSION ARRAY (4,4), MATRIX(5,5,5)

s of 4 Dbytes
This statement defines ARRAY as having 16 real element
ts of 4 bytes
elemen
r
intege
125
having
as
MATRIX
s
define
and
each,
each.

and effect as the DIMENSION
The VIRTUAL statement has the same form ibilit
y with other versions of

statement.

FORTRAN.

is provided for compat

For further information on arrays and on storing array

elements,

Section 2.5.

You can also use array declarators
statements.

1in

type

declaration

and

see

COMMON

However, in each program unit, you can use an array name

in only one array declarator.

You cannot label DIMENSION statements.
Examples of DIMENSION statements are:
DIMENSION BUD(12,24,10)

DIMENSION X(5,5,5),Y(4,85),2(100)
DIMENSION MARK (4,4,4,4)

SPECIFICATION

STATEMENTS

COMMON
5.4
A

COMMON

STATEMENT

statement

storage.

defines

symbolic

one

name

omit

contiguous

identifies

each

areas

block;

(blocks)

however,

you

a symbolic name for one block in a program unit.
This
block
blank
common
block.
COMMON statements also define the order
variables and arrays in each common block.
the

The

COMMON

statement

COMMON

has

[/[cbl/]

the

can

is
of

form:

nlist[[,]/[cb]l/

nlist]...

cb
A

symbolic

the

list

name,

first

called

blank,

variable

names,

common
you

block

can

name.

omit

the

can

first

pair

be
of

blank.
slashes.

nlist
of

separated
A

common

name

cannot

the

array

names,

commas.

block

However,

entry

can

and

array

the

same

variable

have

the

same

name

executable

program.

function,

declarators

array

When you declare common blocks of the same name in
different
units,
these
blocks all share the same storage area when the

units

are

combined

into

You

can

have

only

you

can

have

several

one

The

entities

character

data

and

character

data.

Entities
Thus,

are

blank

named

nlist

assigned

executable
common

should

must

either

all

type.

common

block

program

COMMON
another

executable

program,

storage

common

numeric
cannot

blocks

data

contain

type
both

one-for-one

but

all

numeric

basis.

unit.

For

example,

one

program

unit

contains

the

CENTS
program

INTEGER*2

COMMON

unit

contains

the

statement:

MONEY

When these program units are
incorrect results may occur,
is made to correspond to the
CENTS.

entities assigned by a COMMON statement in one program unit
agree in data type with entities placed in a
common
block
by

statement:

and

program
program

common blocks.

the

another

program.

block

hame.

subroutine,

combined
into
an executable
program,
because the 2-byte integer variable MONEY
high-order 2 bytes of the
real
variable

SPECIFICATION

example

Main

COMMON

the

COMMON

statement

STATEMENTS

follows.

Program

Subprogram

HEAT,X/BLK1/KILO,Q

SUBROUTINE

COMMON

CALL

FIGURE

/BLK1/LIMA,R/ /ALFA,BET

RETURN

END

The COMMON statement in
common
block, and puts

the main
KILO and

program puts
Q in a named

HEAT and X in
common block,

the
blank
BLK1l.
The

COMMON statement in the subroutine makes ALFA and
BET
correspond
to
HEAT
and X in the blank common block, and makes LIMA and R correspond
to

KILO

and

You

can

use

array

arrays.

BLKI1.
declarators

the

COMMON

statement

define

SPECIFICATION

STATEMENTS

EQUIVALENCE
5.5

STATEMENT

EQUIVALENCE

The EQUIVALENCE statement partially or totally associates two or

entities in the same program unit with the same storage location.

The EQUIVALENCE

statement has the form:

EQUIVALENCE
nlist

(nlist)

[,(nlist)]...

A list
of
variables,
array
elements,
arrays,
and
character
You must specify at
separated by commas.
references,
substring
least two of these entities in each list.

entities
the
of
all
allocates
statement
The EQUIVALENCE
parenthesized list beginning at the same storage location.

one

subscript
a
in
expression
In an EQUIVALENCE statement, each
substring reference must be an integer constant expression.

of
or
type
numeric data
of
either
The entities in nlist must be
You cannot make numeric entities and character
type.
data
character
entities equivalent.

You can equivalence variables of different numeric data types.

If you

multiple components of one data type can share storage with a
do,
you
if
example,
For
single component of a higher-ranked data type.
make an integer variable equivalent to a complex variable, the integer
variable shares storage with the real part of the complex variable.
Examples of EQUIVALENCE

statements

follow.

DOUBLE PRECISION DVAR
IARR(4)
INTEGER*2

EQUIVALENCE

(DVAR,IARR(1))

integer
the
elements of
four
This EQUIVALENCE statement makes the
IARR occupy the same storage as the double-precision variable
array
DVAR.

CHARACTER KEY*16, STAR*10
EQUIVALENCE (KEY,STAR)

This EQUIVALENCE statement makes the first character of the
character
variables KEY and STAR share the same storage location.
The character
variable STAR is equivalent to the substring KEY (1:10).

5.5.1

Making Arrays Equivalent

an element of
When you make an element of one array eguivalent to
sets equivalences
also
statement
EQUIVALENCE
the
array,
another
the
1f
Thus,
between the corresponding elements of the two arrays.

first
elements
of
two
equal-sized arrays are made eguivalent, both
arrays share the same storage
space.
If
the
third
element
of
a
seven-element array is made equivalent to the first element of another

array, the last five elements of the first
five

elements of

the

second

array.

array

overlap

the

first

SPECIFICATION STATEMENTS

same

assign

You must not attempt to use the EQUIVALENCE statement

You

For

also must not attempt to assign memory locations
inconsistent

way

that

with

the

normal linear storage of array elements.

element

another

example, you cannot make the first element of one array equivalent
first

the

array,

equivalence between the second element of

and

the

then

first

attempt to set an

array

and

the

array.

the other

sixth element of
For

the

storage location to two or more elements of the same array.

example:

DIMENSION TABLE (2,2), TRIPLE (2,2,2)
EQUIVALENCE (TABLE(2,2), TRIPLE(1,2,2))

As a result of these statements, the entire array TABLE shares part of
Figure 5-1 shows how
space allocated to array TRIPLE.
storage
the
these

statements align the arrays.

Array TABLE

Array TRIPLE

Array
Element

Element
Number

TRIPLE(1,1,1)
TRIPLE(2,1,1)
TRIPLE(1,2,1)

1
2
3

TRIPLE(2,2,2)

TRIPLE(2,2,1)
TRIPLE(1,1,2)
TRIPLE(2,1,2)
TRIPLE(1,2,2)

Figure

5-1

4
5
6
7

Array
Element

Element
Number

TABLE (1,1)
TABLE (2,1)
TABLE (1,2)
TABLE (2,2)

1
2
3
4

Equivalence of Array Storage

The following statements also align the two arrays as shown in

Figure

5-1:

EQUIVALENCE
EQUIVALENCE

(TABLE,TRIPLE(2,2,1))
(TRIPLE(1,1,2), TABLE(2,1))

Similarly, you can make arrays equivalent with nonunity lower bounds.
an array defined as A(2:3,4) is a sequence of eight
For example,
A reference to A(2,2) refers to the third element 1in the
values.
To make array A(2:3,4) share storage with array B(2:4,4),
sequence.

you can

use

the

EQUIVALENCE

statement:
(A(3,4),

B(2,4))

The entire array A shares part of the storage space allocated to array
Figure 5-2 shows how these statements align the arrays.
B.

SPECIFICATION STATEMENTS

Array

Array
Element

Element
Number

B(2,1)
B(3,1)

1
2

B(4,1)
B(2,2)
B(3,2)
B(4,2)
B(2,3)
B(3,3)
B(4,3)
B(2,4)

3
4
5
6
7
8
9
10

B(3,4)
B(4,4)

11
12

Figure 5-2
The

following

Array
Element

Element
Number

A(2,1)
A(3,1)
A(2,2)
A(3,2)
A(2,3)
A(3,3)
A(2,4)
A(3,4)

1
2
3
4
5
6
7
8

Equivalence of Arrays with Nonunity Lower Bounds
statements

also align

EQUIVALENCE

(A,B(4,1))

EQUIVALENCE

(B(3,2),

the

arrays

shown

in Figure

5-2:

A(2,2))

In the EQUIVALENCE statement only, you can identify an
array
element
with
a
single
subscript
(that is, the linear element number), even
though the
array
was
defined
as
a
multidimensional
array.
For
example,
the
following
statements
align the two arrays as shown in
Figure 5-1:
DIMENSION TABLE (2,2),
EQUIVALENCE (TABLE(4),

5.5.2

Making

Substrings

TRIPLE (2,2,2)
TRIPLE(7))

Equivalent

When you make one character substring equivalent to another
character
substring,
the
EQUIVALENCE
statement also sets equivalences between
the other corresponding characters in the character entities.
For

example:

CHARACTER NAME*16 ID*9
EQUIVALENCE (NAME(10:13),

ID(2:5))

SPECIFICATION STATEMENTS

As a result of these statements, the character variables NAME
share space as illustrated in Figure 5-3.

and

1ID

NAME
Character

ID
Character

WIN|—

Position

QNP

o
-—

—
-—

(43}

—_
ey

-d

Figure 5-3

—

QioivN|loOi|dIWIN|—

Position

Equivalence of Substrings

The following statement also aligns the arrays as shown in Figure 5-3:
EQUIVALENCE

(NAME(9:9),ID(1:1))

If the character substring references are
equivalences
sets
statement
EQUIVALENCE

corresponding characters in the complete arrays.

Character elements of arrays can overlap at
For

any

the
array elements,
other
the
between

character

position.

example:

CHARACTER FIELDS®100)*4, STAR(5)*5
EQUIVALENCE (FIELDS(1l) (2:4), STAR(2) (3:5))

As a result of these statements, the character arrays FIELDS and
share storage space as shown in Figure 5-4.

STAR

SPECIFICATION

STATEMENTS

STAR
Character
Position

Subscript

2
3

FIELDS
Subscript

Character
Position

5
1

1
2
3
4

1
2

A/ N\

100

2
3
4

Figure

5-4

Equivalence of

Character Arrays

You cannot use the EQUIVALENCE statement to assign
the
same
storage
location
to
two or more substrings that start at different character
positions in the same character variable or character array.

You also
cannot
use
the
EQUIVALENCE
statement
to
assign
memory
locations in a way that is inconsistent with the normal linear storage
of

character

variables and

arrays.

SPECIFICATION STATEMENTS

5.5.3

EQUIVALENCE and COMMON Interaction

When you make components equivalent to entities
block,

the

boundaries.

previously

stored

common

original
it s
beyond
extended
can
block
common
the
of
element
1 ast
But it can only extend beyond the
You cannot extend the common
block.
common
established

block in such-a way as to place the extended portion before the
elemént

the

existing

common

block.

first

following examples show

The

valid and invalid extensions of the common block:
Vvalid

A(l) | A(2) | A(3) | A(4)

DIMENSION A(4) ,B(6)
COMMON A

EQUIVALENCE

B(1l)

(A(2),B(1l))
N

B(2)

B(3)

B(4)

Existing
Common

B(5)

B(6)

Extended
Portion

Invalid
A(l) | A(2) | A(3) | A(4)

DIMENSION A (4) ,B(6)
COMMON A

EQUIVALENCE

(A(2),B(3))

B(1) | B(2) | B(3) | B(4) | B(5) | B(6)

\\/_J\

Extended
Portion

Existing Common

J\\/'_J

Extended
Portion

If you assign two components to common blocks, you cannot make them
Note also that character data can be
equivalent to each other.
equivalenced only with character data.

SPECIFICATION

STATEMENTS

EXTERNAL
5.6
The

EXTERNAL
EXTERNAL

arguments
An

The

statement

external

subprogram

STATEMENT

other

procedure

EXTERNAL

can

FORTRAN

statement

EXTERNAL

[*]v

EXTERNAL

external

statement

procedure

argument

name

library
has

the

use

external

procedure

names

user-supplied

function

actual

subroutine

function.
form:

of a subprogram or
subprogram name.
declares

name.

that

Such

subprogram;

you

[,[*]v]...

The symbolic name
associated with a
The

lets

subprograms.

the

function

each
name

subprogram

reference

the

name

can

then

can

use

CALL

dummy

the
appear

the

argument

1list

an actual
associated dummy

statement.

If an asterisk
user-supplied
Use

the

(*) precedes a name in the list, the name identifies
a
function or subprogram, not a FORTRAN library function.
asterisk only when a user-supplied function or subprogram
has

same

name

additional
Note,

however,

(such

as,

name.
An

information

FORTRAN

library

FORTRAN

function.

library

(See

6.3

for

that a complete function reference used as an
argument
SUBR(A,SQRT(B),C)) represents a value, not a subprogram
function name need not be defined in an EXTERNAL statement.

CALL

The

example

Main
EXTERNAL

the

EXTERNAL

statement

follows.

Program
SIN,COS,*TAN,SINDEG

Subprograms
SUBROUTINE
Y

TRIG

(X,F,Y)

F(X)

RETURN

CALL

TRIG

(ANGLE,SIN,SINE)

CALL

TRIG

(ANGLE,COS,COSINE)

CALL

TRIG

(ANGLE,TAN,TANGNT)

END

FUNCTION
.

TAN

SIN(X)

(X)
/

COS(X)

RETURN

CALL

Section

functions.)

END

TRIG

(ANGLED,SINDEG,SINE)

FUNCTION

SINDEG

SIN

RETURN

END

(X)

(X*3.14159/180)

(

SPECIFICATION STATEMENTS

The CALL statements pass the name of a function to the subroutine
The function reference F(X) subsequently invokes the function
TRIG.
in the second statement of TRIG. Depending on which CALL statement
invoked TRIG, the second statement 1is equivalent +to one of the
following:

Y
Y
Y
Y

= SIN(X)
= COS(X)
= TAN(X)
= SINDEG (X)

The functions SIN and COS are examples of trigonometric functions
supplied in the FORTRAN library. The function TAN is also supplied in
the library. But the asterisk in the EXTERNAL statement specifies
that the user-supplied function be wused, instead of the library
function. The function SINDEG is also a user-supplied function.
Because no library function has the same name, no asterisk is
required.

SPECIFICATION

STATEMENTS

DATA
5.7
The

DATA
DATA

STATEMENT
statement

elements

before

The

statement

DATA
DATA

assigns

1initial

values

variables

and

array

program execution.
has

the

form:

nlist/clist/[[,]lnlist/clist/]...

nlist
A

list

names

one

variable

character

names,

substring

array

names,

Subscript expressions and
expressions
must be integer constant expressions.

names,

array

separated

element

substring

commas.

references

clist
A

list

constants;

clist

constants

have

one

the

following

forms:

value
n

value

Used when you specify clist as n * value.
Specifies
the
number
of
times the same value is to be assigned to successive entities
in the associated nlist.
The value of n is a
nonzero,
unsigned
integer constant or the symbolic name of an integer constant.
The

DATA
entities
order as

The

statement

1in
they

number

entities

assigns the constant values in each
<c¢list
to
the
the
preceding nlist.
Values are assigned one by one in
appear, from left to right.

constants

must

the preceding

correspond

exactly

When an unsubscripted array name appears in a DATA
are
assigned to every element of that array.
The
list must therefore contain enough values to f£ill
elements

are

filled

If both the constant
have
numeric
data
rules:
®

The

type
®¢

constant

When an
variable

the

order

value

types,

value

the
the

variable

octal
or
or
array

subscript

When a Hollerith
numeric
variable
characters

that

clist

and the
conversion is

converted,

being

array

right with spaces.
than can be stored,

entity

based

necessary,

the

nlist
following

the

data

hexadecimal
constant
is
assigned
to
a
element
that is longer than 4 bytes, the

or
or

can

statement,
values
associated constant
the
array.
Array

initialized.

character
numeric

assigned

element,

variable
or
is truncated

a
of

depends

the

If the constant
the constant is

5-16

1If the
constant

constant
1is
assigned
to
array
element,
the number

the
component
(see Table 2-2).
If the
constant contains fewer characters than

variable

number

progression.

leftmost digits are initialized to zero.
array
element is less than 4 bytes, the
on the left.
e

the

nlist.

the

data

type

Hollerith or character
the
capacity
of
the

constant

contains
truncated

extended

the

more
characters
on the right.

SPECIFICATION STATEMENTS

If the constant value in the clist and the entity in the nlist are
both character data type, the conversion is based on the following
rules:

‘

If the constant contains fewer bytes than the length of the
entity, the rightmost character positions of the entity are

initialized with

spaces.

If the constant contains more bytes than

the

length

the

entity, the character constant is truncated on the right.

If the constant value is numeric data type and the entity in the nlist
is

character

data

restrictions:

these

type, the constant and the entity must conform to
:

The character entity must have a length of 1 character.

The

constant

must

integer,

octal,

hexadecimal

constant and must have a value in the range 0 through 255.

When the constant and the entity conform to

these

restrictions,

the

entity 1is initialized with the character that has the ASCII code
specified by the constant. This permits a character entity to Dbe
initialized to any 8-bit ASCII code.

An example of the DATA statement follows.
INTEGER A (10)

CHARACTER BELL,TAB,LF,FF, STARS*6
DATA A,STARS / 10%0, '****!

DATA BELL,TAB,LF,FF /7,9,10,12/

The DATA statements assign 0 to all ten elements of array A, and 4
asterisks followed by 2 spaces to the character variable STARS. ASCII
control character codes are assigned to the character variables

TAB,

LF,

FF.

BELL,

SPECIFICATION

STATEMENTS

PARAMETER
5.8

PARAMETER STATEMENT

The

PARAMETER

statement

assigns

The

PARAMETER

statement

has

the

form:

[,p=c]

...

PARAMETER

p=c

symbolic

name.

constant,

the

constant

name

(p)
or

becomes a
constant

A compile-time
which:
Each

or
®

constant

operand

expression

constant,

a compile-time

Each

constant,

operand

constant.

constant

the

symbolic

arithmetic

name

Each

operator

is valid

only

integer,

+,-,*,/,

the

real

a symbolic name is defined
position in which a constant is
the constant were written there

The

symbolic

constant.

exponent

Once

double

operator.

type

precision

The

constant,

expression.

value
valid

expression

type.

compile~time

constant and is defined as the
expression
(c);
c can be any

constant.

name

expression.

Each symbolic name
of
the
constant
FORTRAN

symbolic

data

operator

integer.

as a constant, it can
appear
in
any
allowed.
The effect is the same as if
instead of the symbolic name.

name of a
But it can

constant
appear as

cannot
a real

name

PARAMETER

appear
as
part
or imaginary part

of
of a

another
complex

constant.

You

can

the

symbolic

name

use
can

symbolic

name's

defined

same

program

unit.

The

symbolic

name

corresponding
constant's

data

type

only

constant

name

does

corresponding

not

parameter

once

constant

constant
affect

that

PARAMETER

assumes

expression.
its

constant

statement

data

statements

the

The

only

program

data

initial

identify

unit.

Such

within

the

type

1letter

rof

type.

type

You cannot specify
declaration statement.

Examples
PARAMETER

PI=3.1415927,

PARAMETER

PIOV2=PI/2,

DPIOV2=DPI/2

PARAMETER

FLAG=.TRUE.,

LONGNAME='A

DPI=3.141592653589793238D0

STRING

CHARACTERS'

its

the
the

SPECIFICATION STATEMENTS

PROGRAM
5.9

PROGRAM STATEMENT

The PROGRAM statement assigns a symbolic name to a main program unit.
The PROGRAM statement has the form:
PROGRAM
nam

nam

symbolic

name.

The PROGRAM statement is optional. If you use it, it must be the
first statement in the main program. The symbolic name must not be
the name of any entity within the main program. It also must not be
the same as the name of any subprogram, entry, or common block in the
same executable program.

SPECIFICATION

STATEMENTS

BLOCK DATA
5.10

BLOCK

DATA

STATEMENT

The BLOCK DATA
statement,
followed
by
a
series
of
specification
statements,
assigns initial values to entities in named common blocks
and, at the same time, establishes and defines these blocks.
The

BLOCK

DATA

BLOCK

statement

DATA

[nam]

symbolic

name.

has

the

form:

nam

A
You

can

use

EQUIVALENCE,

type
and

declaration,

DATA

statements

IMPLICIT,

following

The

specification statements that
follow
the
establish
and
define
common
blocks, assign
these blocks, and assign initial values to the

BLOCK

DATA

statement

must

BLOCK

DATA

statement

and

BLOCK

—If

you

DATA
use

program

BLOCK

unit

DATA

have

1its

comprise
a
special
kind
BLOCK DATA program unit is
A

not

associated

not

contain

statement

DATA

statement
arrays to
arrays.

label.

specification

any

COMMON,

statement.

BLOCK
DATA
variables and
variables and

statement

of
program unit.
an END statement.
must

DIMENSION,

BLOCK

The

last

executable

initialize

statements

statement

any

statements.
entity

stateients to establish the entire block, even though some of the

labeled

common

entities

1in

block,

the

same

BLOCK

than

one

common

example
BLOCK

LOGICAL
DOUBLE

must

not

block

DATA

program

provide

appear

unit

complete

a DATA

define

set

specification

statement.

initial

values

You

can

for

block.

BLOCK

DATA

INTEGER

you

DATA

program

unit follows.

BLKDAT

S,X

T,W
PRECISION

DIMENSION

R(3)

COMMON /AREAl/R,S,T,U/AREA2/W,X,Y
DATA R/1.0,2*%2.0/,T/.FALSE./,U/0.214537D-7/,W/.TRUE./,Y/3.5/
END

use

CHAPTER 6
SUBPROGRAMS

from another
Subprograms are program units that can be invoked tion
on behalf
computa
y-used
commonl
some
program, usually to perform
of

the other program.

Subprograms are either user-written, or supplied as part of the VAX-1ll1
User-written subprograms include:

FORTRAN IV-PLUS system.
@

Arithmetic statement functions

Functions

Subroutines

Subprograms supplied with the FORTRAN system include:
®

Processor-defined functions

Generic functions

Character functions

known
Generally the program that invokes the subprogram passes values,
actual
the
wuses
which
am,
subprogr
the
to
s,
argument
actual
as
arguments to compute the results.

6.1

SUBPROGRAM ARGUMENTS

arguments.
Subprogram arguments are either dummy arguments or actualram.
Actual
subprog
the
define
you
when
ed
specifi
are
ts
Dummy argumen
are
and
ram,
subprog
the
arguments are specified when you invoke

associated

with

the

corresponding

arguments when control is

dummy

transferred to the subprogram. This means that each dummy argument
takes on the value of the corresponding actual argument; and any

the

dummy

argument

also

assigned

value

assigned

6.1.1

Actual Argument and Dummy Argument Association

corresponding actual argument.

the

ts.
There must be agreement between actual arguments and dummy argumen
the
with
type
data
and
number
order,
in
agree
Actual arguments must
can
ts
argumen
Actual
ted.
associa
are
they
which
with
ts
argumen
dummy
expressions, arrays, array elements,
variables,
constants,
be
character substrings, alternate return specifiers, or subprogram

ions,
The dummy arguments specified in subprogram definitripted
names.
unsubsc
as
appear
s,
argument
actual
onding
corresp
nting
represe
variable names.

SUBPROGRAMS

Although

dummy
arguments
are
not
actual
variables,
arrays,
or
subprograms,
each
dummy argument may be declared as though
it were a

variable,
A

dummy

array,

argument

actual

declared

argument

type.

array

that

Section

subprogram.

that

a dummy argument
is

the

actual

6.1.1.1)

subprogram.

an
an

array
array

array,

argument.

process

can
or

You

associated

only

array

element

the

same

larger

than

the

arrays

(see

it must

can

adjustable

different

use

sized

arrays

with

an
data

single

The

length of a dummy argument of type character
must not
be
greater
than
the
1length of its associated actual argument.
Note that if the
character dummy argument's length is specifi
ed
as
*(*),
the
length
used
1is
exactly the length of the associated actual
argument.
(This
is known as a passed length character argumen
t.
See Section 6.1.1.2.)

Section

6.1.1.3

constants

describes

actual

the

Section

6.1.1.4

6.1.1.1

Adjustable Arrays

subprograms.
the

describes

The

reference

2.5.1)

dummy

use

character

arguments.

alternate

return

- Adjustable

dimensions

arrays

the actual arguments, variables
specified in the array declarator.

The size of
size of the

in the following example
two-dimensional array.
Note the

control

the

iteration.

FUNCTION

SUM(A,M,N)

DIMENSION

SUM

argument
actual

common

the

use

the

calls

SUM:

A(M,N)

=1 'N

=1 M

SUM =

computes

SUM + A(I,J)

RETURN
END

The

arguments

are

following
DIMENSION

statements
A1(10,35),

SUM1
SUM2

=
=

SUM(A1,10,35)
SUM(A2,3,56)

SUM3

SUM(A1,10,10)

are

sample

A2(3,56)

determined

in
Section

that

are

used in the
array
argument
and each

blocks,

0.0

Hollerith

declarator
must
have
a
evaluated using the values

the adjustable array must be less than
array that is its corresponding actual

function

dummy

to the subprogram.
The array
declarator
(see
an adjustable array can contain integer
variables
arguments or variables in a common block.

are

array

for

variable in a common block used in an array
defined value.
The dimension declarator is

The

and

arguments.

adjustable

When the subprogram is entered, each dummy
declarator
must
be
associated
with
an

constants

and

or
equal
argument.
sum of

dummy

constants

the

the elements
arguments M and

SUBPROGRAMS

The upper and lower dimension bound values are
determined
once
each
time
a
subprogram is entered.
These values do not change during the
execution of that subprogram even if the values of variables contained
in the array declaration are changed.
For example:
DIMENSION

CALL

ARRAY

(11,5)

SUB (ARRAY,L,M)

END

SUBROUTINE

SUB(X,I,J)

DIMENSION

X(-I/2:1/2,3J)

END

this

entry

that

example,
to

the

adjustable

subroutine

SUB.

declaration.

adjustable

currently

array

array X

The

undefined

associated

with

declared

assignments

actual

as
and

dummy

argument

array,

X(-4:4,5)
J

not

array
or

1is

any

affect

not
the

variables
in
the
adjustable
array
declarator
are
not
currently
associated with an actual argument or are not in a common block.
Note
that argument association is not retained between one reference
to
a
subprogram

and

the

SUBROUTINE
DIMENSION

A(I)

reference

that

subprogram.

For

example:

S(A,I,J)
A(I)

RETURN

ENTRY

A(I)

(I,A,K,L)

= A(I)

RETURN

END

this

example,

real

array with

statement

sets

B(2)=3:

DIMENSION

The

following
CALL

The

S(B,2,3)

6.1.1.2

Passed
must

the

For

passed

B (5)

the

length

the

Character

dummy

Arguments

argument.

actual

argument.

character

string,

length

passed

specification

use

(see

A passed
length

length

character

argument

asterisk

Section

enclosed

6.2.2.2).

transfers to a subprogram, each passed
length
argument
must
be associated with a character actual
length of the dummy argument is the length of the actual

dummy
The

control

Length

length

parentheses

When

increments

S1(5,B,3,2)

argument

has

elements:

B (10)

statement

CALL

character
argument.
argument.

SUBPROGRAMS

The length
length.
A character array dummy argument can have passed
elements in
the
of
length
the
is
nt
argume
dummy
of each element in the
the array declarator
the actual argument. The passed length and
character array. A
length
passed
together determine the size of the
passed length character array can also be an adjustable array.

gram uses a passed length
The following example of a function subprothe
position of the character
character argument. The function finds
length of the passed
the
uses
1t
value;
code
with the highest ASCII
ion. (Note that the
iterat
the
length character argument to control
the length of the
ine
determ
to
used
is
LEN
on
functi
processor—-defined
LEN function.)
the
of
argument. See Section 6.3.4 for a description
INTEGER FUNCTION ICMAX (CVAR)

(*) CVAR
CHARACTER*
ICMAX =

DO 10 I = 1, LEN(CVAR)

10 IF (CVAR(I:I)

.GT. CVAR (ICMAX:ICMAX)) ICMAX=I

RETURN
END

each time control
The length of the dummy argument is determinedactual
argument can be
the
of
length
transfers to the function. The
substring, or
t,
elemen
array
le,
variab
ter
charac
a
of
the length
specifies a
ces
referen
n
functio
Each of the following
expression.
different length for the dummy argument.
CHARACTER VAR*10, CARRAY (3,5)*20

I1 =

ICMAX(VAR)

I2 = ICMAX(CARRAY (2,2))
I3 = ICMAX(VAR(3:8))

.
I4 = ICMAX(CARRAY(1,3) (5:15))
I5 = ICMAX (VAR(3:4)//CARRAY(3,5)

Arguments -Actual
6.1.1.3 Character and Hollerith Constants as Actual
agree in data
must
ts
argumen
dummy
g
pondin
corres
their
and
nts
argume
example,
(for
nt
consta
ith
Holler
a
is
type. If the actual argument
VAX-11
In
type.
data
numeric
of
be
must
t
argumen
dummy
the
,
4HABCD)
(for
nt
consta
er
charact
a
is
nt
argume
FORTRAN IV-PLUS, if an actual
either
have
can
t
argumen
dummy
g
pondin
corres
the
,
'ABCD')
example,

numeric or character data type. If the dummy argument has a numeric
converted to a
data type, the character constant '"ABCD' is, in effect,
linker.
Hollerith constant by the FORTRAN compiler and the

subroutine name 1is
An exception to this occurs when the functionle orto determ
ine at compile
possib
itself a dummy argument. It is not
t or Hollerith constant

time or link time whether a character constan
In this case, a character constant actual argument can
is required.
only correspond to a character dummy argument. For example:
SUBROUTINE S (CHARSUB,HOLLSUB,A,B)
EXTERNAL CHARSUB,HOLLSUB

B
'STRING')
RSU
CALL CHA(A,
G)
B
6HSTRIN
LSU
(B,
CALL HOL

SUBPROGRAMS

this

example,

themselves

dummy

the

subroutine

arguments

the

names

CHARSUB

subroutine

and

HOLLSUB

Therefore,

the

argument 'STRING'
in
the
«call
to
CHARSUB
must
correspond
character
dummy
argument,
while the actual argument 6HSTRING
call to HOLLSUB must correspond to a Hollerith dummy argument.

6.1.1.4

Alternate

argument

Return Arguments

dummy

argument

list,

specify

place

alternate

asterisks

the

are

actual
to
a
in the

return

list.

For

example:

SUBROUTINE MINN(A,B,*,*,C)
The

actual

return

the

argument

arguments

list
the

passed

the

corresponding

CALL

must

positions.

include
These

alternate

arguments

have

form:

*label
or

&label

You can use
either
an
asterisk
alternate
return
argument in an
specify

for

label

must

issued

the

CALL.

6.1.2

Built-In Functions

the

or
an
ampersand
to
actual argument list.

label

statement

Built-in functions perform
utility
operations
communicating
with
subprograms
written
in
FORTRAN.

There

two

kinds

built-in

Argument list built-in functions

23LOC

6.1.2.1

written
may

are

Argument

need

uses.

built-in

pass

You

can

the
use

-- in the argument
change the form of
You must not
procedure or

The three
argument
functions
function

use

program

that

the

that
are
languages

useful
in
other
than

functions:

Built-In

other
actual
three

Functions - To
call
subprograms
FORTRAN (such as system services), you
arguments in a different form than FORTRAN

than

built-in

list of a CALL
the argument.

these

built-in

user-supplied

functions

statement

functions

subprogram

$%VAL,

invoke

written

$REF,

function

You

cannot

use

them

any

and

$DESCR

reference

FORTRAN

library

FORTRAN.

argument
1list
built-in
functions
specify
should
be
passed
to
the
subprogram.
You
only in the actual argument list
of
a
CALL
reference.

an
you

function

List

languages

indicate
The value

other

the
way
the
can use these
statement
or

context.

SUBPROGRAMS

The argument list built-in functions are:
Effect

Function

$VAL (a)

Pass the argument as a 32-bit value.

$REF (a)

Pass the argument by reference.

$DESCR(a)

Pass the argument by descriptor.

In these functions,

a is an actual argument.

See the VAX-11 FORTRAN IV-PLUS User's Guide for

information

argument-passing mechanisms.

Table 6-1 lists the FORTRAN argument-passing defaults and the
uses of %VAL,

%REF,

and %DESCR.

Table

6-1

Argument List Built-In Functions and Defaults

Actual Argument

Data Type

Functions Allowed

Default
VAL

$REF

$DESCR

Expressions

LOGICAL

REF

Yes

INTEGER

REF

Yes

REAL*4

REF

Yes

REAL*8

REF

Yes

COMPLEX

REF

Yes

CHARACTER

DESCR

Yes

Hollerith

REF

Numeric

REF

Yes

CHARACTER

DESCR

Yes

Numeric

REF

Yes

CHARACTER

DESCR

Yes

Array

Name

Procedure Name

allowed

SUBPROGRAMS

6.1.2.2
$LOC Built-In Function - The 3%LOC built-in function
the internal address of a storage element.
It has the form:

computes

$LOC (v)

A variable
name,
substring name, or

array
element
name,
array
external procedure name.

name,

character

The %LOC built-in function produces an INTEGER*4 value that represents
the
1location
of
1its
argument.
It can be used as an element in an
arithmetic expression.

See the VAX-11l FORTRAN IV-PLUS User's Guide for
the %LOC built-in function.

6.2

information

USER-WRITTEN SUBPROGRAMS

A user-written subprogram is a FORTRAN statement or group
of
FORTRAN
statements
that perform a computing procedure.
A computing procedure
can be either a series of arithmetic operations or a series
of FORTRAN
statements.
You can use subprograms to perform a computing procedure
in several places in your program, and thus avoid having to
duplicate
the series of operations or statements in each place.
There are three types of subprograms.
Table 6-2 lists
each
type
of
subprogram,
the
statements
needed
to
define it, and the method of
transferring control to the subprogram.
.

Types

Subprogram

Arithmetic
function

Table 6-2
User-Written Subprograms

Defining

statement

Function

Control Transfer
Method

Statements

Arithmetic statement | Function
function definition

reference

FUNCTION

reference

Function

ENTRY

RETURN

Subroutine

SUBROUTINE

CALL

statement

ENTRY

RETURN

A function reference
function
name
and
returns a value that

is used in an
expression
and
consists
of
the
the
function
arguments.
A
function reference
is used in evaluating the expression in which
it

appears.

Function and subroutine subprograms can change
the
values
arguments;
the calling program can use the changed values.

A subprogram can refer to other subprograms;
directly or indirectly, refer to itself.

but

cannot,

their

either

SUBPROGRAMS

6.2.1

Arithmetic Statement Functions

An arithmetic statement function is a computing procedure defined by a
Its definition statement is similar in form to an
statement.
single
statement
arithmetic
the
to
refer
you
If
assignment statement.

The resulting value is made
the computation is performed.
function,
statement
arithmetic
the
available to the expression that contains
reference.

function

The arithmetic statement function definition statement has the form:
([pl,p)...]1)=e

The name of the arithmetic statement

dummy

function.

argument.

expression.

The expression

(e)

is an arithmetic or logical expression that defines

the computation to be performed.

An arithmetic statement function reference has the form:
£

(Ipl,pl...1)
function.

The name of the
p

actual

argument.

an
in
appears
reference
function
statement
an arithmetic
When
expression, the values of the actual arguments are associated with the
The
dummy arguments in the arithmetic statement function definition.
The resulting value
the definition is then evaluated.
in
expression
the
containing
expression
is used to complete the evaluation of the
function

reference.

The data type of an arithmetic statement function is determined either
implicitly
by
the initial letter of the function name, or explicitly
in a type declaration statement.
The data type
can
be
any
of
the
numeric data

types,

but cannot be

character data type.

Dummy arguments in a statement
function
indicate
only
the
number,
order,
and
data type of the actual arguments.
You can use the names
of the dummy arguments to represent other entities
elsewhere
in
the
program
unit.
Note
that,
except
for
data
type,
declarative
information associated with an entity is not associated with the dummy

That is, declaring an
arguments in the arithmetic statement function.
a
affect
not
does
entity to be an array or to be in a common block
dummy argument with

the

You cannot use the

name

same

name.

the

arithmetic

statement

represent any other entity within the same program unit.

function

can
definition
function
statement
arithmetic
The expression in an
a reference to another arithmetic
If
references.
function
contain
statement function appears in the expression, you must have previously
defined that function

in the same program unit.

SUBPROGRAMS

Any reference to an arithmetic statement function must appear
same program unit as the definition of that function.
An arithmetic statement function
of,
an
expression.
You cannot
an assignment statement.
Actual arguments must agree in
corresponding dummy arguments.
Examples

arithmetic

VOLUME (RADIUS)
SINH(X)

The
a

following

dummy

function definitions

type with

their

follow.

4.189*RADIUS**3

invalid definition.

You cannot

use

a constant

(A+B+C)/3.

arithmetic

statement

function

references

follow.

the definition:

AVG(A,B,C)
These

and data

argument.

Examples
is

order,

(EXP(X)-EXP(-X))*0.5

AVG(A,B,C,3.)

This

the

reference must appear as, or be
part
use the reference as the left side of

number,

statement

are

the

(A+B+C)/3.

references:

GRADE

= AVG(TEST1,TESTZ2,XLAB)
IF (AVG(P,D,Q).LT.AVG(X,Y,Z))
FINAL = AVG(TEST3,TEST4,LAB2)

300

The last of these three references is invalid, because the
data
of the third argument does not agree with the dummy argument.

6.2.2

type

Function Subprograms

function

subprogram

program

function

subprogram

returns

unit

consisting

FUNCTION

statement
followed
by a series of statements that define a computing
procedure.
You use a function reference
to
transfer
control
to
a
function
subprogram,
and a RETURN statement to return control to the
calling program unit.
a

single

value

the

unit
by
assigning that value to the function's name.
name determines the data type of the value returned.

<calling

The

program

function's

SUBPROGRAMS

6.2.2.1

Numeric Functions - The FUNCTION statement has the

[typ]
typ

form:

FUNCTION nam[*m] [([p[,p]...]1)]

One of the numeric data type specifiers.

nam

The name of
m

Unsigned,
data

function.

nonzero integer constant specifying the length

the

statement

has

type.

dummy

argument.

Character Functions - The CHARACTER

6.2.2.2
the

the

FUNCTION

form:

CHARACTER[*n]

FUNCTION nam

[([p[,p]...1)]

Unsigned, nonzero integer constant, or (*)
indicating
a
passed
If you specify CHARACTER*(*), the function
length function name.
assumes the length declared for
it
in
the
program
unit
that
invokes
it.
A
passed
length
character
function
can
have
different lengths when it is invoked by different program
units.
If
n
1is an integer constant, the value of n must agree with the
length of the function specified in the program unit that invokes
If you do not specify n, a length of 1 is assumed.
the function.
nam

The name of the

dummy

6.2.2.3
control

function.

argument.

Function
Reference - A
function
reference
to a function subprogram has the form:

nam

([pl[,pl...])

The

symbolic name of

that

transfers

nam

actual

the

function.

argument.

the
the values of
When control transfers to a function subprogram,
in the function reference are associated
(if any)
arguments
actual
The
statement.
FUNCTION
the
in
with the dummy arguments (if any)
A value must be
then executed.
subprogram are
in the
statements
is
Finally, a RETURN statement
assigned to the name of the function.

executed
in
that function and returns control to the calling program
to
available
The value assigned to the function's name is now
unit.
and is used to
reference,
function
the
expression containing
the
complete the evaluation of that expression.

SUBPROGRAMS

The data type of a function name can be specified
explicitly
in
the
FUNCTION
statement or in a type declaration statement, or implicitly.
The function name defined in the function
subprogram
must
have
the
same data type as the function name in the calling program unit.
The FUNCTION statement must be
the
first
statement
of
a
function
subprogram.
It
must
not
have
a
statement
1label.
A
function
subprogram
cannot
contain
a
SUBROUTINE
statement,
a
BLOCK
DATA
statement,
or
another
FUNCTION
statement.
ENTRY statements can be
included, to provide multiple entry points to
the
subprogram.
(See
Section 6.2.4.)
Examples

function

FUNCTION
2

subprograms

follow.

ROOT (A)

X =1.0
EX = EXP (X)

EMINX

ROOT =
IF

1./EX

((EX+EMINX)*.5+COS(X)-A)/((EX -

(ABS(X-ROOT) .LT.1E-6)

EMINX)*.5-SIN(X))

RETURN

X = ROOT
GO TO 2
END

The function in this example uses the
to obtain the root of the function:
F(X)

cosh(X)

The value of A
this root is:

cos(X)

is passed

Newton-Raphson

iteration

method

argument.

The

iteration

formula

for

cosh (Xi)+cos(Xi)-A

Xi+l

sinh (Xi)-sin(Xi)

This is calculated repeatedly until
is
1less than 1.0E-6.
The function
EXP, SIN, COS, and ABS (see Section

the difference between Xi and Xi+l
uses the FORTRAN library functions
6.3).

The following example
returns
the value of
function.

length
repeated

is
its

a
passed
argument,

CHARACTER*
(*) FUNCTION
CHARACTER*1 CARG
DO
10

I=1,

REPEAT

character
function.
to fill the length of

It
the

(CARG)

LEN(REPEAT)

(I:I)

CARG

RETURN
END

Within any
character
example,

given program
function
must

the

REPEAT

unit
have

function has

CHARACTER*1000 REPEAT,
MANYAS = REPEAT('A')
MANYZS = REPEAT('Z2')

all
the
a

MANYAS,

references
to
same
length.

length
MANYZS

1000.

a
In

passed
1length
the following

SUBPROGRAMS

However, another program unit within the executing program can specify
a different length.
1In the following example, the REPEAT function has
a length of

CHARACTER HOLD*6, REPEAT*2
HOLD = REPEAT('A') // REPEAT('B')

6.2.3

// REPEAT('C')

Subroutine Subprograms

A subroutine subprogram is a program unit consisting of
a
SUBROUTINE
statement
followed
by a series of statements that define a computing
procedure.
You
use
a
CALL
statement
to
transfer
control
to
a
subroutine subprogram, and a RETURN statement to return control to the
calling program unit.
The

SUBROUTINE

statement has

SUBROUTINE

nam

the

form:

[([pl[,p]l..-]1)]

nam

The

name

A dummy

the

subroutine.

argument.

alternate

You

can

specify

dummy

return argument by placing an asterisk

argument

in the argument

list.

Section

4.5 describes

the CALL

statement.

When control transfers to
the
subroutine,
the
values
of
the
actual
arguments
(if
any) in the CALL statement are associated
with the corresponding dummy arguments (if any) in the SUBROUTINE
statement.
The statements in the subprogram are then executed.
The SUBROUTINE
statement
must
be
the
first
subroutine.
It must not have a statement label.

statement

A subroutine
BLOCK
DATA

subprogram cannot contain a
FUNCTION
statement,
a
statement,
or
another SUBROUTINE statement.
ENTRY

subroutine.

(See

statements are allowed,

to specify multiple entry points

Section

the

6.2.4.)

Examples:

The subroutine in the following example computes the volume of
a
regular
polyhedron,
given the number of faces and the length of
one edge.
It uses the computed
GO
TO
statement
to
determine
whether
the
polyhedron
1is
a
tetrahedron,
cube,
octahedron,
dodecahedron, or icosahedron.
The GO TO statement also transfers
control
to
the proper procedure for calculating the volume.
If
the number of faces is not 4, 6, 8, 12,
or
20,
the
subroutine
displays an error message on the user's terminal.

SUBPROGRAMS

Main Program

COMMON NFACES,EDGE,VOLUME
ACCEPT *, NFACES,EDGE
CALL

TYPE

PLYVOL

'VOLUME=',6 VOLUME

STOP
END

Subroutine
SUBROUTINE PLYVOL
COMMON NFACES ,EDGE,VOLUME
CUBED = EDGE**3

corTo

GOTO

(6,6,6,1,6,2,6,3,6,6,6,4,6,6,6,6,6,6,6,5),NFACES

0.11785

VOLUME = CUBED *

0.47140

VOLUME = CUBED
RETURN

VOLUME

CUBED

RETURN

VOLUME = CUBED *
RETURN

VOLUME = CUBED *

7.66312

2.18170

RETURN

TYPE

100,

NFACES

(' NO REGULAR POLYHEDRON HAS
100 FORMAT

',I3,'FACES.'/)

VOLUME=0.0
RETURN
END

example

following

The

illustrates

the

wuse

alternate

return

specifiers to determine where control is transferred on completion of
The SUBROUTINE statement argument list contains two
the subroutine.
actual
the
to
corresponding
return arguments,
alternate
dummy
The RETURN
arguments *10 and *20 in the CALL statement argument list.
Thus,
taken depends on the value of Z, as computed in the subroutine.
O,
to
equal
if
taken;
1is
return
normal
the
0,
than
less
is
2
if
if more than 0,
1is to statement label 10 in the main program;
return
return is to statement label 20

in the main program.

Main Program

CALL CHECK
TYPE

GO TO

20
30

Subroutine

SUBROUTINE CHECK

(A,B,*10,%*20,C)

'VALUE LESS THAN

ZERO'

TYPE *, 'VALUE EQUALS ZERO'
GO TO 30
TYPE *, 'VALUE MORE THAN ZERO'
CONTINUE

50
60
70
80

IF(z) 60,70,80
RETURN
RETURN 1
RETURN 2
END

(X,Y,*,*,Q)

SUBPROGRAMS

ENTRY
6.2.4

ENTRY

The

ENTRY
subprogram.

Statement

statement
It

provides

multiple
executable and can
after
the
FUNCTION

entry

not

points

appear

within

function

subroutine
program
or
SUBROUTINE
statement.
Execution of a subprogram referred to by an entry name begins with the
first executable statement after the ENTRY statement.
The

ENTRY

statement

ENTRY

nam

has

the

form:

[([p[,p]...]1)]

nam

The

Use

entry

name.

dummy

argument.

the

CALL

statement

refer

subprograms.

function

Use
function
subprograms.

name

entry

declaration
You

can

actual

You

function

entry

argument;
use

the

name

you

entry

cannot

names

appearance

entry

in
the

names

refer

subprogram

statement.

specify

cannot

precede

within

references

EXTERNAL

use

executable
entry

name

can

use

number,

dummy

type,

arguments

ENTRY

names

appear

use

that

ENTRY

statements

and

within

type

argument.

statements
in

subroutine

entry

statement
dummy

You can include alternate return
arguments
in
placing
asterisks
in the dummy argument list.
specify alternate return arguments can
be
used
subprograms.
You

within

physically

statement.

ENTRY
ENTRY
only

that

statements
by
statements that
in
subroutine

differ

and

order,

name
from
the
dummy
arguments
you use in the
FUNCTION,
SUBROUTINE,
and
other
ENTRY
statements
in
the
same
subprogram.
However,
each
reference
to a function, subroutine, or
entry

and

must

type

use

dummy

ENTRY

dummy
not

argument

the

first

argument
currently

retention of
to the next.
You

cannot

loop.

actual

the

SUBROUTINE,

with

list

that

list

agrees

the

order,

corresponding

number,

FUNCTION,

statement.

specified.

associated

argument

can be referred to only in executable statements that
SUBROUTINE, FUNCTION, or ENTRY statement im which the

argument

use

argument

dummy

A dummy
with

association

ENTRY

statement

argument
actual

undefined

argument.

from one

reference

within

block

There

1is

is
no

subprogram

construct

SUBPROGRAMS

entry names within a
6.2.4.1 ENTRY in Function Subprograms - All
function subprogram are associated with the name of the function
subprogram. Therefore, defining any entry name or the name of the
function subprogram defines all the associated names of the same data
all associated names that are of different data types become
type;
The function and entry names need not be of the same data
undefiped.
data
type, but they all must be either numeric data type or character
type. At the execution of a RETURN statement or the implied return at
subprogram must be defined.

refer

the end of the subprogram, the name used

the

function

If the function is of character data type, the entry name must also be
of character data type and must have the same length specification as

Note that the length specified must also agree with the
the function.
If
length specified in the program unit referring to the entry name.
an asterisk enclosed in parentheses is used to specify the length of
the entry name has a passed length (see Section
the entry name,
6.1.1.2).

Figure

6-1

illustrates

hyperbolic functions sinh,

subprogram

function

cosh,

and tanh.

that

computes

REAL FUNCTION TANH (X)
C

STATEMENT FUNCTION TO COMPUTE TWICE SINH
TSINH(Y)

= EXP(Y)

EXP

(-Y)

C
C

STATEMENT FUNCTION TO COMPUTE TWICE

TCOSH(Y)

= EXP(Y)

COSH

+ EXP(-Y)

C
C

COMPUTE

TANH

TANH = TSINH(X)

/ TCOSH (X)

RETURN
C

COMPUTE

SINH

ENTRY

SINH (X)

SINH = TSINH(X)

/ 2.0

RETURN

COMPUTE

COSH

ENTRY

COSH =

COSH (X)

TCOSH(X)

2.0

RETURN
END

Figure 6-1

Multiple Functions in a Function Subprogram

the

SUBPROGRAMS

6.2.4.2

ENTRY

name

point

name

in Subroutine Subprograms - To refer to an entry
subroutine, issue a CALL statement that includes the
defined in the ENTRY statement.
For example:

point

entry

Main Program
CALL

SUBA(A,B,C)

Subroutine
SUBROUTINE

ENTRY

SUB

(X,Y,Z)

SUBA(Q,R,S)

In this example, the CALL is
subroutine (SUB).
Execution
ENTRY

CALL

SUBA

(Q,R,S),

statement.

statements.

using

Note

For

that

to
an
begins

the

SUB

ENTRY

(J,K,*,*,X)

SUBC

RETURN

1
2

arguments

returns

(SUBA)
within
the
statement following

(A,B,C)

can

example:

SUBROUTINE

RETURN

actual

alternate

entry
point
with the first
be

passed

specified

the

ENTRY

(K,*,*)

END

If you issue a CALL to entry
alternate return arguments.
CALL

this

RETURN
program.

SUBC,
example:

you

control

statement

must

include

actual

SUBC(M,N,*100,*200,P)

case,
2

point
For

RETURN

transfers

transfers
control

statement

label

200,

100,
the

and

calling

SUBPROGRAMS

6.3

FORTRAN LIBRARY FUNCTIONS

tions are provided to perform commonly used
FORTRAN library funcions
.
mathematical computat
Function
The FORTRAN library functions are listed in Appendix B.

the same way
ary functions are written inwritt
references to FORTRAN libr
en. For
are
function references to user-defined functions

example:

R = 3.14159 * ABS(X-1)

of X-1 is calculated
ence, the absolute value
As a result of this refer
is assigned to the
t
resul
the
59;
3.141
and multiplied by the constant
variable R.
ion and of ther
type of each librarydescfunct
Appendix B gives the data
the characte
s
ribe
also
4
6.3.
Section
actual arguments.
functions.

Processor-Defined Function References

6.3.1

called processor—-defined
function names are defin
The FORTRAN library
ed functions include
ssorthat the proce
function , names. Note ions
Functions defined
nal
Exter
Basic
the
and
both the Intrinsic Funct
in ANS FORTRAN.
d function names
table of processor-definename.
Normally, a name in the libr
However, the
that
with
ion
funct
ary
AN
refers to the FORTRuser
following
the
of
any
r
-defined function unde
name can refer to a
conditions:
declaration statement specifying a
e The name appears in a typeshown
in the table.
different data type than

in an
e The name is prefixed with an asterisk and appears
EXTERNAL statement.

ion reference with arguments of a
e The name is used in a funct
shown in the table._
different data type than

and are prefixed by
in an EXTERNAL statement local
Except when they are used
to the program
ed function names are for other
an asterisk, processor-definThus,
purposes
used
be
can
they
unit that refers to them. .
of a
type
data
the
In addition,
units
in other program tion
ICIT
IMPL
an
use
you
if
e
chang
not
does
processor-defined func
statement to change the implied data type rules.

t with
an EXTERNAL statemention
fined function name inname
Use of a processor-de
or
func
a
to
s
refer
specifies that the nal subprogram.
an asterisk prefix will
provide as an exter
subroutine that you
6.3.2

Generic Function References

ng some of the
provide a way of calli
Generic function references
n of the actual
ctio
sele
the
that
such
s
tion
FORTRAN mathematical is func
ent in the
argum
the
of
type
based on the data
used
library routinerenc
able, the
vari
real
a
is
X
1if
ple,
exam
For
e.
function refe
tion.
sine
ed
valu
real
) refers to the the function referencfunc
function reference leSIN(X
(D)
SIN
e
able,
But if D is a doub precisionn vari
e
writ
not
need
You
.
tion
sine func

refers to the double precisio

.
DSIN(D)

6-17

SUBPROGRAMS

Generic

function selection
occurs
independeritly
for
each
function
reference.
Thus,
you
could use both the function references
in the
example above, SIN(X) and SIN(D), in
the same program unit.

Table

6-3
only with

You

cannot

lists

the generic
argument data

the
use

the

them

you

use

the

type

names

function

types

Table

6-3

unit

a program
declaration

name

dummy

names.
These names
shown in the table.
generic

the

function

following

used

selection

ways:

statement

arithmetic

argument

for

any

can

name,

statement

common

block

name

function

name,

variable

array

Using a generic name in an EXTERNAL stateme
nt does not affect
generic
function
selection
for
function references.
When you use a generic
function name in an actual argument list
as the name of a function
to
be
passed,
no
generic
function
selection
occurs,
as there is no
argument

list

according
above

Generic

them.

which

base

the
rules
Section 6.3.1.

function

Thus,

names

they

can

are

nongeneric

local

used

selection.

for

Table
Generic

Generic

the

other

Function

program

unit

Name

Type

Integer

Double
Complex

INT,

ANINT

NINT

DBLE

EXP,

LOG,

MIN,

SIN,

SIGN,

COS,

DIM

LOG10,

TAN,

ATAN,

ATAN2,

ACOS,

SINH,

COSH,

TANH

ASIN,

Type

Result

Integer

Real
Double
Real
Real

Integer

Double

Integer

Integer
Double

Real

Integer

SQRT

Data

Double

Integer

MAX,

Real

Real
MOD,

refers

programs.

Double
Real

SNGL

that

other

treated
described

Summary

Real

AINT,

1is

6-3

Argument

ABS

name

functions

purposes

Data

Name

The

FORTRAN

Real

Double

Double
Integer

Real

Double

Real

Double

Complex

Real
Double

SUBPROGRAMS

6.3.3

Processor-Defined

and

Figure

6-2 shows the use
In this figure,
four distinct ways:
name

the

generic

processor-defined

user-defined

Using the name in
properties of the
In Figure 6-2, the
follow the figure.

arithmetic

function

these
name.

Function

Usage

of
processor-defined
and
generic
a single executable program uses the

names.

Generic

statement

function
name SIN

function

name
function

name

function

four

ways

parenthetical

emphasizes

notes

are

the

keyed

local

the

and

notes

global

that

QOO0

SUBPROGRAMS

COMPARE

WAYS

PROGRAM

SINES

PARAMETER PI

COMPUTING

SINE.

3.141592653589793238D0

REAL*8

COMMON

V (3)

DEFINE
SIN(X)

SIN AS A STATEMENT FUNCTION
= COS(PI/2-X)

X =

REFERENCE
10
100

-PI,

PI,

CALL

COMPUT (X)

THE

2*PI/100

STATEMENT

WRITE (6,100) X,V,
FORMAT (5F10.7)

(Note

FUNCTION

SIN

(Note

SIN(X)

END

SUBROUTINE COMPUT (Y)
REAL*8 Y
MAKE PROCESSOR-DEFINED
EXTERNAL
COMMON

SIN

EXTERNAL

(Note

REFERENCE

V(l) = SIN(Y)
PROCESSOR-DEFINED

DOUBLE

FUNCTION

PRECISION
SINE AS

SINE

(Note

END

SUBROUTINE

SUB(A,S)

DECLARE SIN AS NAME OF USER FUNCTION.
EXTERNAL *SIN
DECLARE SIN AS
REAL*8 A, SIN

TYPE

REAL*8

(Note

(Note 6)

COMMON V (3)
EVALUATE PROCESSOR-DEFINED FUNCTION
V{(2) = S(A)
EVALUATE USER DEFINED SIN FUNCTION.

SIN

(Note

SIN(A)

OO0

END

DEFINE

THE

REAL*8

FUNCTION

INTEGER

SIN
1

USER

SIN

FUNCTION.

(Note

10)

SIN(X)

FACTOR

X -

X**3/FACTOR(3)
X**7/FACTOR(7)

X**5/FACTOR(5)

END

INTEGER FUNCTION
FACTOR =1

DO
10

FACTOR

I=N,
=

FACTOR(N)

-1

FACTOR

END

Figure

6-2

Multiple

ACTUAL ARGUMENT.

SUB(Y,SIN)

V(3)

V (3)

GENERIC

CALL

FUNCTION

SIN

Function Name

Usage

(Note

SUBPROGRAMS

NOTES

An arithmetic statement function
named
SIN
is
defined
in
terms
of
the generic function name COS.
Since the argument
of COS is
double
precision,
the
double
precision
cosine
function will be evaluated.

The

The
name
precision,
an

arithmetic

actual

statement

function

SIN

called.

SIN
is
declared
external
so
that
the
single
processor—-defined
sine function can be passed as
argument

The generic function name
precision sine function.

The single precision, processor—-defined
as an actual argument.

sine

)

The

name

SIN

function

The

type

SIN

The

single precision

The user-defined SIN

6.3.4

declared
is

SIN

used

user-defined

declared

sine

refer

the

function

double

used

name.

double precision.

function passed

function

evaluated.

The user—-defined SIN function is defined as a
simple
Taylor
series
using
a
user-defined function FACTOR to compute the
factorial function.

Character Library Functions

Character
library
functions
are
functions
that
take
arguments
or
return
character
values.
There
are
four
functions provided with FORTRAN:
e

character
character

LEN

The LEN
The LEN

function
function

returns
has the

the length of
form:

a character

expression.

LEN (c)
o]

A character expression.
many bytes there are in
®

The value returned
the expression.

indicates

how

INDEX
-

The
INDEX
function
searches
for
a
substring
(c2). in
a
specified
character
string
(cl),
and,
if
it
finds
the
substring, it returns the substring's starting
position.
If
c2
occurs
more than once in cl, the starting position of the
first occurrence is returned.
If c2 does not occur in cl, the
value zero is returned.
The INDEX function has the form:
INDEX

(cl,

c2)

SUBPROGRAMS

cl
A

character

searched

expression

for

the

specifying

substring

the

specified

string

c2.

c2
A

character

the
¢

location

specifying

the

substring

for

which

determined.

ICHAR

The

ICHAR

function

equivalent
the form:

ASCII

ICHAR

The

character
than

returned.
e

converts

code

and

character

returns

expression

the

ASCII

value.

an ASCII

ICHAR

has

(c)

longer

byte,

The

converted
only

the

value

remainder

are

ignored.

converts

ASCII

code.

the

first

byte

CHAR

The

CHAR

function

character

the

Examples

value,

and

returns

the

integer

character

(1)

integer

illustrating

Function

expression.

the

LEN

and

Example:

SUBROUTINE REVERSE
(S)
CHARACTER T, S*(¥*)
J = LEN(S)
IF (J .GT.

THEN

I=1,

J/2

10
T

S(I:1I)

S(I:I)

S(Jd:J)

J
10

J-1

CONTINUE
ENDIF
RETURN

END

S(J:J)

INDEX

functions

value

value.

form:
CHAR

LEN

expression

starting

follow.

CHAR

has

SUBPROGRAMS

INDEX Function Example:

GS S)
(SUB,
SUBROUTINE FINDSUBSTRIN

(*)
CHARACTER*
CHARACTER*132

SUB,S
MARKS

I =1
K=1

MARKS
10

J =

INDEX(S(I:),

.NE.

SUB)

THEN

I =1+ (J-1)
MARKS (I:I) = '#'
K=1

I+l

.LE.

LEN(S))

GO TO 10

ENDIF

100

WRITE(7,100) S, MARKS(:K)
FORMAT( 2(/1X, A) )
END

CHAPTER

INPUT/OUTPUT STATEMENTS

FORTRAN programs

WRITE, TYPE, and
statements
are

use

READ

and

PRINT
statements
used
with
format

statements

for
output.
specifiers

translation and editing of the data to and

and

external

character

(readable)

form.

for

Some
that

input;

these
the

(binary)

form

from internal

Each READ or WRITE statement refers to the logflcal

and

forms of
control

unit

which data is to be transferred.
A logical unit can be connected
device or file by the OPEN statement (see Section 9.1).

from

The ACCEPT, TYPE, and PRINT statements do not refer to logical
units;
rather, they transfer data between the program and an implicit logical
unit.
The ACCEPT and TYPE statements are normally
connected
to
the
default
input
or
output
device and the PRINT statement is normally
connected

Input/ouytput
categoriies:
@

the default output device.

(I/0)

statements

are

grouped

into

the

following

Formatted Sequential
I/0O -- transmits
character
data
using
format specifiers to control translation of data from internal
(binary) form to character (readable) form on output, and vice

versa oninput.
@

List-Directed Sequential I/0 -- transmits character data.
The
data type of the I/O list elements controls the translation of
data

internal

form

input

and

character

form

output.

Unformatted Sequential

Formatted Direct Access I/0 -- transmits character data to and
from
direct
access
files using format specifiers to control
translation of data from internal form to
character
form
on
output, and vice versa on input.

Unformatted Direct Access I/0 -- transmits binary data without
translation to and from direct access files.

Internal
transfer

without

translation.

I/0 --

transmits data

I/0 -- ENCODE and
DECODE
data
between
variables

statements
and
arrays

internal

form

translate
and
in the FORTRAN

program.

These

categories of

through

7.7.

Section

7.1

I/0

statements

below describes

are

the general

described

components of

1in

I/0

Sections

7.2

statements.

INPUT/OUTPUT STATEMENTS

7.1

I/0 STATEMENT COMPONENTS

The

following

record

sections

numbers,

error

7.1.1

describe

format

end-of-file

Logical

logical

specifiers,

condition

unit

I/O

numbers,

lists,

and

occurs.

direct

transfer

access

control

Unit Numbers

I/0 statements use logical unit numbers to
refer
to
files
and
I/0
devices.
A logical unit number is an integer expression with a value
in the range 0 through 99.
It is converted, if necessary, to
integer
data type before use.
You

can

use

OPEN

device

1logical

User's

Guide

logical

Some

(see

wunit.

describes

other

Chapter

ways

the

connect

I/O

connect

VAX-11

file

unit.

file

FORTRAN
I/O

I/O

IV-PLUS

device

statements do

Instead,

these

number.

‘The

implicit

logical

7.1.2

statement

not

statements

VAX-11

contain
use

FORTRAN

unit

Direct Access

IV-PLUS

numbers

Record

explicit

1logical

system-specific,

User's

in greater

Guide

unit

implicit

number.

logical

describes

the

unit

use

detail.

Numbers

When using direct access I/0, you must indicate which record is to
be
read or written.
To do so, you specify a direct access record number,
which is a numeric expression.
If necessary,
the
record
number
is
converted

direct

access

less

than

For
the

a more
VAX-11

integer

record
equal

general
FORTRAN

7.1.3

Format

Format
of the

specifiers
following:

data

number
the

type

used.

than

number

records

defined

discussion of record numbers,
IV-~PLUS User's Guide.

The

equal
for

refer

value

the
to

and

file.

Chapter

Specifiers

are

used

FORMAT

formatted

label

character

variable

character

array

name

character

array

element

character

substring

numeric

array

FORMAT

describes

In sequential I/O
format
specifier

statements

and

can

any

can

name

reference
which

format

contains

value

that

statements.

statements,
to
denote

list-directed

I/O

statement

name

interpreted

describes

greater

Chapter

before

must

1/0.

you can use an
1list-directed

asterisk
instead
of
a
formatting.
Section 7.3

INPUT/OUTPUT STATEMENTS

Input/Output Records

7.1.4

The amount of
I/0 statements transfer all data in terms of records.
data that one record can contain, and the way records are separated,

depend on how the data is transferred.

In unformatted I/0, the I/0 statement specifies the amount of data to
the 1I/0 statement and 1its
In formatted 1I/0,
transferred.
be
associated format specifier jointly determine the amount of data to be
transferred.

of a new
Executing an input or output statement initiates transfer
the
record,
a
of
part
only
requires
statement
input
an
If
record.
©Normally, the data transferred by
remainder of the record is ignored.
formatted
However,
record.
constitutes one
statement
I/0
an
sequential I/0 statements can transfer more than one record.
|

Input/Output Lists

7'1F5
!

The I/0 list in an input or output statement contains the names of
and character substrings whose
array elements,
arrays,
variables,
The I/0 list in an output statement can
values are to be transferred.
also contain constants and expressions.

An I/0 list has the

form:

s[,s]...

A simple list or an implied DO list.

from,
The I/0 statement assigns input values to, or transfers values
in the order in which they appear, from left to
the 1list elements
right.

7.1.5.1 Simple Lists - A simple I/O list element can be a single
substring reference,
character
array element,
array,
variable,
constant, or expression. A simple I/O 1list consists of either a
simple I/O 1list element or a group of two or more simple I/0 list
elements separated by commas and enclosed in parentheses.

a READ or
When you use an unsubscripted array name in an I/O 1list,
ACCEPT statement reads enough data to fill every element of the array;
array.
a WRITE, TYPE, or PRINT statement writes all the values in the

Data

transfer

begins

proceeds in the order of

subscript

with

the

subscript

varying most rapidly.

two-dimensional

initial

element

progression,

with

the array and
the

leftmost

For example, the following defines a

array:

DIMENSION ARRAY (3,3)

statement,
If the name ARRAY, with no subscripts, appears in a READ
to
input record(s)
the
from
values
assigns
statement
that
through
on,
so
and
ARRAY(l,2),
ARRAY(3,1),
ARRAY (1,1), ARRAY(2,1),
ARRAY (3,3).

INPUT/OUTPUT STATEMENTS

In
in

READ
array

or ACCEPT statement, variables in the I/0 list can
be
subscripts
or substring references later in the list.

used
For

example:

1250

The

input

READ (1,1250) J,K,ARRAY (J,K)
FORMAT (Il,X,Il,X,F6.2)

record

contains

the

following

values:

1,3,721.73

When the READ statement is executed,
to
J
and
the second to K, thereby

values

for

ARRAY (1,3).
must appear

An output
However,
on

the

ARRAY (J,K).
Then
the
value
721.73
1is
assigned
to
Variables
that are to be used as subscripts in this way
before (to the left of) their use in the array subscript.

statement
I/O0O
1list
this
expression must

same

the first input value is assigned
establishing the actual subscript

logical

unit.

expression
cannot refer
the same logical unit.

An input statement I/O
subscript
expression
substring reference.

For
to

<can
<contain
any
valid
expression.
not attempt any further I/0 operations
example,

function

output

subprogram

list cannot contain an
1in
an array reference

statement
that

I/0

performs

1list
I/0

expression, except as
or as an expression in

a
a

7.1.5.2
1Implied DO Lists - Implied DO lists
are
used
to:
specify
iteration within an I/0 list;
transfer part of an array;
or transfer
array elements in a sequence different than
the
order
of
subscript
progression.
This type of list element functions as though it were a
part of an I/0 statement in a DO loop;
it uses the
control
variable
of
the implied DO statement to specify the value(s) to be transferred
during each iteration of the loop.
An

implied

list

has

the

form:

(list,i=el,e2[,e3])

list
An

I/0

integer,

list.

el,e2,e3
Arithmetic

real,

double

precision

variable.

expressions.

The variable i, and the parameters el, e2, and e3 have the same
forms
and the same functions that they have in the DO statement (see Section
4.3).
The list can refer to the variable
i
(which
1is
the
control
variable)
as
1long
as
it
does not change the value of i.
The list
specifies the range of the implied DO list.
For example:
WRITE

(3,200)

WRITE

(6,15)

READ

(1,75)

(A,B,C,

I=1,3)

L,M,(I,(J,P(I),0Q(I,d),d=1,L),I=1,M)
(((ARRAY(M,N,I),

I=2,8),

N=2,8),

M=2,8)

INPUT/OUTPUT STATEMENTS

innermost
the
The first control variable definition is equivalent to
DO loop of a set of nested loops, and therefore varies most rapidly.
example:

For

150

This

WRITE

FORMAT

statement

(6,150)

(F10.2)

L=1,10),

((FORM(K,L),

K=1,10,2)

to:

1is similar

150

50 K=1,10,2
50 L=1,10
WRITE (6,150) FORM(K,L)
FORMAT (Fl10.2)

CONTINUE

Because the inner DO loop is executed ten times for each iteration
of
the second subscript, L, advances from 1 through 10
1loop,
outer
the
This is the reverse of the
for each increment of the first subscript.
In addition, K is incremented by 2,
subscript progression.
of
order
so only the odd-numbered rows of the array are output.

The entire list of the implied

control variable is incremented.
READ

(5,999)

(p(I),

(Q(I1,J),

1list

For example:
J=1,10),

transmitted

before

I=1,5)

This statement assigns input values to the elements of arrays P and
in

the

order:

P(l), Q(1,1), Qo(1,2),
P(2), Q(2,1), Q(2,2),

...
...

P(5), Q(5,1), Q(5,2),

...

, Q(5,10)

When processing multidimensional arrays, you can use a combination of
fixed subscripts and subscripts that vary according to an implied DO
list.

For

READ

example:

(3,5555)

(BOX(1,J),

J=1,10)

through BOX(1l,10),
to BOX(1l,1)
This statement assigns input values
then terminates without affecting any other element of the array.
The value of the control variable can also be

output

example:

WRITE

(6,1111)

(I,

I=1,20)

This statement simply prints the integers 1 through 20.

directly.

For

INPUT/OUTPUT STATEMENTS

7.1.6
Any

Transferring

READ

another
error

WRITE

Control

statement

condition.

respectively,

the

The

for

End-of-File

can

specify

I/0

statement

that

specifications

end-of-file

and

Error
control

transfer

encounters

end-of-file

have

error

Conditions

the

following

forms,

conditions:

END=s

ERR=s
—

The label
transfer.

executable

statement

which

control

A sequential READ or WRITE statement can contain either
above

the

specifications,

unit

number,

record

any

order.

number,

The

and/or

both

specification(s)

format

must
specification.

the

An end-of-file condition occurs
when
no
more
records
exist
in
a
sequential
file,
or
when an end-file record produced by the ENDFILE

statement
is
encountered.
If
a
READ
statement
encounters
an
end-of-file condition during
an I/0 operation, it transfers control to
the
statement
named
in
the
END=s
specification.
If
no
END=s
specification is present, an error condition occurs.
If

READ

I/0

WRITE

operation,

appears

present,

the

statement

executable

encounters

transfers

control

specification.
If
error terminates program

label

the

statement

END=s

within

error

ERR=s

I/O

The

statement

ERR=s
the

the

condition

during

statement whose

ERR=s
execution.

label

specification

specification must refer
program
unit as the

I/0

same

statement.

An END= specification in a WRITE statement or
a
direct
statement
is ignored.
If you attempt to read or write a
a record number dgreater than the maximum
specified
for

unit,

The
that
also

error

condition occurs.

access
record
the

READ
using

1logical

VAX-11

FORTRAN IV-PLUS User's Guide describes
system
subroutines
you
can use to control error processing.
These subroutines can
obtain information from the I/0O system on the type of error
that

occurred.

Examples

READ

I/0

statements

(8,END=550)

follow.

(MATRIX(K),K=1,100)

This

statement transfers control
condition occurs on logical unit
WRITE

This

to
8.

statement

550

when

390

when

end-of-file

(6,50,ERR=390)

statement

transfers

control

statement

occurs.,

and

(1,FORM,ERR=150,END=200)

statement
to

error

READ

This

transfers

statement

200

when

control
the

ARRAY

statement

end-of-file

7-6

150

when

condition

error

occurs.

occurs

INPUT/OUTPUT STATEMENTS

FORMATTED SEQUENTIAL INPUT/OUTPUT

7.2

Formatted sequential I/O statements are used with format specifiers to
output data from internal (binary) form to external
translate
character (readable) form; and to translate input data from external
to

form.

internal

READ

ACCEPT
Formatted Sequential Input Statements

7.2.1

the
A formatted sequential READ statement transfers data from
specified logical unit. If a formatted sequential READ statement does
not have a logical unit number, it uses an implicit logical unit.
The formatted sequential ACCEPT statement is similar to a formatted
sequential READ statement except that it always uses an implicit
logical

unit number.

Formatted sequential input statements have the forms:
READ

(u,f[,END=S][,ERR=s])[list]

READ

£f£[,list]

ACCEPT
u

list

f[,list]

A logical unit number.

A format specifier.

The label of an executable statement.
An I/0 list.

A statement of the following form
system-defined logical unit:
READ 200,

data

read

from

ALPHA,BETA,GAMMA

transferred

Characters

causes

formatted

sequential

statements

are

translated to the internal form specified by the format specifier.
The resulting values are assigned to the elements of the I/0 list,

If the number of list elements is less than the number of input record
fields, the excess portion of the record is ignored.

Usually a single formatted record is transferred by the execution of a
formatted sequential input statement. However, the format specifier
can specify that more than one record is to be read during execution
of a single

input statement.

formatted input statement
character constant, input data is read and

If the FORMAT statement associated with a

contains

Hollerith

stored directly into the format specification.

INPUT/OUTPUT STATEMENTS

I/0

record

list

and

100

format

READ

(5,100)

FORMAT

REVIEW

The

FORMAT

100

the

occurs
example:

only

betveen

the

characters

from

the

record

statements

follow.

logical

SECTIONS

becomes:

(lS5HREVIEWASECTIONS)

formatted

READ

300

These

read

statement

transfer

For

15 characters are:

FORMAT

Examples

data

specifier.

(l5HADATAAGOESAHERE)

These statements

unit

present,

the

sequential

(1,300)

FORMAT

input

ARRAY

(20F8.2)

statements

read

ARRAY.

a record

from logical

unit

and assign

fields

READ

(5,50) CHARV
FORMAT (AZ25)

These

statements read a record from
logical
character field to character variable CHARV.
READ

100

100,

(I5,

These

statements

read

assign

fields

BETA.

F8.2,

assign

logical

wunit,

variables

ALPHA

and

CHARAR(5)

200,

FORMAT

(5A10)

These statements
assign fields to

and

F5.1)

a record from
an
implicit
integer variable ICOUNT and real

CHARACTER*10
200

ICOUNT,ALPHA,BETA

FORMAT

wunit

CHARAR

read a record
the character

from
an
implicit
array CHARAR.

logical

wunit,

and

WRITE

TYPE
PRINT
7.2.2
The

Formatted
formatted

specified

Sequential
sequential

logical

unit.

Output

Statements

WRITE

statement

The formatted sequential TYPE and PRINT
formatted
sequential
WRITE statement,

implicit

logical

unit.

transfers

data

the

statements are similar to
the
except that output is directed

INPUT/OUTPUT STATEMENTS

The formatted sequential output statements have the form:
WRITE
TYPE

PRINT
u

(u,f[,ERR=s]) [list]
f[,list]

f[,list]

A logical unit number.

A format specifier.

The label of an executable statement.

list

An I/O0 list.

d to
The I/0 list specifies a sequence of values that are converte
r. If no
characters and positioned as specified by the format specifie
I/0 list is present, data transfer occurs only between the format
specifier and the record.

nt
The data transferred by a formatted sequential output thestateme
format
However,
record.
d
formatte
one
normally constitutes
specifier can specify that additional records are to be written during
execution of a single output statement.

during the
Numeric data output under format control is isrounded
ently input
subsequ
data
such
(If
format.
l
externa
to
ion
convers
To avoid
result.
may
on
precisi
of
loss
for additional calculations,
loss of precision, use unformatted output.)

nt must not exceed
The records transmitted by a formatted WRITE statemeFor
example, a line
accept.
can
the 1length that the specified device
characters.
printer typically cannot print a record longer than 132

Examples of formatted sequential output statements follow.
650

WRITE

FORMAT

(6,

650)

('AHELLOATHERE')

These statements write the contents of the FORMAT statement to logical
unit

CHARACTER*16 NAME,JOB
PRINT 400, NAME, JOB

400

FORMAT

('NAME=',A,'JOB=',A)

These statements write one record consisting of four fields.
95

WRITE (1,95) AYE,BEE,CEE
FORMAT (3F8.5)

These statements write one record,
logical unit 1.

consisting

three

fields,

INPUT/OUTPUT STATEMENTS

950

WRITE

(1,950)

FORMAT

(F8.5)

These

statements

each,

In
of

write three
unit 1.

the last example, format
the
FORMAT
statement

output.

new

logical

AYE,BEE,CEE

Each

record

time

this

initiated.

separate

control
before

occurs,

Thus,

the

records,

consisting

one

field

reaches the rightmost
all
elements
of the
current

three

record is
separate records

parenthesis
I/0 1list are
terminated and a

are

written.

INPUT/OUTPUT STATEMENTS

7.3

LIST-DIRECTED SEQUENTIAL

INPUT/OUTPUT

List-directed sequential I/O statements provide a way to obtain simple
formatted
sequential input or output without using FORMAT statements.
On input, values are read from the logical unit, converted to internal
form, and assigned to the elements of the.I/0O list.
On output, values
in the I/0 list are converted to character form and written in a fixed
format
according
to
the
data
type
of the value.
The I/0 list is
required.
Records written by list-directed
list-directed
input
statements
logical values.
Both formatted
to
the
same

sequential

output statements
if
they
contain

can
be
input
only
numeric

by
and

and list-directed sequential I/O statements
can
refer
1logical
unit.
All operations permitted for formatted

I/0

are

permitted

for

1list-directed

I1/0.

However,

backspacing
over
list-directed
records
1leaves
the
file
position
undefined.
When files
are
read
that
contain
both
formatted
and
list-directed records, the user program should ensure that each record
is read with the same kind of formatting that was used to write it.

READ

ACCEPT
7.3.1

List-Directed

Input

Statements

list-directed READ
statement
transfers
data
from
the
specified
logical
wunit,
translates
data
from
external to internal form, and
assigns the input values to the .elements of the I/0 list in the
order

which

they

appear,

statement does not have
logical unit number.

The list-directed ACCEPT
statement except that it
The

list-directed

input

from
a

unit

right.

statements

(u,*[,END=s][,ERR=s])

READ

*,list

have

the

list

*,list

logical

unit

number.

Indicates

list-directed

The

formatting.

label

list
An

I/0

list.

number,

statement is similar to
always uses an implicit

READ

left

logical

executable

statement.

list-directed
uses

READ

implicit

a list-directed
READ
logical unit number.

forms:

INPUT/OUTPUT STATEMENTS

The

external

separators.

contains

record

A value can be:

sequence

values

value

and

constant

A null

A repetition of constants in the form r*c

A repetition of null values in the form r*

value

The following paragraphs describe these values.

Each input constant has the form of the corresponding FORTRAN
A complex constant has the form of a pair of real or
constant.
in parentheses.
integer constants separated by a comma and enclosed
Spaces can occur between the opening parenthesis and the first
constant, before and after the separating comma, and between the

second

A logical constant is

the closing parenthesis.

and

constant

A character constant 1is delimited by
either T (true) or F (false).
any apostrophe in the character constant is represented
apostrophes;
and hexadecimal
octal,
Hollerith,
by two consecutive apostrophes.
constants are not permitted.

no
A null value is specified by two consecutive commas with
A null
Spaces can occur between the commas.
intervening constant.
value specifies that the corresponding list element remains unchanged.
A null value cannot be used for either part of a complex constant, but
can represent an entire complex constant.

1is a nonzero,
r
The form r*c indicates r occurrences of ¢ where
Spaces are not
integer constant and ¢ is a constant.
unsigned
permitted except within the constant c as specified above.

input statement
tabs -- terminates processing on the
all remaining I/0 list elements are unchanged
record;

and

they

are

The form r* indicates r occurrences of a null
unsigned

integer

constant.

A value

separator

can be:

value

where

tabs

One or more spaces or

A comma, with or without surrounding spaces or tabs

slash,

When any of the

with

above

appear

considered part of the constant,

without

surrounding

character

spaces

constant,

not value separators.

The end of a record is equivalent to a space character except when it
In this case, the end of the record
occurs in a character constant.

1is continued with the next
constant
is ignored and the character
in the previous record is
the last character
That is,
record.
followed immediately by the first character of the next record.

Spaces at'the beginning of a record are ignored unless they

are

part

1In this
of a character constant continued from the previous record.
case, the spaces at the beginning of the record are considered part of
the

constant.

Input constants can be any of the following data types:
complex, and character.
logical,
double precision,
real,
type of the constant determines the data type of the value
translation from external to internal

7-12

form.

integer,
The data
and the

INPUT/OUTPUT STATEMENTS

A numeric
a

list

character

element

list

can

element

correspond
can

only

correspond

only

a numeric
to

constant,

character

and

constant.

If the data types of a numeric
1list
element
and
its
corresponding
numeric
constant
do
not match, conversion is performed according to
the rules for arithmetic assignment (see Table 3-1).
Each input statement will read one
satisfy
the I/O list.
If a slash

or
more
separator

exhausted

before

all

the

values

the

record

ignored.

example

list-directed

The

program

unit

consists

CHARACTER*14
DOUBLE

The

D,E

LOGICAL

L,M

I,R,D,E,L,M,J,K,S,T,C,A,B

record

read

contains:

(3.4,4.2),

following

I/0

used,

the

remainder

follows.

(1,*)

6.3

are

of:

values

(3,

are

)

T,F,,3*14.6

assigned

the

,'ABC,DEF/GHI''JK'/

I/O list

elements:

List

OWH

Value
4

ORI

Element

statements

READ

The

input

record

to
is

PRECISION

COMPLEX

records
as
required
occurs or the I/0 list

.TRUE.

and

6.3
(3.4,4.2)
(3.0,2.0)
.FALSE.

14.6
14.6D0

ABC,DEF/GHI'JK
J

will

unchanged.

WRITE

TYPE
PRINT
7.3.2

List-Directed

Output

The list-directed WRITE
list
to
according

Statements

statement transfers the elements

the
specified
wunit,
translating
to the data type of the value.

and

editing

The list-directed
TYPE
and
PRINT
statements
are
list-directed
WRITE
statement,
except that output
implicit logical unit.

the

1I/0

each value

similar
to
the
is directed to an

INPUT/OUTPUT STATEMENTS

The

list—-directed

output

WRITE

(u,*[,ERR=s])

TYPE

*,list

statements

have

the

forms:

list

logical

unit

number.

Indicates list-directed

formatting.

The

label

executable

statement.

character

constants,

the

list
An

Except

I/0

for

list.

output values

have

the

same

forms

as
the
input
constant
values
described
above
in
Section 7.3.1.
Character constants are transferred
without
delimiting
apostrophes;
each
internal
apostrophe
1is
represented
by
only
one apostrophe.
Consequently, records containing list-directed character output can be
printed but cannot be used for list-directed input.
Table 7-1

lists the default output

formats

Table 7-1
List-Directed Output

Data

Type

for

each data type.

Formats

Output

Format

LOGICAL*1

LOGICAL*2
LOGICAL*4

L2
L2

INTEGER*2

INTEGER*4
REAL* 4’
REAL*8

I12
1PG15.7
1PG25.16

COMPLEX*8
CHARACTER

1X,'(',1pG14.7, ',', 1PG1l4.7,")"
1X,An
(n is the length of the
character expression)

The list-directed output statements do not produce octal values,
null
values,
slash
separators,
or repeated forms of values.
Each output
record
begins
with
a
space
for
carriage
control.
Each
output

statement
writes
one or more complete records.
Each output value is
contained within a single record, except for character constants
that
are

longer

than

record.

An example of list-directed output statements
PRINT *,'THEAARRAYAZAIS',Z
TYPE *, 'THEAANSWERAIS', (I,XX(I),I=1,10)

follows.

INPUT/OUTPUT STATEMENTS

program

unit consists of:

DIMENSION A(5)

DATA A/5%*3.4/
WRITE
WRITE
This

(1,*)
(1,*)

program

'ARRAYAVALUESAFOLLOW'
A,5

unit writes

the

following

records:

ARRAYAVALUESAFOLLOW

AAA3.400000AAAAAAA3.400000AAAAAAA3.400000AAAAAAA3.400000

AAA3.400000AAAAAAAAAAS

INPUT/OUTPUT STATEMENTS

7.4

UNFORMATTED SEQUENTIAL

INPUT/OUTPUT

Unformatted sequential I/0 transfers data in internal (binary) format.-.
That
1is,
no
conversion
or editing takes place.
Unformatted I/O is
generally used when data output by
a
program
will
be
subsequently
input
by
the
same
(or
a
similar) program.
Unformatted I/0 saves
execution time by eliminating the data translation process,
preserves
greater
precision
in
the
external data, and usually conserves file
storage

space.

READ
7.4.1
The

Unformatted Sequential Input Statement

unformatted

sequential

READ

statement

inputs

one

unformatted

record
from
the specified logical unit, and assigns the untranslated
fields of the record to the I/0 list elements in the
order
in
which
they
appear,
from
left
to
right.
The
data type of each element
determines the amount of data that element receives.
The

unformatted

READ

sequential

READ

statement

has

the

form:

(ul,END=s][,ERR=s])[list]

A logical

unit

number.

The label of an executable statement.

)

list

I/0

list.

An unformatted sequential READ statement reads exactly one record.
1If
the I/0 list does not use all the values in the record (that is, there
are more values
in
the
record
than
elements
in
the
1list),
the
remainder
of the record is discarded.
If the number of list elements
is greater than the number of values in the record, an error occurs.

If an unformatted sequential READ statement contains no I/O
skips over one full record, positioning the file to read the
record on the next execution of a READ statement.
The unformatted
records created
Examples
READ

sequential READ statement can only
be
used
by unformatted sequential WRITE statements.

unformatted

(1)

FIELDl,

sequential

This

statements

from

logical

unit

(8)

statement

read

follow.

FIELD2

This statement reads one record
to variables FIELD1 and FIELD2.
READ

input

1list,
it
following

advances

logical

one

record.

and

assigns

values

INPUT/OUTPUT STATEMENTS

WRITE
7.4.2

Unformatted Sequential Output Statement

untranslated
The unformatted sequential WRITE statement transfers the
values
of the elements in the I/O list to the specified logical unit.
That is, one unformatted record is output.

The

unformatted
WRITE

sequential WRITE

statement has

the

form:

(u[,ERR=s]) [list]

A logical

unit

The

number.

label

executable

statement.

list

I/0

list.

If an unformatted WRITE
record is output to the

statement
contains
specified unit.

Examples of

sequential

WRITE

This
LIST

unformatted
(1)

I/0

statements

list,

one

null

array

follow.

(LIST(K) ,K=1,5)

statement outputs the
to logical unit 1.
WRITE

output

contents of elements

through

(4)

This statement writes a null

record on logical

unit 4.

INPUT/OUTPUT STATEMENTS

7.5

FORMATTED DIRECT ACCESS

INPUT/OUTPUT

Formatted direct access I/0 transfers character data
to
and
from
a
file
on a direct access device.
The OPEN statement (see Section 9.1)
establishes the attributes of the direct access file.
Each
READ
or
WRITE statement contains a direct access record number.

Formatted Direct Access

Input Statement

The formatted direct access READ
record
from
the
direct
access
specified unit.
The
<characters
internal

form

according

are

assigned

The

formatted direct
READ

the

format

the elements of
access

statement
transfers
the
specified
file
currently
connected
to
the
in
the
record
are
translated
to

READ

the

specifier.

The

resulting

values

I/0 list.

statement

has

the

form:

(u'r,f[,ERR=s])[list]

A logical

unit number.

The

direct

access

record

number.

£
A format

specifier.

The

list

label

executable

statement.

I/0 list.

If the I/0 list and format
specify
more
characters
than
contains, or specify additional records, an error occurs.

record

INPUT/OUTPUT STATEMENTS

WRITE
7.5.2

Formatted Direct Access Output Statement

The formatted direct access WRITE statement
transfers
record to the direct access file currently connected to
list specifies a sequence of values that are translated
and positioned as specified by the format specifier.
The

formatted direct
WRITE

access WRITE

statement

has

the

the
specified
the unit.
The
to
characters

form:

(u'r,f[,ERR=s])[list]

A logical

unit

number.

The

direct

access

record

number.

f
A

format

specifier.

The

label

executable

statement.

list
An

I/0 list.

If the values specified by the list
the unused portion of the record is
If

the

I/0

record,

list

and

format

and format do not
filled with space

specify more

specify additional

characters

records,

error

fill the record,
characters.

than

can

occurs.

fit

into

INPUT/OUTPUT STATEMENTS

7.6

UNFORMATTED DIRECT ACCESS INPUT/OUTPUT

Unformatted direct access I/0 transfers data in internal (binary) form
The OPEN or DEFINE FILE statement
from a direct access file.
and
to

(see Sections 9.1 and 9.7) establishes the
direct

Each

record number.

access

attributes

the

file.

READ or WRITE statement contains a direct access

Unformatted Direct Access Input Statement

The unformatted direct access READ statement transfers

the

specified

the
to
file currently connected
from the direct access
record
to
record
specified unit, and assigns the untranslated fields of the
the

I/0 list elements.

The unformatted direct access READ statement has the form:
(u'r[,ERR=s])

READ
u

A logical
r

label

The
list

unit number.

The direct access

[list]

record number.

of an executable

statement.

I/0 list.

is,
(that
record
If the I/0 list does not use all the fields in the
there are more fields in the record than elements in the list), the
1If the number of list elements
remainder of the record is discarded.
is greater than the number of record fields, an error occurs.

Examples of unformatted direct access input statements follow.
READ

(1'10)

LIST(1l),LIST(8)

This statement reads record 10 of

file

1logical

wunit

and

file

logical

wunit

and

assigns two integer values to specified elements of array LIST.
READ

(4'58)

(RHO(N) ,N=1,5)

This statement reads record 58 of

assigns five real values to array RHO.

INPUT/OUTPUT STATEMENTS

WRITE
7.6.2

Unformatted Direct Access Output Statement

The
unformatted
direct access
WRITE
statement
transfers
the
untranslated
values
of the elements in the I/0 list to the specified
record of the direct access file currently connected to the
specified
unit.

The unformatted direct access WRITE statement has the form:
WRITE

(u'r[,ERR=s])

[list]

A logical

unit number.

The

direct

access

record

number.

executable

I/0

statement

1label.

list
list.

If the values specified by the list do not fill
portion of the record is filled with zeros.
If

the

list

specifies more data

than can

the

fit into

record,

the

record,

unused

error

occurs.

Examples of unformatted direct access output
WRITE

(2'35)

(3'J)

follow.

(NUM(K) ,K=1,10)

This statement outputs ten integer
connected to logical unit 2.
WRITE

statements

values to

record

the

file

ARRAY

This statement outputs the entire contents
of
ARRAY
to
the
record
indicated by the value of J in the file connected to logical unit 3.

INPUT/OUTPUT STATEMENTS

ENCODE
DECODE
ENCODE AND DECODE STATEMENTS

7.7

The ENCODE and DECODE statements transfer
data
according
to
format
specifiers,
translating the data from internal to character form, and
vice versa. ‘Unlike conventional formatted
1I/0
statements,
however,
these
data
transfers take place entirely between variables or arrays
in

the

FORTRAN program.

The ENCODE and DECODE

statements have the

forms:

ENCODE (c,f,b[ ,ERR=s]) [list]

DECODE (c,f,b[,ERR=s8])
[1list]
c

An integer expression.
1In the ENCODE statement, ¢ is the
number
of characters (bytes) to be translated to character form.
1In the
DECODE statement, ¢ is the number of characters to be
translated
to

internal

A format specifier.
error

form.

If more than one

record

specified,

occurs.

The name of an
array,
array
element,
variable,
or
character
substring
reference.
In
the
ENCODE statement, b receives the
characters after translation to external
form.
In
the
DECODE

statement, b contains the characters to be translated to internal
form.

The
list

label

executable

statement.

An I/0O list.
1In the ENCODE statement, the I/O list contains
the
data
to
be
translated
to
character
form.
In
the
DECODE
statement, the
1list
receives
the
data
after
translation
to
internal

The ENCODE

form.

statement translates the list elements

character

form

according
to
the
format
specifier, and stores the characters in b,
similar to
a
WRITE
statement.
If
fewer
than
¢
characters
are
transmitted, the remaining character positions are filled with spaces.

The DECODE statement translates the character data in
b
to
internal
(binary)
form
according
to
the
format
specifier,
and stores the
elements

If b

the

list,

is an array,

similar

a READ

statement.

its elements are processed

in the order of subscript

progression.

The number of characters that
the
ENCODE
or
DECODE
statement
can
process depends on the data type of b in that statement.
For example,
an INTEGER*2 array can
contain
2
<characters
per
element,
so
the
maximum
number
of characters is twice the number of elements in that

array.

A character variable or character array

characters

equal

number

1its

1length.

element

can

contain

A character array can

INPUT/OUTPUT STATEMENTS

contain characters equal
the

number

elements.

in number

to the length of each element times

The interaction between the format specifier and
same as for a formatted I/0 statement.
An example of

the ENCODE and DECODE

statements

the I/0 list

1is

the

follows.

DIMENSION K (3)
CHARACTER*12 A,

B
DATA A /'123456789012'/
DECODE (12,100,A) K
FORMAT (3I4)
ENCODE (12,100,B) K(3), K(2),

100

K(1)

The DECODE statement translates the 12 characters in A to integer form
(specified by statement 100), and stores them in array K, as follows:
K(1)
K(2)
K(3)

=
=
=

The ENCODE

character
as

1234
5678
9012

statement translates

form

follows:

B =

the values

(K(3),

K(2),

and

K(1l)

and stores the characters in the character variable B

'901256781234"

CHAPTER 8
FORMAT

STATEMENTS

FORMAT statements are nonexecutable statements used with formatted I/O
statements

and with ENCODE and DECODE statements.

A FORMAT statement

describes the format in which data is to be transferred, and what data
conversion and editing are required to achieve that format.
FORMAT

statements have

FORMAT

q
f

(qlf

form:

f.s_

...

fnqn)

slash

(/)

record

Zero or more

the

terminators.

A field descriptor or a group of field

descriptors

enclosed

parentheses.
S

field

separator.

The entire list of field descriptors and field
separators,
including
the parentheses, is called the format specification.
The list must be
enclosed

in parentheses.

A field descriptor

in a format specification has the

form:

[r]lcw][.d]
r

The

repeat count for

assumed
c

A format code

the

field descriptor.

If you omit r,

(I1,0,%2,F,E,D,G,L,A,H,X,T,P,Q,S$,

|is

:).

The external
d

The

field width.

The number of characters to the right of the decimal point.
terms

and d must all

unsigned

integer

constants;

and

must
be less than or equal to 32767, and 4 must be less than or equal
to 255.
The r term is optional;
however, you cannot use it
in
some
field
descriptors
(see
Section
8.1.18).
The
d term is required in
some field descriptors and is invalid in others.
You are not
allowed
to use PARAMETER constants for the terms r, w, or d.

FORMAT

The

field

descriptors

are:

Integer

Iw,

Logical

Real,

Character

double

Editing,
nX,

TTMn,

STATEMENTS

Ow,

precision,

and

complex

and
nP,

character
Q,

. Section 8.1 describes each
The

first character in an
control information:
see

and

Hollerith

number

field descriptor
output

record

Section

8.2

The field separators are comma
terminator.
Sections 8.3 and

field

You

can

and
8.4

separators.

create

format
instead
formats.

Fw.d,

Ew.d,

Dw.d,

Gw.d

nH,

‘'...',

-- Aw

positions)

the

format

during

FORMAT

constants
of

character

contains
information.

carriage

in detail.

generally

for

slash.
A slash is
describe in detail

program

statement.

characters

execution

Section

8.5

alsoc
a
record
the functions of

using

describes

run-time

During data transfers, the format specification is scanned
from
left
to right.
Data conversion is performed by correlating the elements in
the I/0 list with the corresponding field
descriptors.
In
H
field
descriptors
and
character
constants,
data
transfer
takes
place
entirely
between
the
field
descriptor
and
the
external
record.
Section
8.6
describes
1in
detail the interaction between the format
specifier and the I/0 list.

Section

8.1

8.7

summarizes

the

rules

for

writing

FORMAT

statements.

FIELD DESCRIPTORS

A field descriptor describes the
several
data
items;
each data
an external field.

size
item

and format of a data item
in the external medium is

or
of
called

The following sections describe
each
of
the
field
descrlptors
in
detail.
The
field descriptors ignore leading spaces in the external
field, but treat embedded and trailing spaces as zeros.

8.1.1
The

Field

field

Descriptor

descriptor

transfers

decimal

integer

values.

has

the

The corresponding I/0
logical data type.

list

element

must

either

integer

FORMAT STATEMENTS
w characters
In an input statement, the I field descriptor transfers
external field and assigns them, as a decimal value, to the
the
from
The external data
corresponding I/0 list element as an integer value.

must

have

the

form

integer

decimal point or exponent fields.

constant;

1t cannot contain a

the
range of
the
exceeds
field
the external
the value of
If
If the first nonblank
an error occurs.
element,
list
corresponding
treated
character of the external field is a minus sign, the field is
1If the first nonblank character is a plus sign,
as a negative value.
or if no sign appears in the field, the field is treated as a positive
Blanks
treated as a value of zero.
1is
field
all-blank
An
value.
following the first non-blank character are treated as zeros.
Input Example:

Internal Representation

External Field

Format

2788
-26
312

2788
-26
AAAAAA312
2AA8

14
I3
I9
14

2008

In an output statement, the I field descriptor transfers the value
of
the
corresponding
I/0
1list element, right justified, to an external
field,
the
fill
not
value does
the
If
field w characters long.

if the value exceeds the field width,
inserted;
spaces are
leading
1list
the
If the value of
the entire field is filled with asterisks.
leftmost,
its
as
sign
minus
a
have
will
field
the
negative,
is
element
to
therefore be 1large enough
The term w must
nonblank character.
Plus signs, however, are
a minus sign, when necessary.
for
provide
suppressed.

Output Example:

Format

Internal Value

284
-284
AA174
**
*kk
Not permitted:

284
-284
174
3244
-473
29.812

I3
14
I5
12
I3
17

8.1.2

External Representation

error

O Field Descriptor

The O field descriptor transfers octal values.

It has the form:

The corresponding I/0 list
logical

data

type.

element

must

either

integer

In an input statement, the O field descriptor transfers
w
characters
assigns them as an octal value to the
and
field
external
the
from

The external field can contain only
corresponding I/0 list element.
it cannot contain a sign, a decimal point,
the numerals 0 through 7;
of
value
a
as
An all-blank field is treated
or an exponent field.
following the first non-blank character are treated as
Blanks
zero.
the
range of
If the value of the external field exceeds the
zeros.
corresponding

list element,

an error occurs.

FORMAT

Input

STATEMENTS

Example:

Format

External

Field

32767

04
06
03

16234
13AAAA
97A

Internal
Representation

Octal

32767

Not

1623
130000
permitted:

In an output statement, the O field
descriptor
value
of
the
«corresponding I/0 list element,
external field w characters long.
No signs are

error

transfers
the
octal
right justified, to an
output;
a
negative

value is transmitted in its octal (2's complement) form.
If the value
does not fill the field, leading spaces are inserted;
if
the
value
exceeds the field width, the entire field is filled with asterisks.
Output

Example:

Format

8.1.3
The

(Decimal)

06
06

32767
-32767

14261

13.52

Internal

Value

External

Representation
ATT7777
100001
* %
AA33

Not

permitted:

error

Field Descriptor

field

descriptor

transfers

hexadecimal

values.

has

the

form:

The

corresponding

logical

data

I/0

list

element

must

either

integer

type.

In an input statement, the Z field descriptor transfers
w
characters
from the external field and assigns them as a hexadecimal value to the
corresponding I/O list element.
The external field can
contain
only
the
numerals
0
through
9
and
the letters A through F;
it cannot
contain a sign, a decimal point, or an exponent field.
An
all-blank
field
1is
treated
as
a
value
of zero.
Blanks following the first
non-blank

external

character

field

error

occurs.

Input

Example:

Format

are

exceeds

treated

the

range

External

A94
A23DEF

9AFAAAA

95.AF2
statement,

zeros.

the

corresponding

value

list

the

element,

Internal
Hexadecimal Representation

Field

z3
Z5

output

A94
A23DE
9AF0000
Not
Z

field

permitted:

descriptor

error

transfers

the

hexadecimal
value
of
the
corresponding
I/0
1list
element,
right
justified, to an external field
w
characters
1long.
No
signs
are
output;
a
negative
wvalue
1is
transmitted
in its hexadecimal (2's
complement)
form.
If the value
does
not
fill
the
field,
1leading
spaces are inserted;
if the value exceeds the field width, the entire
field is filled with asterisks.

FORMAT STATEMENTS

Output Example:

Format

Internal

(Decimal) Value

7FFF
A8001
10

32767
-32767
16

z4
z5
Z2

Not permitted: error

25.2

External Representation

F Field Descriptor

§.1.4

The F field descriptor transfers real or double precision values.
has

the

form:

Fw.d

real or double
The corresponding I/0 list element must be of either
or it must be either the real or the imaginary
precision data type;
part of a complex data type.

In an input statement, the F field descriptor transfers w characters
from the external field, and assigns them, as a real or double

precision value, to the corresponding I/O list element. If the first
nonblank character of the external field is a minus sign, the field is
treated as a negative value. If the first nonblank character is a

plus sign, or if no sign appears in the field, the field is treated as
a positive value. An all-blank field is treated as a value of =zero.
In all F field descriptors, w must be greater than or equal to d+l.

If the field contains neither a decimal point nor an exponent, it is
treated as a real number of w digits, in which the rightmost 4 digits
If the field contains an
are to the right of the decimal point.
explicit decimal point, the location of that decimal point overrides
the location specified by the field descriptor. If the field contains
an

exponent

(as

described

Section

2.3.2 for real constants or

Section 2.3.3 for double precision constants), that exponent is used
to establish the magnitude of the value before it is assigned to the
list

element.

Input Example:

Format

External Field

Internal Representation

F8.5
F8.5
F8.5
F5.2

123456789
-1234.567
24.77E+2
1234567.89

123.45678
-1234.56
2477.0
123.45

In an output statement, the F field descriptor transfers the value of
the corresponding I/0 list element, rounded to d decimal positions and
If the value
right justified, to an external field w characters long.
if the value
does not fill the field, leading spaces are inserted;

exceeds the field width, the entire field is filled with asterisks.

(when
The term w must be large enough to include a minus sign
left
the
to
digit
one
least
at
,
suppressed)
are
signs
plus
necessary;
of the decimal point, the decimal point, and d digits to the right of
the decimal.

Therefore, w must be greater than or equal to d+3.

FORMAT

Output

STATEMENTS

Example:

Format

Internal

F8.5
F9.3
F2.1
F10.4

External

A2.35472
A8789.736
* &

-23.24352
325.013

AA-23.2435
kkkkk

-.2

Field

The E field
exponential

Representation

2.3547188
8789.7361
51.44

F5.2
F5.2

8.1.5

Value

-0.20

Descriptor

descriptor transfers real
form.
It has the form:

double

precision

values

Ew.d

The corresponding I/0 list
precision
data
type;
or
part

complex

data

element must be of either
it must be either the real

type.

real
or
double
or the imaginary

In an input statement, the E field descriptor transfers
w
characters
from the external field and assigns them as a real or double precision
value to the corresponding I/0O list element.
This is exactly the same
way that the F field descriptor interprets and assigns data.
Input

Example:
Format

External

Field

E9.3
El2.4
E15.3

734.432E3
AA1022.43E-6
52.3759663AAAAA

E12.5

210.5271D+10

Note that, in the last example,
exponent
field indicator as an
single precision.
In

Internal

output

statement,

the

Representation

734432.0

1022.43E-6
52,.3759663
210.5271E10

the E field
E indicator

field

descriptor
if the I/0

descriptor

treats
the
list element

transfers

the

value

D
is

the
corresponding
I/0 list element, rounded to d decimal digits, and
right justified, to an external field w characters long.
If the value
does
not
fill
(the field, leading spaces are inserted;
if the value
exceeds the field width, the entire field is filled with asterisks.
When

you

standard

use

the

form.

field
This

descriptor,
form

data

consists

output
a

minus

transferred

sign

value

point,

the

negative (plus signs are suppressed),
a
zero,
a
decimal
digits to the right of the decimal points, and a 4-character
The exponent has one of the following two forms:

exponent.

E+nn

E-nn
nn

2-digit

integer

constant.

The d digits to the right of the decimal
value, scaled to a decimal fraction.

8-6

point

represent

the

entire

FORMAT STATEMENTS

(when
sign
a minus
include
to
1large enough
The term w must be
suppressed), a zero, a decimal point, d
signs are
plus
necessary;
equal
or
Therefore, w must be greater than
digits, and an exponent.
to d+7.

Output Example:

Internal Value

Format

A0 .48E+06
A0.47587E+06
AAAO.690E-03

475867.222
475867.222
0.00069

E9.2
E12.5
E12.3

-0.556E+00
ok ok
*kk

-0.5555
56.12

E10.3
E5.3

8.1.6

External Representation

D Field Descriptor

The D field descriptor transfers real or double
exponential

form.

It has the

form:

precision

values

Dw.d

double
or
real
The corresponding I/O list element must be of either
or it must be either the real or the imaginary
type;
precision data
part of a complex data type.

w characters
In an input statement, the D field descriptor transfers
from the external field, and assigns them as a double precision value
the
This is the same way that
to the corresponding I/O list element.
F field descriptor interprets and assigns data.
Input Example:

Format

External Field

Internal Representation

D10.2
D10.2
D15.3

12345AAAAA
AA123 .45AA
367.4981763D+04

12345000.0D0
123.45D0
3.674981763D+06

In an output statement, the D field descriptor has the same effect

the

used

E field descriptor, except that the D exponent field indicator is
in place of the E

indicator.

Output Example:

Format
D14.3

D23.12
D9.6

Internal Value
0.0363

5413.87625793
1.2

External Representation
AAAAA0.363D-01

AAAAAQ.541387625793D+04
kkkkkkkkk

FORMAT

8.1.7
The

G Field Descriptor

form

the

STATEMENTS

field

descriptor

transfers

that,

combines

effect,

form:

real
the

double

and

precision values

field

descriptors.

in
It

a
has

Gw.d
The corresponding I/0 list element must be of either
precision
data
type;
or it must be either the real
part of a complex data type.

real
or
double
or the imaginary

In an input statement, the G field descriptor transfers
w
characters
from
the
external
field,
and
assigns
them
as
a
real or double
precision value to the corresponding I/0 list element.
This
1is
the
same way the F field descriptor interprets and assigns data.
In

the

output

statement,

corresponding

right

which
value,

I/0

the

list

field

element,

justified,

descriptor

transfers

the

value

rounded

positions

and

The form
the magnitude of

the

to an external field
the
value
is
written
is
a
as described in Table 8-1.

Effect

Data

Table
Magnitude

Data

decimal
w characters long.

function of

8-1
on G

Magnitude

are

Format

Effective

Conversion

0.1

<m<«<

1.0

F(w-4).d,

1.0

10.0

F(w=4).(d-1),

10d-2

< m

104-1

F(w=4).1,

'AAAA!

10d-1

< m

10d

F(w=4).0,

'AAAA!

m >

10d

the

numeric

Conversions

Ew.d

'AAAA'

'AAAA!

Ew.d

The 'AAAA' field descriptor, which is,
field
descriptor
for
values
within
spaces

data

in effect, inserted
by
the
G
its range, specifies that four
representation.
-

necessary;

the

plus

decimal

signs

point,

decimal
point,
and
field descriptor)
a
greater than or equal

are

the

suppressed),

decimal

point,

least

one

digits

(for values outside the
4-character
exponent.
to 4 plus 7.

digit

the

right

left

the
range of the G
Therefore,
w
must
be

effective

FORMAT STATEMENTS

Output Example:

Internal Value

Format

External Representation
A0.123457E-01

0.01234567

Gl13.6

-0.123457AAAA
AA1.23457AAAA
AA12.3457AAAA
AA123.457AAAA
A-1234.57AAAA
AA12345.7AAAA
AA123457.AAAA

-0.12345678
1.23456789
12.34567890
123.45678901
-1234.56789012
12345.67890123
123456.78901234

Gl3.6
Gl3.6
Gl3.6
Gl3.6
G13.6
G1l3.6
Gl3.6

-0.123457E+07

-1234567.89012345

Gl3.6

Compare the above example with the following example, which shows
same values output using an equivalent F field descriptor.
External Representation

Internal Value

Format

AAAAAD.012346

0.01234567

F13.6

AAAA-0.123457
AAAAAL. 234568
AAAAL2.345679

~0.12345678
1.23456789
12.34567890

F13.6
F13.6
F13.6

AAA123.456789
A-1234.567890
A12345.678901
123456.789012
gk ke dkod ek ok ok ok

123.45678901
-1234.56789012
12345.67890123
123456.78901234
~1234567.89012345

F13.6
F13.6
F13.6
F13.6
F13.6

the

L Field Descriptor

8.1.8

The L field descriptor transfers logical data.

It has the form:

The corresponding I/O list
logical

data

element

must

either

integer

type.

In an input statement, the L field descriptor transfers w characters
1If the first nonblank character of the field
from the external field.

1is assigned to the corresponding
is the letter T, the value .TRUE.
character of the field is the
nonblank
first
the
If
I/0 list element.
is
.FALSE.
letter F, or if the entire field 1is blank, the value
assigned.

Any other value in the external field produces an error.

In an output statement, the L field descriptor

transfers

either

the

letter T (if the value of the corresponding I/0 list element is
to an external
.TRUE.), or the letter F (if the value 1is .FALSE.)
The letter T or F is in the rightmost
field w characters 1long.

position of the field, preceded by w-1 spaces.
Output Example:

Format

Internal Value

.TRUE.

.FALSE.

External Representation
AAAAT

FORMAT

8.1.9
The
has

STATEMENTS

A Field Descriptor

A field descriptor
the form:

transfers

character

Hollerith

values.

The

corresponding

character

data

any

other

data

type,

The

value

In an
from

list

I/O

w must

list

element

can be of any data type.
If it
character data is transmitted.
If it is
Hollerith data is transmitted.
type,

less

than

input statement, the A
the
external
record

element.

depends

size

elements,

the

size

substring
elements,

the

size

the

I/O

32767.

the

transfers
w
characters
to the corresponding I/0

characters

list

element.

is the length of the
reference, or character array

depends

field descriptor
and assigns them

The maximum number
the

equal

data

I/0 List

is
of

that
For

character

can

character

variable,

stored
I/0

list
character

element.
For numeric
type, as follows:

I/0

list

Maximum Number

Element

Characters

BYTE

LOGICAL*1
N

LOGICAL*2
LOGICAL*4

0 0O CO > o> DN

INTEGER*2
INTEGER*4
REAL

REAL*8
DOUBLE

PRECISION

COMPLEX

If w is greater
stored
in
the
characters
are
characters
are

than

the maximum number
of
characters
that
can
be
corresponding
I/0
1list
element, only the rightmost
assigned
to
that
element.
The
leftmost
excess
ignored.
If w is less than the number of characters
can be stored, w characters are assigned
to
the
1list
element,
justified, and trailing spaces are added to fill the element.

that
left
Input

Example:
Format

Field

Internal

Representation

(CHARACTER*1)

PAGEA#

EA#

PAGEA#

(CHARACTER*3)

PAGEA#

(CHARACTER*6)

PAGEA#AA

PAGEA#

(CHARACTER*8)

(LOGICAL*1)

A®

PAGEA#

(INTEGER*2)

GEA#

(REAL)

PAGEA#

PAGEA#AA

(DOUBLE

output

the

appears
than

statement,

corresponding

characters

less

External

long.

the
transferred.

the

field,

1list

the

I/0

field

1list

greater

right

element,

descriptor

element

than

justified,
only

8-10

the

with

PRECISION)

transfers

1list

the
external

element,

leading

leftmost

spaces.

contents

field

the

data

If w is
characters
are

" FORMAT STATEMENTS

Output Example:

External Representation

Internal Value

Format

AOHMS

OHMS

VOLTS
AMPER

VOLTSAAAA
AMPERESA

A5
A5

is supplied.
If you omit w in an A field descriptor, a default value
If the I/0 list element is of character data type, the default value
If the I/0 list element is of
is the length of the I/O0 list element.
is the maximum number of
the default value
numeric data type,
characters that can be stored in a variable of that data type.

8.1.10

H Field Descriptor

The H field descriptor transfers data between the external record and
It has the form of a Hollerith
itself.
the H field descriptor
constant:
4

anlczc3

The number of characters to be transferred.
An ASCII

character.

In an input statement, the H field descriptor transfers n characters
from the external field to the field descriptor. The first character

appears immediately after the letter H.

Any characters in

the

descriptor before input are replaced by the input characters.

field

In an output statement, the H field descriptor transfers n characters
following the letter H from the field descriptor to the external
field.

An example of H field descriptor usage follows.
TYPE

100

FORMAT

200

The

100

FORMAT

(41HAENTERAPROGRAMATITLE
, AUPATOA20ACHARACTERS)
200

(20HAATITLEAGOESAHEREAAA)

transfers

statement

TYPE

descriptor

statement

100

the

from

characters
wuser's

the

statement accepts the response from the keyboard,

the

terminal.

placing

field

the

input

The ACCEPT

The new characters’
in the H field descriptor in statement 200.
data
less than 20
enters
user
the
If
HERE.
GOES
TITLE
words
replace the

characters,

remainder

the

spaces to the right.

8.1.10.1

instead

Character

the H field descriptor is filled with

Constants - You

field

function identically.

descriptor.

can

use

character

constant

Both types of format specifier

In a character constant, the apostrophe is written as two apostrophes.

For

example:

FORMAT

('TODAY''SADATEAIS:A',I1I2,'/',12,'/',12)

A pair of apostrophes used this way is considered a single character.
8-11

FORMAT STATEMENTS

8.1.11
The

X Field Descriptor
field descriptor

a positional

specifier.

It has

the

form:

The term n specifies how many character positions
are
over.
The value of n must be greater than or equal to

to
1.

In an input statement, the X field descriptor specifies
n characters in the input record are to be skipped.
In

the

output

statement,

external

record.

the

For

field

descriptor

passed

that

the

spaces

transfers

example:

WRITE (6,90) NPAGE
FORMAT (13H1PAGEANUMBERA,I2,16X,23HGRAPHICAANALYSIS,ACONT.)

90
The WRITE
PAGE

statement
NUMBER

prints

record

similar

to:

GRAPHIC

ANALYSIS,

CONT.

The term nn is the current value of the variable NPAGE.
The numeral 1
in the first H field descriptor is not printed, but is used to advance
the printer paper to the top of a new
page.
Section
8.2
describes
printer carriage control.

8.1.12

The

T Field

field

Descriptor

descriptor

tabulation

specifier.

has

the

form:

The term n indicates the character position of
the
external
record.
The
value
of
n
must be greater than or equal to 1, but not greater
than the number of characters allowed in the external record.

In an input statement, the T field descriptor positions
the
record
to
its nth character position.
For example, a READ
inputs a record containing:

external
statement

ABCAAAXYZ

This

record
10

the

control

the

FORMAT

statement:

(T7,A3,T1,A3)

the

READ

statement would

input

the

characters

XYz

first,

characters ABC.

output

terminal,

transfer

under

FORMAT

execution,

then
In

statement

the

field

to begin at

devices

other

descriptor

than

specifies

the nth character

the

1line

that

position

printer

subsequent

the

data

external

record.
On
output
to
a
printer, data transfer begins at position
(n-1).
The first position of a
printed
record
is
reserved
for
a
carriage
control character, which is never printed (see Section 8.2).
For

example:

PRINT 25
FORMAT (T51, 'COLUMNA2" 'COLUMNAl"')
',T21,

FORMAT STATEMENTS

These statements would print the following line:
Position 50

Position 20
COLUMN 1

8.1.13

COLUMN 2

Q Field Descriptor

The Q field descriptor obtains the number of characters in
the
input
record
remaining
to
be transferred during a READ operation.
It has
the

form:

The corresponding I/O list element must be of integer or logical

data

type.

For

example:

1000

READ (4,1000) XRAY,KK,NCHRS, (ICHR(I),I=1,NCHRS)
FORMAT (E15.7,14,Q,80Al)

These input statements read two fields into the variables XRAY and KK.
The
number
of characters remaining in the record is stored in NCHRS
By
ICHR.
array
the
and exactly that many characters are read into
the Q descriptor first in the format specification, you can
placing
determine the actual length of the input record.
no
In an output statement, the Q field descriptor has
that the corresponding I/0 list element is skipped.

8.1.14

Dollar

effect

except

Sign Descriptor

The dollar sign character ($) in a format specification modifies
the
carriage
control
specified by the first character of the record.
1In
an output statement, the $ descriptor suppresses the
carriage
return
if the first character of the record is a space or a plus sign.
1In an
input statement, the $ descriptor is ignored.
The
$
descriptor
is
intended
primarily for interactive I/0;
it leaves the terminal print
position at the end of the text (rather than returning it to the
left
a typed response will follow the output on the same
that
so
margin)
line.

Thus,

the

100
200

statements:
TYPE

100

FORMAT

('AENTERARADIUSAVALUEA',S)

200

FORMAT

(F6.2)

produce a message on the

terminal

in the form:

ENTERARADIUSAVALUE

Your response

(for example, 12.)

ENTER RADIUS VALUE

12.

can then go on the same line, as:

FORMAT STATEMENTS

8.1.15

Colon Descriptor

The colon character (:) in a format
specification
terminates
format
control
if
no more items are in the I/O list.
The :
descriptor has
no effect if I/0 list items remain.
For example:
PRINT
PRINT
1

FORMAT ('AI="',I2,

'AJ=',I2)
FORMAT ('AK=",I2,:,"'AL=",12)

These

1,3
2,4

statements print

the

following

two

lines:

I=A3Ad=
K=A4

Section

8.6 describes

8.1.16

Complex Data Editing
-

A complex value
or
output
of

descriptors,

is
a

format

control

an ordered pair
complex
value

in detail.

of real values.
Therefore,
is
governed
by
two
real

using any combination of the forms Fw.d,

Ew.d,

input
field

Dw.d,

Gw.d.

input

statement,

complex

I/0

the

1list

respectively.
Input

successive
as

fields

its

are

real

read

and

assigned

imaginary parts,

Example:
Format

External

F8.5,F8.5
E9.1,F9.3
In

two

element

Field

Internal

1234567812345.67
734.432E8123456789

output

statement,

the

two

Representation

123.45678,
734.432E8,

parts

12345.67
12345.678

complex

value

are

transferred
under
the
control
of
repeated
or
successive
field
descriptors.
The two parts
are
transferred
consecutively,
without
punctuation or spacing, unless the format specifier states otherwise.
Output

Example:

Format

Internal

2F8.5
E9.2,'A,A" ,E5.3

8.1.17

Scale

Value

2.3547188,
47587.222,

External

3.456732
56.123

Representation

A2,35472 3.45673
AO.48E+06A, A*****

Factor

The
of

scale factor
the
decimal

two

parts

The

scale

factor

lets you
point in

complex

has

alter, during input or output, the
1location
real and double precision values, and in the

values.

the

form:

npP

A signed or
+127.
It
right, that

unsigned integer constant in the range
specifies
the
number
of
positions,
the

decimal

point

8-14

to move.

-127
to the

through
left or

FORMAT STATEMENTS

A scale factor can appear anywhere in a format specification, but must
precede the field descriptors that are to be associated with it. For
example:

nPGw.d

nPDw.d

nPEw.d

nPFw.d

On input, the scale factor in any of the above field descriptors
multiplies the data by 10**-n and assigns it to the corresponding I/0
list element. For example, a 2P scale factor multiplies an input
value by .01, moving the decimal point two places to the left. A -2P
scale factor multiplies an input value by 100, moving the decimal

two

point

places

the

right.

However,

the external field

contains an explicit exponent, the scale factor has no effect.
Input Exauple:

Form;t

External Field

Internal Representation

3PE10.5
3PE10.5
-3PE10.5

AAA37.614
AA37.614E2
AAAA37.614

.037614
3761.4
37614.0

On output, the effect of the scale factor depends on the type of field
descriptor associated with it. For the F field descriptor, the value
of the I/0 list element is multiplied by 10**n before transfer to the
Thus, a positive scale factor moves the decimal
external record.
point to the right; a negative scale factor moves the decimal point
to

the

left.

For the E or D field descriptor, the basic real constant part of the
I/0 list element is multiplied by 10**n, and n is subtracted from the
exponent. Thus, a positive scale factor moves the decimal point to
the

right

and decreases the exponent;

a negative scale factor moves

the decimal point to the left and increases the exponent.
Output Example:

Format

Internal Value

External Representation

1PE12.3
1PE12.2
-1PE12.2

-270.139
-270.139
-270.139

AA-2.701E+02
AAA-2.70E+02
AAA-0.03E+04

The effect of the scale factor for the G field descriptor is suspended

if the magnitude of the data to be output is within the effective
range of the descriptor, because the G field descriptor supplies 1its
own scaling function. The G field descriptor functions as an E field
descriptor if the magnitude of the data value is outside its range.
In this case, the scale factor has the same effect as for the E field

- descriptor.

On input, and on output under F field descriptor control, a scale
factor actually alters the magnitude of the data. On output, a scale
factor under E, D, or G field descriptor control merely alters the
form in which the data is transferred. 1In addition, on input, a
positive scale factor moves the decimal point to the 1left and a
negative scale factor moves the decimal point to the right; and, on
output,

the effect is the reverse.

8-15

FORMAT

If you do not specify a
scale
default
scale
factor
of
0
factor,
however,
it
applies
precision
field
descriptors
another scale factor appears.

STATEMENTS

factor
with
a
field
descriptor,
a
1is
assumed.
Once you specify a scale
to
all
subsequent
real
and
double

in
For

the
same
example:

FORMAT

statement,

unless

DIMENSION A(6)

A(I)

=1,6

25.

TYPE

100,A
FORMAT('
',F8.2,2PF8.2,F8.2)

100

produces:

25.00

2500.00

If a second scale factor appears in the
control from the first scale factor.
Format reversion
A
scale
factor
specification.

has
of

8.1.18

Repeat Counts

You

apply most

can

no
0

the

following

and

field

descriptors

can apply a group of
enclosing

them
with
1is
called a

statements

are

these

field

an
unsigned
group repeat

descriptors

to
and

equivalent:

FORMAT (I8,f8,F8.3,E15.7,F8.3,E15.7,F8.3,E15.7)

If you do
assumed.

not

repeatedly

FORMAT

specification.

parentheses

or X field
enclosed

For example,

integer
constant.
count.
For example,

The
integer
the following

(218,3(F8.3,E15.7))

equivalent:

(3E12.4,415)

takes

Section
8.6).
an explicit OP

is called a repeat count.

are

FORMAT

constant

preceding
two

constant

statements

(except

(E12.4,E12.4,E12.4,15,15,1I5,1I5)

you

(see
by

fields by preceding that field descriptor
constant
specifying
the
number
of

FORMAT

fields

statement,

Group Repeat Counts

Similarly,
data

This

two

FORMAT

effect on the scale factor
can
only
be
reinstated

number
of
successive
data
with
an
unsigned
integer

repetitions.

2500.00

descriptor,

in
Thus,

which
parentheses

specify a

could
group

could

and

3
not

repeated

repeat

otherwise

treated

count,

desired
a

default

repeated,

group

number
count

of
of

can

repeat
times.
1

Section 8.6 discusses how to use parentheses when the number of values
to be formatted is greater than the number of format specifications.

FORMAT STATEMENTS

8.1.19

Variable Format Expressions

You can use an expression in a FORMAT statement wherever you

can

use

an integer (except as the specification of the number of characters in
the H field) by enclosing it in angle brackets.
For example:
(I<J+1>)

FORMAT

field
transfer with a
data
This statement performs an I (integer)
than the value of J at the time the format is
width one greater
encountered
is
The expression is re-evaluated each time it
scanned.

in
the
normal format scan.
If the expression is not of integer data
You can use any
type it is converted to integer data type before use.
valid
FORTRAN
expression, including function calls and references to

dummy arguments.

Figure 8-1 shows an example of a variable format expression.

The value of a variable format expression must obey
on magnitude applying to its use

in the format,

the

restrictions

or an error occurs.

DIMENSION A(5)

100

DO 10 I =1,10
WRITE (6,100) I
FORMAT (I<MAX(I,5)>)

CONTINUE

DO 201 =1,5
WRITE (6,101)

(A(I),J=1,I)
(<KI>F10.<I-1>)

101

FORMAT

CONTINUE
END

On execution, these statements produce the following output:
1
2
3
4

5
6
7
8

9
10

1.
2.0

3.00
4.000
5.0000

2.0
3.00
4.000
5.0000

Figure 8-1

3.00
4.000
5.0000

4.000
5.0000

Variable Format Expression

5.0000

Example

FORMAT STATEMENTS

8.1.20

Default Field Descriptors

If you write the field descriptors 1, O, %2,
L,
F,
E,
D,
G,
or
A
without specifying a field width value, default values for w and d
are
supplied based on the data type of the I/0 list element.
Table

8-2

lists

the default values

Default

INTEGER*2
INTEGER*4
LOGICAL

REAL,

DOUBLE

COMPLEX
PRECISION

LOGICAL*1

LOGICAL*2,INTEGER*2
LOGICAL*4,INTEGER*4

REAL,
-

COMPLEX

DOUBLE

PRECISION

CHARACTER*n

for

the A

corresponding

field

I/0

descriptor,

list

the

default

element.

the

actual

length

CARRIAGE CONTROL

The first
printed.
The

7
12

8.2

that

the

Values

List Element

Note

w and

Table 8-2
Field Descriptor

Field Descriptor
I,
I,

for

character

FORTRAN

control

of
every record transferred to a printer is not
it is interpreted as a carriage control character.
system
recognizes
certain
characters
as
carriage

Instead,
I/0

characters.

Table

effects.

8-3

1lists

these

characters

and

Table 8-3
Carriage Control Characters

Character

Zffect

(space)

Advances

one

line

(zero)

Advances

two

lines

(one)

Advances

top of

(plus)

(dollar

page

Does not advance
(allows overprinting)

sign)

Advances one
and

line before printing

suppresses

return

record

8-18

the

carriage
end

the

their

FORMAT

Any character other than
space,
and
1is
deleted

those
from

listed in Table 8-3
the
print
1line.

accidentally omit the carriage
of the record is not printed.

8.3

control

character,

is
treated
Note
that

the

descriptors

commas.

field

You

and

character

statement

slash

(/)

terminates

new

record.

is equivalent

WRITE

(6,40)

FORMAT

(306)

WRITE

(6,50)

use

specification

the

are generally separated

record

input

For

terminator

output

result

K,L,M

N,O,P

slashes

output.

slash

The

bypass

first

terminates

slashes

current

to:

input

records

the

slash

first

terminates

skipped

the

beginning

blank

records.

end

This

output

between two
or
(n-1)

field
blank

the

current

record;

blank

record,

and

format

on.

However,

the

(I6,2F8.4)

multiple

are

second

separate

example:

blank
records.
If
n
consecutive
slashes appear
descriptors,
(n-1) records
are
skipped
on
input,
records

K,L,M,N,O,P
(306/16,2F8.4)

FORMAT

can

format
use

(6,40)

FORMAT

also

initiates a

WRITE

This

can

descriptors.

record,

the

first

as
a
if you

FORMAT SPECIFICATION SEPARATORS

Field

You

STATEMENTS

specification

because the opening and
closing parentheses of
the
format
specification
are
themselves
a
record initiator and terminator, respectively.
For example:

The

above

WRITE (6,99)
FORMAT ('1',T51, 'HEADINGALINE'//T51,'SUBHEADINGALINE'//)

statements

Column

50,

top

produce
of

page

HEADING

(blank

line)

(blank
(blank

line)
line)

the

LINE

SUBHEADING

LINE

following

output:

FORMAT STATEMENTS

8.4

EXTERNAL FIELD SEPARATORS

A field descriptor such as Fw.d specifies that an input statement is
to read w characters from the external record. If the data field in
the external record contains less than w characters, the input
statement would read characters from the next data field in the
external record, unless the short field is padded with leading zeros
When the field descriptor is numeric, you can avoid
or spaces.
The
padding the input field by using a comma to terminate the field.
n.
specificatio
width
field
descriptor's
field
the
overrides
comma
This is called short field termination, and 1is particularly useful
when you are entering data from a terminal keyboard. You can use it
For example:
with the I, O, 2, F, E, D, G, and L field descriptors.
(5,100) I,J,A,B
FORMAT (2I6,2F10.2)
READ

100

The above statements read the following record:
1,-2,1.0,35

On execution, the following assignments occur:
I

-2

A =1.0

0.35

Note that the physical end of the record also serves as a field
and that the d part of a w.d specification is not
terminator;
affected by an external field separator.

You can use a comma to terminate only fields less than w characters
If a comma follows a field of w characters or more, the comma
long.
is considered part of the next field.

Two successive commas, or a comma after a field of exactly w
characters, constitutes a null (zero-length) field. Depending on the
0.0,
field descriptor specified, the resulting value assigned is 0,
0.D0,

.FALSE..

You cannot use a comma to terminate a field that is controlled by an
A, H, or character constant field descriptor. However, if the record
reaches its physical end before w characters are read, short field
termination occurs and the characters that were input are assigned
successfully. Trailing spaces are appended to fill the corresponding
I/0 list element or the field descriptor.

FORMAT

8.5
You

RUN-TIME
can

STATEMENTS

FORMAT

store

format

specifications

character

variables,

character

arrays,
character
array
elements,
character substrings, or numeric
arrays.
Such a format specification is called a run-time format,
and
can be constructed or altered during program execution.

A run-time format in an array has the same form as a FORMAT statement,
without
the
word
FORMAT
and
the statement label.
The opening and
closing parentheses are required.
Variable format expressions are not
permitted.
In

the

following

parenthesis
parenthesis

to
and

for

|use.

1later

example,

the
three

inclusion in the

the

DATA

statement

character
variable
field descriptors to

The

proper

field

descriptors

format specification.

assigns

left

FORCHR, and assigns a right
four
character
variables
are

then

selected

The selection is based on

for

the

magnitude
of
the
individual
elements
of the array TABLE.
A right
parenthesis is then added to the format specification just before
the
WRITE
statement uses it.
Thus, the format specification changes with
each iteration of the DO loop.
REAL

TABLE (10,5)

CHARACTER*5 FORCHR(0:5), RPAR*1,
DATA FORCHR(0) ,RPAR/'(',')'/
DATA
DO

FBIG,FMED,FSML

FBIG,FMED,FSML/'F8.2,','F9.4,','F9.6,"'/

I=1,10

J=1,5

(TABLE(I,J)

.GE.

100.)

THEN

FORCHR
(J) =FBIG
ELSE

(TABLE(I,J)

.GT.

0.l1)

THEN

FORCHR
(J) =FMED
ELSE
FORCHR
(J) =FSML
END

CONTINUE
FORCHR
(5) (5:5)=RPAR
WRITE

(6,FORCHR)

(TABLE(I,J),

J=1,5)

CONTINUE
END

NOTE

Format specifications stored
in
arrays
are
recompiled
at
run
time each time
they
are
character
READ

statement

format

back

used.
If
a
Hollerith
run-time
format is used
to

itself,

into

the

read

that

data

original

will
not
be available
using that array as
a
specification.

array.

into

not

or
a

the

copied

Thus,

subsequently for
run-time
format

FORMAT STATEMENTS

8.6

FORMAT CONTROL INTERACTION WITH INPUT/OUTPUT LISTS

I/O statement.
formatted
Format control begins with execution of a
information provided
takes depends on
The action format control
the
and
exists)
jointly by the next element of the I/O list (if one
Both the I/O list
next field descriptor of the format specification.

and the format specification

are

interpreted

except when repeat counts are specified.

irom

1left

right,

1least
If the I/0 statement contains an I/O list, you must specify at
F, E, D, G, L, A, or Q field descriptor in the format
%2,
O,
I,
one
An error occurs if these conditions are not met.
specification.
the
from
record
On execution, a formatted input statement reads one
Thereafter, additional
specified unit and initiates format control.
Format
records can be read as indicated by the format specification.
control requires that a new record be input when a slash occurs in the
the
1last closing parenthesis of
format specification, or when the
and I/0 list elements remain to be
reached
is
format specification
Any remaining characters in the current record are discarded
filled.
when the new record

is read.

to the
On execution, a formatted output statement transmits a record
Records can also be
specified unit as format control terminates.
if a slash appears in the format
written during format control
if the last closing parenthesis is reached and more
specification or
I/0 list elements remain to be transferred.

The I, O, 2, F, E, D, G, L, A, and Q field descriptors each correspond
No list element corresponds to an H,
one element in the I/O list.
to
character
In H and
X, P, T, or character constant field descriptor.
field descriptors, data transfer occurs directly between the
constant
external record and the format specification.

When format control encounters an I, O,

%2, F, E, b,

field descriptor, it determines whether a corresponding element exists
transfers data
format control
If one does,
I/0 list.
in the
as appropriate, to or from external format) between the
(translated,
then proceeds to the next £field
record and the 1list element,
If no corresponding
descriptor (unless the current one is to repeat).
list element remains,

format control terminates.

is
format specification
When the last closing parenthesis of the
format control determines whether more I/O list elements are
reached,
1if
However,
terminates.
format control
If not,
to be processed.
format
the
all of
part or
remain,
elements
list
additional
specification is reused in a process called format reversion.
is
In format reversion, the current record is terminated, a new one
repeat
group
the
to
reverts
control
format
and
initiated,
specification whose opening parenthesis matches the next-to-last
If the format does
closing parenthesis of the format specification.
to
returns
not contain a group repeat specification, format control
Format
the initial opening parenthesis of the format specification.
control continues from that point.

FORMAT STATEMENTS

8.7

SUMMARY OF RULES FOR FORMAT STATEMENTS

The following sections summarize the rules for constructing and
using
the
format
specifications and their components, and for constructing
external fields and records.
Table 8-4 summarizes the FORMAT codes.

8.7.1

General

Rules

1‘

In a field descriptor
n
must
be
unsigned

FORMAT

statement must

(They cannot

statements.)

such

can

omit

labeled.

rIw or

integer
assigned

names

You

always

nX,

the

terms

constants
greater
to
constants
in

the

repeat

count

and

than zero.
PARAMETER
field

width

specification.
In a field descriptor such as
unsigned
1integer
constant.

Fw.d, the term
d
must
be
an
You must specify 4 in F, E, D,
and G field descriptors even if it is zero;
and
the
field
width
specification
(w) must be greater than or equal to d.
The decimal point is also required.
You must either
specify
both w and d, or omit them both.
In

field descriptor

characters

must

printing ASCII
In

scale

such

character

factor

the

as
the

this

nHclc2

...

cn,

format

code.

You

field descriptor.

form nP,

must

exactly
can

use

any

signed

unsigned
integer
constant
in
the
range
-127 through 127
inclusive.
The scale factor affects the F, E, D, and G field
descriptors
only.
Once
you
specify
a
scale
factor, it
applies to all subsequent
real
or
double
precision
field
descriptors
in that format specification until another scale
factor appears.
You must explicitly specify OP to
reinstate
a scale factor of zero.
Format reversion does not affect the
scale factor.
No repeat count is permitted in H, X, T or character constant
field
descriptors
unless
these descriptors are enclosed in
parentheses and treated as a group repeat specification.
If the
format

associated I/O statement contains
an
I/0
specification
must
contain
at
least

descriptor.
or
A

However,

character

format

the same as a
including the

cannot be

constant.

specification

reference,

this descriptor

1list,
the
one
field

variable,

array element, or an
format specification

opening

and

character

array must

substring

constructed

in
a
FORMAT
closing parentheses.

statement,

8-23

FORMAT STATEMENTS

Input Rules

8.7.2
l.

A minus sign must precede a negative
value
in
an
external
input
field;
a
plus
sign
1is
optional before a positive
value.

On input, an external field under I field descriptor
control
must
be
an
integer
constant.
It cannot contain a decimal
point or an
exponent.
An
external
field
under
O
field

descriptor
control
must contain only the numerals 0 through
7.
An external field input under Z field descriptor
control
must
contain only the numerals 0 through 9 and the letters A
through F.
An external field under O or Z
field
descriptor
control
must
not
contain
a
sign,
a decimal point, or an
exponent.
You cannot use octal and hexadecimal constants
in
the

form

'777'0 or

'AF9'X

external

records.

On input, an external
field
under
F,
E,
D,
or
G
field
descriptor
control
must be an integer constant or a real or
double precision constant.
It can contain
a
decimal
point
and/or an E or D exponent field.
If an external field contains a
decimal
point,
the
actual
size
of
the
fractional
part of the field, as indicated by
that decimal point, overrides
the
d
specification
of
the
corresponding real or double precision field descriptor.
If an external field contains an exponent, the
(if any) of the corresponding field descriptor
for the conversion of that field.
The
field
width
specification
must
accommodate both the numeric character

field and any other characters that

sign,

decimal point,

scale
factor
is inoperative

be
1large
enough
to
string of the external

are

and/or exponent).

allowed

(algebraic

A comma is the only character you
can
use
as
an
external
field
separator.
It
terminates
input
of
fields
(for
noncharacter data types) that are shorter than the number
of
characters
expected.
It also designates null (zero-length)
fields.

8.7.3

Output Rules
1.

A format specification cannot specify more output

characters

printer
record
cannot
contain
more
than
including the carriage control character.

characters

than

the

external

record can contain.

For example, a line
133

The field width specification (w) must
be
large
enough
to
accommodate
all
characters
that
the
data
transfer
can
generate, including an algebraic
sign,
decimal
point,
and
exponent.
For example, the field width specification in an E
field descriptor
should
be
1large
enough
to
contain
d4+7
characters.

The first character of a record output to a line
printer
or
terminal
is
wused
for carriage control;
it is not printed.
The first character of such a record should be a space,
0,1,
$,
or
+.
Any
other
character
is treated as a space and
deleted

from the

record.

8-24

FORMAT

Summary

STATEMENTS

Table 8-4
of FORMAT Codes

Code

Form

Effect

Transfers decimal

Oow

Transfers octal

Transfers hexadecimal

Fw.d

Transfers

Ew.d

Transfers real or
(E exponent field

double
precision
indicator)

values

Dw.d

Transfers

double

values

real

(D exponent
G

Gw.d

Transfers

integer

integer values

integer

values

double precision values

field

real

values

precision

indicator)

double

precision

on
input,
acts
1like
F code;
acts like E code or F code.
L

Transfers

logical data:

characters;

Transfers

on output,

character

input,

transfers

transfers
T

nHc...cC
or -’

Transfers data between an external record
and the H field descriptor, or character

'c...c!

constant

Specifies

that

Tabulation

specifier

Obtains the number of characters
remaining
to be transferred in an input record.

suppresses
interactive

Terminates

values:

on output,

Hollerith values

characters

skipped
(on input) or that n
be transmitted (on output)

carriage
I/0.

format control

exhausted.

are

spaces

return

the

are

be
to

during

I/0

1list

CHAPTER

AUXILIARY INPUT/OUTPUT STATEMENTS

The
auxiliary
input/output
functions.
These statements
OPEN

associates

statements
are:

FORTRAN

logical

perform

wunits

file

with

management

files.

OPEN

establishes
a connection between a logical unit and
device, and declares the
attributes
required
for
write operations.

a file or
read
and

CLOSE -- terminates
a file or device.

unit

REWIND,

BACKSPACE,

the

and

connection between

FIND

perform

logical

file

and

positioning

functions.
ENDFILE -- writes a special form
of
record
end-of-file
condition
(and
END=
transfer)
statement reads the record.
DEFINE FILE -and associates

that
causes
an
when an
input

describes an unformatted,
direct
access
the file with a logical unit number.

file

AUXILIARY INPUT/OUTPUT STATEMENTS

OPEN
9.1

OPEN STATEMENT

An OPEN statement either connects an existing file
or creates a new file and connects it to a logical

OPEN can specify file attributes
subsequent processing.
The

OPEN

statement

has

the

that

control

to a logical
unit,
unit.
1In addition,

file

creation

and/or

form:

OPEN (par | ,par]...)
par

keyword

key
key =

specification

one

the

following

forms:

value

key
A

keyword,

described

the

keyword,

below.

value
Depends
Keywords

are

divided

Keywords

that

into

described

several

identify