Intel fortran-80 User Manual

FORTRAN-SO

PROGRAMMING

MANUAL

Manual

Order

Number: 9800481 A

Copyright

Intel

I

Corporation,

3065 Bowers Avenue, Santa Clara,

Corporation

California

95051

. I

The information in this

document

is

subject to change without notice.

Corporation

Intel to, the implied warranties

makes

no

warranty

of

merchantability

of

assumes no responsibility for any errors commitment to

part

No written consent

The following are

and the combination

Additional copies

Literature Intel

update

of

this

document of

ICE iCS In

site Intel Intellec

Department

Corporation

nor

to keep current the

may be copied

Intel

Corporation.

trademarks

of

Intel

ICE, iCS, iSBC, MCS,

this

manual

Corporation

or

other

3065 Bowers A venue Santa Clara,

CA

95051

any kind with regard to this material, including,

and

that

or

fitness for a particular purpose. Intel

may

appear

information

reproduced in any form

iSBC Library MCS Megachassis Micromap Multibus

or

in this document. Intel

contained in this document.

or

by any means without the prior

and

may be used only

Manager

to

RMX and a numerical suffix.

describe

PROMPT Promware RMX UPI ",Scope

Corporation

Interproducts:

Intel literature may be obtained from:

but

not

limited

Corporation

makes no

ii

PRINTED

IN

U.S.A./A151/1279/5K

CP

This

manual

for

programming

ANSI

FORTRAN FORTRAN sions

of

FORTRAN

describes the Intel-developed

the 8080

77

full language;

and

8085 microcomputers.

subset. In some instances, it

FORTRAN-80

77.

FORTRAN

also has features

language (FORTRAN-80)

FORTRAN-80

incorporates

is

based

features from the

that

exceed

both

PREFACE

on

the

ver-

Appendix found features features are clearly not tability

The or

extensions related to a for operator's system environments (such as ISIS-II

The experienced reviewing the summaries in the appendixes should

programming

a Chapter

Chapter statements in detail. These statements lustrative examples. programming ing

as TRAN bibliography

D lists

in the

that

be used

among FORTRAN-80

that

compiler. They are summarized in Appendix F

read it

1 has a

2 introduces

a science. We also suggest

tutorials. Some recommended

FORTRAN-80

FORTRAN

go

beyond

if

you

processors.

manual

through

reference, it does contain some instructional material. short

techniques

at

the end

both

marked

want

to preserve total compatibility with

language

particular

also includes considerations when running in different

FORTRAN

from the beginning. While the

program

FORTRAN

Chapter

and

of

extensions to the

77

full set;

versions

in the text

is

described in its entirety in this

programmer

designed

7 suggests some guidelines for improving recommends sources for further study

that

Chapter

it

of

FORTRAN

of

this

compiler

and

RMX-80).

to

provide

concepts.

are

grouped

the beginning

introductory

7.

FORTRAN

also lists those

77. In

manual

are

can

of

Chapters

(shadowed in grey)

described in the

possibly begin

this

manual.

manual

an

intuitive feel for the language.

3-6 describe

functionally

programmer

texts

77 subset

FORTRAN-80

addition,

FORTRAN

manual.

operator's

of

this

programming

The

is

primarily written as

and

read

are

that

these superset

and

77

Limitations

manual.

new

programmer

FORTRAN-80

include

FORTRAN

of

programm-

one

included in the

can be

superset

should

and

por-

manual

The

operating

after

brief

two

FOR-

il-

Finally, all users necessary:

ISIS-II ISIS-II X3.9-1977

The latter

The 1430 Broadway New

10018

FORTRAN-80 System

American National

York, New York

User's

FORTRAN

document

of

this

Compiler

Guide

can

manual

Operator's

be

ordered

Standards

should refer to the following

Manual

from:

Institute, Inc.

documents

9800480 9800306

as

iii

PREF'ACE

GLOSSARY

CHAPTER 1 INTRODUCTION TO FORTRAN

1.1

An

Introductory

1.1.1

Comment

1.1.2 Type Statement

1.1.3

Input

1.1.4 Value Assignment Output

1.1.5

L 1.6

Program

1.2

Summary

1.2.1 Executable

1.2.2 Nonexecutable

1.2.3

Order

CHAPTER

2

Example

Lines

Statement

Statements. . . . . . . . . . . . . . . .

Termination.

Of

FORTRAN-SO

Statements.

of

Statements . . . . . . . . . . . . . .

................

....................

...................

..................

. . . . . . . . . . . .

Statements.

. . . . . . . . . . . .

Statements.

. . . . . . . . .

. . .

.. .. .. .. .. ..

FORTRAN CONCEPTS

2.1

FORTRAN

2.1.1

2.1.2

2.1.3 Statements

2.2

FORTRAN

2.2.1

2.2.2

2.2.3

2.2.4 Expressions

2.2.5 Scope Notational

2.3

CHAPTER

Program Program The

PROGRAM

Statement Character Constants Arrays

....................

of

Conventions . . . . . . . . . . . . . . .

Structure

units

and

Procedures

Statement

and

Lines. . . . . . . . . . . . . .

Elements.

Set.

.....................

and

Variables

and

Operators

Symbols

.................

.............

............

3

.........

..........

. . . . . . . . .

, . . . .

..........

.. ..

..

DEFINING VARIABLES, ARRA YS,AND MEMORY

3.1

TypeStatements

3.1.1

REAL

3.1.2

INTEGER

3.1.3

LOGICAL

3.1.4

CHARACTER

3.1.5

IMPLICIT

3.2

Array

Definition.

3.2.1

DIMENSION

3.2.2 Kinds

3.2.3 Properties

3.2.4 Referencing

3.3 Assignment Statements

3.3.1 Arithmetic Assignmnt 3,,3.2 Logical Assignment Statement.

3.3.3

Character

3.3.4

ASSIGN

3.3

.5

DATA

........................

Type

Statement

Type Statement . . . . . . . . .

Type Statement

Type

Statement

. . . . . . . . . . . . . . . . . . . . .

Statement

of

Array

Declarators

of

Arrays

Array

Assignment

Statement.

........

Statement.

..............

............

................

Elements . . . . . . . .

..................

Statement.

. . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

,.....

..... , .....

. . . . .

..........

......

........

. . . .

..

.. ..

..

,

1-1 1-1

,

1-2

,

1-2 1-2 1-2 1-3 1-3 1-3 1-3 1-4

2-1 2-1

, 2-2

2-2 2-3

'2-3

, 2-4

2-6

, 2-7

2-12 2-13

3-1 3-1 3-1 3-2 3-3

, 3-3

3-4 , 3-5 , 3-5

3-6

3-6 , 3-7

3-S

3-9

3-10

CONTENTS I

3.4

Memory

3.4.1

3.4.2

3.4.3

3.4.4

CHAPTER

PROGRAM EXECUTION CONTROLS

4.1

Transferring

4.1.1 Unconditional

4.1.2

4.1.3 Assigned

4.1.4 Arithmetic

4.1.5 Logical

4.1.6

4.1.7 Block

4.1.S ELSE

4.1.9

4.1.10

4.2

Loop

4.2.1

4.2.2

4.2.3 Program

4.3

4.3.1

4.3.2

4.3.3

CHAPTERS FUNCTIONS

5.1 Intrinsic

5.1.1 Intrinsic Functions

5.1.2

5.1.3 Statement Functions

5.2 External Procedures

5.2.1

5.2.2 Subroutines

5.2.3

5.2.4 5 .2.5 SAVE

5.2.6

5.2.7

5.3 Arguments

5.3.1

5.3.2

5.3.3 Association

CHAPTER

INPUT/OUTPUT

6.1 Records, Files,

6.1.1 Record Properties

6.1.2 File Properties

6.1.3

Definition. EQUIVALENCE COMMON BLOCK BLOCK

. . . . . . . . . . . . . . . . . .

Statement

Statement DATA DATA

..............

Subprograms Statement

4

Program

Computer

IF

IF,

ELSE

IF IF

ELSE

Statement . . . . . . . . . . . . . . . . .

END

IF

Control

Operation DO CONTINUE

PAUSE STOP END

Statements . . . . . . . . . . . . . . .

Statement.

Termination

Statement.

Statement . . . . . . . . . . . . . . . . .

Statement

AND

And

Statement Functions . . . . . . . .

INTRINSIC

FUNCTION

SUBROUTINE

RETURN

Statement. EXTERNAL CALL

Statement.

And Common Dummy

and

Control

GO

TO

Statement

GO

TO

Statement.

GO

TO

Statement.

IF

Statement . . . . . . . . . . .

Statement.

IF,

and Statement Statement

Statement.

of a DO

Statement.

. . . . . . . . . . . . .

ELSE

.................

................

. . . . . . . . . . . . . .

Loop.

. . . . . . . . . . . . . . . . . .

Statements.

................

...................

SUBROUTINES

.................

Statement

...............

....................

Statement

......................

Statement.

Statement

Common

Blocks

Actual

of

...............

. . . . . . . . . . . . . . . .

Statement.

. . . . . . . . . . . . . . . .

Blocks Revisited

...................

Arguments.

Arguments

6

And

Units

.................

..................

....................

Unit

Properties.

. . . . . . . . . . . . . . . . .

.........

........

...........

.............

......

. . . . . . .

. . . . . . . .

Blocks

.......

. . . . . . . . . .

. . . . . . . . . . .

. . . . . . . .

.............

..............

...........

.............

. . . . .

............

..

3-11 3-11 3-12 3-13 3-14

4-1 4-1

..

4-1

..

4-2

..

4-2

..

4-3

, 4-3

4-4

, 4-4

..

4-5

..

4-5

..

4-6

..

4-6

..

4-6

..

4-7

..

4-7

,

4-S

..

4-S

,

4-S

..

5-1

5-1 , 5-2 , 5-2 , 5-4

5-4 , 5-5

5-6 , 5-6

..

5-7

..

5-S

... , 5-S

5-9

..

5-9

6-1

6-1 ,

6-1

..

6-3

v

6.2 File-Handling

6.2.1

OPEN CLOSE

6.2.2

6.2.3

BACKSPACE

6.2.4

REWIND

6.2.S

ENDFILE

6.3

Data-Transfer

6.3.1

READ

6.3.2

WRITEStatement PRINT

6.3.3 Formatted

6.4

6.4.1

Unformatted

6.4.2

Formatted

6.4.3

FORMAT

6.4.4 List-Directed

CHAPTER

Statement.

And

7

Statements.

Statement.

110

Statement

Unformatted

PROGRAMMING

7.1

Program

7.1.1

7.1.2

7.1.3 Refining the

7.1.4 Final Coding

7.2

FORTRAN

7.2.1 Functions

7.2.2

7.2.3 Crossing

7.2.4

7.2.S

7.3

References....

APPENDIX

FORTRAN-80

A.I

Statement Sequence

A.2 Statement

Development Problem Program

GO

Computing

Reminders........................

Coding.

TO

Statement

Definition.

Unit Lines

A

STATEMENT

Summary.

APPENDIXB INTRINSIC

B.l

Intrinsic Function

B.2 Notes On Intrinsic

FUNCTIONS

. . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

...............

Statement. . . . . . . . . . .

Statement

Statement.

Statements

Data

Statement

Formatting

.................

Data

Transfer.

...............

. . . . . . . . . . . .

.............

Data

Transfer..

.........

..........

...............

...........

GUIDELINES

.................

. . . . . . . . . . . . . .

Documentation.

Problem

....................

. . . . . . . . . . . . . . . . . . .

and

Subroutines . . . . . . . . .

..................

Variables

. . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . .

Definition.

...............

and

Constants. . ..

. . . .

SUMMARY

....................

. . . . . . . . . . . . . . . . . .

Summary.

Functions.

. . . . . . . . . . . .

. . . . . . . . . . .

.. .. " 6-8 ..

6-10

..

6-10 6-10 6-10 6-13 6-14 6-14 6-14

, 6-1S

6-16

, 6-22

" .. .. .. " .. ..

" 7-4

..

, A-I

..

.. ..

6-4 6-4

6-9

7-1 7-1 7-1 7-2 7-3 7-4 7-4 7-4

7-4

7-S 7-S

A-I

B-1 B-3

APPENDIX HOLLERITH

C.l

Hollerith As A

C.2

The Hollerith C.2.1 Hollerith

C.2.2

C.3 Hollerith C.4

'A'

APPENDIX EXTENSIONS

D.l

Standard

D.2

Nonstandard

D.3 More Specific Semantics

D.4 Differences

APPENDIX ASCII

CODES

APPENDIX 8080/8085

F.l

Processor Limitations

F.2

Compiler Extensions F .2.1 Lowercase Letters

F .2.2 F .2.3 Reentrant Procedures F .2.4 Free-form Line F.2.S F .2.6 Default F .2.7 Including Source Files F .2.8

F.2.9 Flexibility In

F.3

Unit Preconnection

C

DATA

Statements

Hollerith

Statements

Format

Editing

Of

TYPE

Data

Type. . . . . . . . . . . . . . .

Constant

Constants

......................

Constants

........................

Specification. . . . . . . . . . .

Hollerith

D

TO

ANSI

Extensions

Subset.

FORTRAN

FORTRAN.

. . . . . . . . . . . . . . . . . . . . . . . . .

From

To

Extensions

......................

. . . . . . . . . . . . . . . . . . . . .

1966

. . . . . . . . . . . . . . . . . . . . .

E

F

PROCESSOR

...................

Port

Input/Output.

Format.

Interpretation

RECL

Files. . . . . . . . . . . . . . . . . . . . . . . . . . . .

F.2.9.1 Association

Locations. F.2.9.2 Partially Initialized Arrays F.2.9.3 Transfers

of

Data

Lengths.

Specification

Standard

. . . . . . . . . . . . . . . . . . .

....................

INDEX

..................

In

DATA

In

CALL

Data.

. . . . . . . . . . .

FORTRAN

1977

To

1977

Than

1977

DEPENDENCIES

On

Language

................

. . . . . . . . . . . . . .

DO Statements

For

of

Memory

Into

An

.......

.............

. . . . . . . . . . .

.....

. . . . . . . . . . . .

.............

Sequential

Restrictions

....

IF Block

....

...

..

C-I C-l

"

C-l

..

C-2.

..

C-2

..

D-l

" D-l

..

0-2

..

0-2

" F-l "

F-l " F-2 ..

F-2 " F-2 ..

F-2 " F-3 ..

F-3 " F-3

..

F-4 " F-4

..

F-4 " F-4 " F-4 " F-4

1-1 1-2 2-1 2-2 Type, Length,

vi

Batting Average

Order

of

FORTRAN

Program

Units

(OPI +OP2)

Program

.......................

and

......................

...............

Statements

Interpretation

.........

of

1-1 1-4

2-1

2-8

ILLUSTRATIONS I

2-3 Type, Length,

(OPI **OP2) 2-4 3-1 3-2

Length Subscript Result

of(OPI . OR.

of

Value.

'v = e'

and

Interpretation

....................

OP2)

. . . . . . . . . . . . . . . . . . .

......................

of

.............

" 2-8

2-11

..

3-7 3-8

GLOSSARY

Argument(s) - A collection

formed. are

Array

selectively (by

Array

Association -

symbols.

tuals

of

are

Common

units.

Compiler -

executable

Equivalenced Expression - A

arithmetic,

File - A collection

ternal

quentially

Functions

replaced

with actual values

-

An

ordered

array

Element -

May

Association

when a procedure

memory

also said

by

to

two

be

Memory

The

software

form.

Memory

character,

(a

character

or

directly.

and

subroutines

set

of

element

An

individual

refer

to

of

arguments

is

referenced.

or

more

associated.

-

Memory

tool

-

Memory

combination

relational,

of

data

variable

of

values

when

data

that

name).

item within

association

items.

shared

for

of

logical,

records.

or

character

and

are

usually

the

functions

can

be

of

is

the

replacement

Association

The

symbol

by items in

translating

shared

by items in

operands,

or

May

be array

variables

defined

or

referenced

an

array.

arguments,

of

names

FORTRAN

operators,

Boolean.

external

element).

on

which a with

subroutines

collectively (by

of

dummy

memory

of

the

same

the

same

(any

Records

computation

dummy

memory

arguments

locations

the

items

or

source

and

ISIS-recognized file)

code

program

parentheses.

can

arguments

are

referenced.

array

locations,

is

the

sharing

different

into

units.

be accessed se-

is

per-

that

name)

or

of

with ac-

sharing

memory

program

machine-

May

be

or

in-

Function - A

Functions

ed. defined, program).

Length

two,

or

Character

routine

single-statement

of

Data -

four

bytes

data

Main Program -

gram

to

be

invoked. DATA first

Number

decimal,

Procedure -

SUBROUTINE

Program -

programs trollines,

statement

statement.

Base -

or

hexadecimal.

The

(FUNCTION,

library

Another

that

are

called

The

number

for

integer

occupies

The

subprograms

entire

procedures,

one

The

main

It

may

as its first

representation

term

executable

SUBROUTINE,

returns

a value

'intrinsic'

function),

of

bytes

and

byte

per

program

not

have a FUNCTION,

statement.

for a function

are

called

program

and

included

to

the

(FORTRAN

or

'external'

occupied

logical items

character.

is

the

first

It

may

used

for

numeric

'external'

includes BLOCK files.

calling

have a PROGRAM

statement

predefined),

(user-defined

by a

data

and

four

part

of

an

SUBROUTINE,

data.

or

subroutine.

procedures.

the

DATA)

when

item.

bytes

entire

May

main

plus

it

is

referenc'statement' FUNCTION

This

can

for

real items.

FORTRAN

or

statement

be

binary,

FUNCTION

program, any

compiler

(user-

sub-

be

one,

pro-

BLOCK

as its

octal,

and

all sub-

con-

vii

Glossary

FORTRAN-SO

Program Unit -

unit

must

Record

- A

Statement - A

pressions,

Statement

be

given a

Another

be

terminated

sequence

etc. A

statement

Label-

symbolic

of

A 1-5 digit

name

Subprogram - A tJlock

DA

T A

statement

Subroutine - A

SUBROUTINE the

last.

Unit - A logical

as its first

group

statement

way

of

name

by

an

values

of

syntactic

has

by

of

code

of

statements

mllst

referring

for a main

END

or

characters.

an

'initial'

integer

the

ASSIGN

having a FUNCTION,

statement

be

to

program

statement.

items:

statement

line

in

columns

statement.

and

the

for

performing a frequently-used

the

first

a file.

Once

or a subprogram.

label,

and

up

1-5

of a statement's

END

statement

statement;

connected,

keyword,

to

nine

'continuation'

SUBROUTINE,

as

its

the

END

statement

it

is

the

Every

arguments,

initial line.

last

statement.

operation.

same

as a file.

program

lines.

or

BLOCK

must

ex-

Can

The

be

viii

This

chapter TRAN also includes a quence.

1.1

An Introductory Example

A

FORTRAN processing the drawn statements

The

example calculates a baseball

opens

with a

a feel

for

the language.

summary

program

data

received,

from the statistical world

for doing these

short

The

of

FORTRAN-80

generally

and

operations.

CHAPTER

11

INTRODUCTION TO FORTRAN

example intended

example

performs

returning

of

the

player's

is statements

three basic

output.

sports

batting

to

give the newcomer

discussed in some d.etail.

and

their

proper

operations:

The

following

fan(atic), shows typical

average using the

receiving

short

equation:

to

FOR-

The

chapter

coding se-

input,

program,

FORTRAN

AVERAGE =

The

baseball statistician, sitting

how

often the is

done

he has

player's

only once.

name

batted, and

and

batting

TIMES AT BAT

at

his console terminal, enters the

his

total

average.

1.1.1 Comment Lines

The

first seven lines

document a program.

A

comment

any

characters blank line sent in line 7

line

is

considered a of

C CALCULATE BATTING AVERAGES C VARIABLES C PNAME = PLAYER'S C AB = TIMES AT BAT CHITS C AVG

C

CHARACTER*12 PNAME

10

20

of

the

example

must

have the letter

accepted by

the example.

FORTRAN

comment

Comment

USED-

= TOTAL BASE HITS

= BATTING AVERAGE

READ FORMAT AVG PRINT 20, FORMAT (A, 5X, F4.3) END

10, PNAME, AB, HITS

(A,

2(2X, F3.0))

= HITS/AB

PNAME, AVG

HITS

hits. To

are

comment

.

'C'

or

in the

line also.

lines have

NAME

name

of

a player,

The

program

keep this example simple, the calculation

lines.

an

asterisk (*) in

remainder

For

example, the

no

Comment

effect

returns

of

on

a listing showing

lines

are

used

column

the line. A completely

program

'C'

need

1 followed by

not

execution.

to

be pre-

Fig. 1-1 Batting Average

Program

1-1

Introduction

To

FORTRAN

1.1.2 Type Statement

FORTRAN-SO

Every variable used in a character, PNAME

No type statement FORTRAN be used

or

Hollerith. The

represents character

variable-naming convention tells us implicitly

to

name real data. This convention

FORTRAN

CHARACTER data

is

needed for the other variables listed (AB, HITS, AVO). The

program has a type - integer, real, logical,

and

may have up to

is

type statement says

12

characters.

that

described later in section 2.2.2.1.

that

these variables are to

1.1.3 Input Statements

Following the The first tells the program to read input data; the second describes the input

data.

The

number found in statement 10. The rest be specified by the person

CHARACTER

10

in the

READ

statement in the example are two input statements.

statement tells the

of

the statement lists the variables whose values will

at

the console.

By

program

that

the input format

default,thein,pu(is reaclfrOl11tbecori-

soleierlllinaL

The

FORMAT to

expect.

having the length

2X, F3.0, 2X, F3.0

and

refers followed by a 3-digit floating-point (F) number whose decimal digits. Clearly, no

statement (labeled statement

'A'

indicates the first field

of

to

the AB and

batter

PNAME,

the player's name. '2(2X, F3.0)'

HIT

comes to the plate 79.3 times

fields.

of

10)

tells the program what kind

data

is

a string

'2X'

indicates two blanks will be entered

of

alphanumeric characters

is

equivalent to

portion

or

has 22.8 hits.

the variable

format

of

contains

data

the

is

'0'

One might ask that

is,

2(2X,13}

The

reason by another. Since AB batting

1.000.

is

at

that

this point why

FORTRAN

~

truncates the remainder when one integer

HITS,

we

didn't

specify these fields

all averages would be

'0'

except for the rare player

1.1.4 Value Assignment Statement

The actual batting average calculation

AVG = HITS/AB

This

is

one form

ed the value

of

assignment statement, in which the

of

the expression

'HITS/

is

done by the next statement:

variable'

AB.'

1.1.5 Output Statements

Following the calculation are two

output

In the

~t~teI11e~t?O:'.flN~¥.E'~nd'AV9'

Qefa;tllt~.theoutpqtis~xi~t~n.

data; the next describes the

PRINT

statement, the

output

format

number

~qtheconsoJe.

lines. The first tells the

of

the

output.

20 indicates

are the items whose values are

that

the

output

to

be 'integer' data;

is

divided

AVO'

is

assign-

program

format

to

is

found in

be printed.

write

lly

1-2

FORTRAN-SO

Introduction

To

FORTRAN

Statement of

characters will be followed by five 'A

VG' decimal

20,

the

FORMAT

of

variable

field consists

digits.

statement,

length, as in

blanks

(5X)

of

four

floating-point

1.1.6 Program Termination

The

final

statement

the

FORTRAN

unit

must

be

terminated

1.2 Summary Of

The

statements classifications. and

some

from statements chapter

Statements do

calculations, ecutable units.

are

references

are

statements

terminates

compiler

by

that

an

the

it

hasTeached

END

FORTRAN·SO

available

These

the

listed in

are

classified as executable

read

in

FORTRAN-80

include all

FORTRAN

the

same

to

this

or

write

define

the

statements

sequence

manual,

110

characteristics

indicates

the

FORMAT

and

then

program.

statement.

are

77

full language.

as in Section 7

however,

or

data,

and

the

'PNAME'

statement

the

digits --

The

the

end

batting

average will be

one

END

statement

of

the

Statements

listed below

available in

not

nonexecutable.

control

or

value

to

To

the

simplify of ANSI

program

of

field will be a

labeled 10.

integer

program.

according

FORTRAN

the

ANSI

standard.

Executable

execution.

data

and

The

printed.

digit

and

is

an

indicator

Every

program

to

their

77 subset

comparison,

standard.

statements

Nonex-

define

program

string

name

The

three

to

main

the

All

1.2.1 Executable Statements

1.

Arithmetic, (Chapter

2.

Unconditional,

3.

Arithmetic

4. Block

5.

CONTINUE

6.

STOP

7.

DO

statement

8.

READ,

9.

REWIND,

(Chapter

10.

CALL

11.

END

statement

logical,

3);

and

IF,

ELSE

statement

and

PAUSE

WRITE,

BACKSPACE,

6);

and

RETURN

and

assigned,

logical

IF,

ELSE,

statements

(Chapter

and

(Chapter

character

and

computed

IF

statements

and

(Chapter

4);

PRINT

statements

ENDFILE,

4).

assignment

(Chapter

END

4);

(Chapter

IF

1.2.2 Nonexecutable Statements

1.

PROGRAM SUBROUTINE

2.

DIMENSION, EXTERNAL,

3.

INTEGER,

4.

DATA

5.

FORMAT

6.

Statement

(Chapter

INTRINSIC,

REAL,

statement

function

2),

(Chapter

COMMON,

LOGICAL,

(Chapter

statement

BLOCK

5)

statements;

EQUIVALENCE,

and

SAVE

CHARACTER

3);

6);

(Chapter

DATA

GO

TO

4);

statements

4);

(Chapter

OPEN,

5);

(Chapter

5).

statements;

statements

(Chapter

6); and

3),

IMPLICIT

5)

statements;

type

statements

ASSIGN

(Chapter

4);

CLOSE

FUNCTION,

(Chapter

statement

4);

statements

and

3),

3);

1-3

Introduction

To

FORTRAN

1.2.3 Order of Statements

The

following

1.

Comment after

The

2. program. as the first statement

FORMAT

3.

IMPLICIT

4.

5. All specification statements (lists 1

DATA

must

6.

The

rules for ordering

order

must

be observed in coding

lines can

the

END

PROGRAM

FUNCTION,

appear

statement.

statement can

before

SUBROUTINE,

of a subprogram

statements

can

appear

statements must precede all

statements, which

must

precede all executable statements.

last line

of a program

FORTRAN

unit must be the

statements are summarized in Figure 1-2.

FORTRAN

or

between statements. They

appear

only as the first statement

and

BLOCK

statements lines:

cannot

of

DATA

can

appear

a main

only

(Section 2.1.1).

anywhere before the

other

specification statements.

and

2 in section 1.2.2) must precede all

END

statement.

precede all statement function statements, which

END

statement.

FORTRAN-SO

PROGRAM,

Comment

FORMAT

Lines Statements

Figure 1-2.

FUNCTION,

END

Statement

Order

of

FORTRAN

SUBROUTINE,

or

Statement

IMPLICIT

Other

Statements

Specification

Statements

DA

T A Statements

Statement

Function

Statements

Executable Statements

Statements

BLOCK

DATA

1-4

CHAPTER 2

The

chapter discusses the concepts

and

elements

2.1

FORTRAN Program Structure

of a FORTRAN

and

program.

terminology used to describe the structure

2.1.1 Program U nits and Procedures

The

scope

of

many

FORTRAN

gram

unit

is

either a main program

A main program can start with a programs start with either a ment. A have any

Subroutines tions are further defined to be created outside a call library. Procedures are discussed in detail in

FORTRAN

number

and

an

external procedure written in

program

of

subprograms

functions are called procedures. Subroutines

FORTRAN

operations

or

PROGRAM

FUNCTION,

must have one

(Figure 2-1).

external procedures. External procedures can be

program also; for example, a

is

defined to be a program unit. A pro-

a subprogram.

SUBROUTINE,

and

PL/M-80

Chapter

FORTRAN

statement, though it need not. Sub-

or

BLOCK

only one main program

and

FORTRAN

and

stored in

5.

an

CONCEPTS

DATA

'external'

program can

ISIS-II system

and

state-

may

func-

MAIN

PROGRAM

[PROGRAM]

•

END

SUBPROGRAMS

SUBROUTINE

•

FEN:

FUNCTION

•

Fig.

..

END

BLOCK

~

END

f'

'-----"

MISC. EXTERNAL PRO-

CEDURES

~

2-1

Program Units

•

DATA

•

• • •

-

~

..

1-

• • •

•

J

P-'S~UB=R=OU=T=IN=E

END

J

FUNCTION

END

J

BLOCK

END

,...

DATA

2-1

FORTRAN

Concepts

2.1.2 The PROG RAM Statement

The

PROGRAM

but

when present must be the first statement

PROGRAM name

where

'name'

is

allowed per program.

FUNCTION,

statement

is

the symbolic name

SUBROUTINE,

is

The

used

to

of

main program can contain any other statement except

BLOCK

name a program. This statement

of

a main program.

the program. Only one

DATA,

SAVE,

or

It

has the

PROGRAM

RETURN.

is

optional,

format

statement

FORTRAN-SO

The scope moment, suffice it program. DATA

of

symbolic names

It

cannot

subprogram,

is

discussed later in this chapter (section 2.5.5).

to

say

that

the program name be the same as the name common

block,

or

'local' symbol name in the main program.

is

of

'global'

an

to the entire executable

external procedure, BLOCK

For

the

2.1.3 Statements and Lines

A FORTRAN-SO source program

statements.

Compiler controls direct the operation it

what

kind

of

output ed in detail in the compiler manual where relevant. In general, controls can be embedded in code

and

are identified by a

All

but

two types

by

that

keyword. the keyword do

not

example

'PROGRAM.'

begin with keywords. The

of

an arithmetic assignment statement.

to

of

FORTRAN

For

example, the

produce, the form

is

made up

operator's

'$'

in the first character position

Only 'assignment'

manual,

statement begin with a keyword

PROGRAM

'AVG = HITS/AB'

of

a particular

of

list

compiler controls

output,

and

a few will be mentioned in this

statement

and

'statement

and

FORTRAN

etc. Controls are discuss-

('column'

just

statement in Figure

compiler, telling

FORTRAN

1).

and

are identified

described begins with

function' statements

source

1-1

is

an

2.1.3.1 Statement Labels. Any statement can be labeled; any statement to be

referenced from elsewhere in the unsigned, nonzero, integer ment's

gram

initial line. No two statements may have the same label within the same pro-

unit.

constant

program

must be labeled. The label

written anywhere in columns

is a 1-5

1-5

of

digit,

a state-

2-2

2.1.3.2 Line

first line

ment are called

FORTRAN

A (referred to as statement begins in column 7. A statement can extend over nine continuation lines (columns 7-72) for a total

An

integer anywhere in columns tains a is

Specific Intel mat.

'0,'

a continuation line. Columns

See Section

Format.

of

a statement

continuation lines.

line

can

'columns'

the line

is

FORTRAN

F.2.4

A

FORTRAN

is

called the initial line; subsequent lines in the same state-

have up to 72 characters. The first six character positions

1-6) contain

of

660 characters.

an

initial line;

1-5

compilers may allow exceptions to the

of

Appendix F for details.

statement consists

information

1-5

is

a statement label.

if

column 6 contains any

of

a continuation line must be blank.

characterizing the line.

of

If

column 6

one

or

more lines. The

is

blank

other

character, the line

standard

The

line for-

actual

or

con-

FORTRAN-80

2.2 FORTRAN Statement Elements

A

FORTRAN

• Statement identifier (keyword), such as

• Function identifiers, such as SQRT(A)

• Constants, such as 3.142857

• Variables, such as A

• Operators, such as *

•

Combinations mathematical expressions, such as A

Statement variables, and

in

Chapter

2.2.1 Character Set

statement can include the following elements:

PROGRAM

or

FLOA

or

'STRING'

or

AB

or

.AND.

of

the above into value assignments, such as X =

*B

+ SQRT(C)

and

function identifiers are the subjects

operators,

3.

and

expressions are described in the remainder

T(I)

or

of

Chapters

INTEGER

3-6.

y*Z,

Constants,

of

this

FORTRAN

or

into

chapter

Concepts

The FORTRAN-80 character set consists 0-9,

and

the special characters listed below.

pr~5~ss~~1~9ItliJ~~t~~p:r~rl~iqg·

l~t~ng.sequ¢nce()fthe¢h~(actersisthat

SPECIAL

= Equal Sign

CHRACTERS

Blank

•• ~ ..

~~

of

the alphabetic characters A-Z, the digits

The

set

ofcnar&ctetsreptesentable

••

~I.gr~l?hi¢s~A9th~blan~cryar~cter

of~he

·AS~TTqhaT(\cter$et(Appendix·

'.

+ Plus

Minus * Asterisk / Slash ( Left Parenthesis ) Right Parenthesis

Comma

Period

Single

Quote

$ Dollar Sign

#Pol.1ndSign

Generally, blanks have no meaning in a improve program readability.

For

FORTRAN

example,

statement

and

should be used

in

The·.

E).

the

col*

to

A =

8*C

+ (D* *2/E)

and

A

are equivalent statements. Blanks are counted in the total characters allowed in a

however. They are also significant in character strings dard

line format. They

8*C

+ (D* *2/E)

and

are

not counted in the memory space occupied by a

FORTRAN

in column 6

statement,

of

the stan-

program.

2-3

:FORTRAN

Concepts

2.2.2 Constants and Variables

FORTRAN-SO

The value next. gram execution

The

of

a constant does not change from one execution

value

of

a variable, on the other hand,

or

between runnings

of

the program.

is

subject to change during pro-

For

of a program

example, in the statement

C=A**2+B

the

'2'

is

constant, whereas A,

an

earlier calculation

or

A constant appears as its actual value. A variable has a

B,

and

C are variable

value assignment.

and

may change as the result

symbolic name

that

1-6 alphanumeric characters. The first character must be alphabetic. Thus, all

following are valid variable names:

K

XYZ

B52

ERROR8

STEP3

Every constant Arithmetic constants

2.2.2.1 Data Types. Arithmetic constants and variables are

real (sometimes called 'fixed

and

variable has a data type

and

variables also have an associated

point'

and

'floating

and

point').

length associated with it.

number

base.

of

type integer

to the

of

can be

of

the

or

An

integer

or

'-' include a decimal point preceded by ne.ed

The internal representation, the precision,

constant

sign (0, 123, - 34, + 5). A

an

not

have a decimal point (4E3). Again, the exponent may be signed.

4.2E3

4.2E+3

4.2E-3

is

written without a decimal point

'+'

is

assumed if no sign

and

optional sign (5.,

'E'

may follow a real constant

X

103 .or 4200)

(4.2 (same as above) (4.2 x

10-3 or

.0042)

and

can be preceded by a

is

present. Real constants

.5,0.5,

the range

.0005).

and,

in this case, the real constant

of

real values

An

integer exponent

conforms

'+'

to the floating-point conventions established for the particular processor being used. See section

An

integer variable name begins with one

'N.'

are assumed using

Constants Logical

logical

F.l.

of

the alphabetic

Variables beginning with an alphabetic character

to

be type real. This implicit naming convention can be circumvented

Type

statements, however (see section 3.1).

and

data

constant

.TRUE

variables can also be

may have only the values

are:

.

of

type logical

'true'

or

'false.'

characters'!'

other

than

I,

type character.

The possible forms

J,

K,

through

L,

M,

or

of

N

. FALSE.

a

2-4

FORTRAN-80

FORTRAN Concepts

Character character the form the string

data

is

a string

is

valid

and

of

a string

is

represented by a double

of

any characters representable in the processor. The blank

significant in a character constant. A character

of

characters

surrounded

apostrophe.

by apostrophes.

An

apostrophe

A character constant can have

constant

1-255 characters.

'ARITHMETIC 'MURPHY'S

For

the sake

of also supports Hollerith type Hollerith type

2.~~~.2~~t3Lellgth.ATeal

hitS).

Atlintegtrm~~pc¢upY?l1e,>~W9,.()rfout

spe~ifiedw~eT\th,e"ariably,

Jf~ou~rnberpas~jS'specified

deciO}alaT\<lits'l~ngthist~e

base

.,is

speci~ied.

~~n~t~js~~t~rmJned

OVERFLOW

LAW'

compatibility with earlier versions

is

summarized in Appendix C.

type is

s~Ille;asthe

for,an.i~teger,,~onstant

implicitly

ERROR'

data

under the guise

v~lu~

defined

[or

an

integer constant, the ,constant

by

the

processor from the base

of

FORTRAN,

of

arithmetic/logical types. The

alwij.yso¢cupies four

bytes.F?ran

(section 3.1)

integervariahle, the length is

Or

by

jntegervaria,bIedefault length.,

(a$,des~rib:d

j~

the, next ,'.subsection).

FORTRAN-80

hytesof

default (section F

is

assumed to he

and

number

pftlJeirrteger.

within

memory

If

a

number

of

digits

has

(32

.2~6).

the

~~~.4.~~~~~e~.~~~~

th~~(#!il11ql~t),Illp:erb(l$.ei$assll~ed

can;bebinary,()etal~"

.•

,<~()r

.~ritliineti9·.·

or hexadecimal

••

(irlte~~r~r/~~~~)c()nst~~tS'.and

with.thefollQwi'Q8.e~(;eption~',

as

well.

·,luteger

',variables,

••

constants

2-5

FORTRAN

Concepts

The

or

where:

possibleJorms

[s]d"

.

[sl#d

... b

s

is

an

of

aninteger

optional

constant are:

+ or

-.

FORTRAN-SO

sign

d

is

a digitor.one

of

the letters

b designates the number base

lfthe

numher base specjfier

can

contain

specifier

is interpreted can·contain spe¢ifiedas

only

'8'

is

present,

asa

only the digits· 0-7 and is

the

digitsO~9artdjsinterpretedas

binary number.

base, the number

the

character

is

absent or is the letter

leo'

can

interpreted as a hexadecimal number. The

following

are valid

integer

constants:

o

23

+64101

-#14010

#100010108

-#100010108

-#4567Q +#AF2CH

2.2.3 Arrays

Frequently, still be is called

able

an

the

programmer

to

refer

array.

will

individually

want

to

elements

A~F

and

is

one

of

>'D/thecharacler

a

decimal

stringm

ust contain only the digits o and 1

or

'Q'

is specified as

interpreted

as

an

contain the digits 0-9

to

refer

to a group

in

the

group

the letters

number

thebase,

octal number

or

the

of

data

as

necessary.

D,B,O;

string

..

If

Q,

the

or

'd

... '

base

H

and

the

numper

..

If

'H'

is

lettersA~Fand

by

one

name

Such a group

is

and

2-6

An

array name is

DIMENSION

array element is

ray

name

qualified

The

following generate and PNAME(3),

four

batting

averages. AB(3),

PNAME

GEHRIG OTT

RUTH

We

can

refer

to

hits

for

'OTT'

AVG(3).

the

symbolic

statement,

one

table

1

~dimensional

member

by

a subscript

could

Thus

HITS(3),

type

be

arrays.

the

statement,

in

produced

and

AB

49

60

54

any

can

element

be

in

referenced

name

the

group

enclosed

The

array

A VG(3).

HITS

14

21

18

these

arrays

as

HITS(2)

assigned

or

COMMON

of

data.

in

parentheses.

by

rewriting

table

lists

definitions

AVG

.286 .350 .333

by

using a

and

the

array

when

statement

An

array element name is

the

program

three

players,

for

these

four

subscript.

the

batting

average

it

is

defined

(Chapter

in

Figure

times

arrays

For

example,

for

at

bat, could

RUTH

3). an

1 ~ 1

in a

An

ar~

to

hits,

be

the

as

FORTRAN-80

2.2.4 Expressions and Operators

An

expression

parentheses appear cessing statements (e.g., IF (CALLSUB(X+l,

in assignment statements (e.g., A = B + C) as controls in certain

is

a combination

and

may also include functions (discussed in

Y).

of

numbers, symbols,

FLAG

.NE.

3

GO

TO

and

250),

operators.

Chapter

and

It

5). Expressions

in subroutine calls

may

data

FORTRAN

include

pro-

Concepts

FORTRAN

2.2.4.1 Character Expressions. A

character

reference. The expression

2.2.4.2 Arithmetic Expressions.

computation.

values representable in the processor (see section

tion produces values outside this range, the results are undefined.

Arithmetic operands must identify values

2.2.4.2.1 Arithmetic

has four kinds

constant,

This

computation

Operator

**

/

*

+

of

character

a character variable reference,

may

be enclosed in parentheses.

An

arithmetic expression performs a numeric

is

limited by the range

of

Operators.

Meaning

Exponentiation Division Multiplication U

nary

Unary

The

arithmetic

or

binary addition

or

binary subtraction

expression consists

F.l).

type integer

operators

or

a character

and

If

any

or

real.

are:

precision

part

of

either a

array

of

numeric

of

the computa-

element

The following expressions calculate the perimeter, square with side

SPERIM = SQAREA = SQDIAG = SQRT (2*(S* *2))

length'S.'

4*S

S**2

area,

and

diagonal length

of

a

2-7

FORTRAN Concepts

FORTRAN-80

Fig.

2w$

·.type,. Length,

andJnterpretatbin

As these figures indicate, mixed-mode arithmetic operands This conversion

In the case

If

the magnitude

the

to the same type (the type

is

unnecessary when a real

of

an

integer divided by

The

value

of

1/3

is

The

value

of

8/3

is

The

value

of

-8/3

of

an

number

of

bytes shown

of

another

0

2

is

-2

arithmetic result

in

these figures, the result

the result) before performing the

number

is

integer, the remainder

is

too

large for the processor

2.2.4.3 Relational Expressions

Relational expressions return a

TRUE

or

2.2.4.3.1 Relational

compare

FALSE

Operators.

two arithmetic

result

of

The relational

or

type logical.

operators

o~(O~1**QP2)

is

done

by converting

raised to

an

integer power.

is

truncated.

to

is

undefined. See section

two

character

expressions

are:

both

operation.

represent in

F.I.

and

2-8

Operator

.LT. .LE.

.EQ.

.NE. .GT. .GE.

Meaning

Less

than

Less

than

Equal

Not

equal Greater Greater

or

than than

equal

or

equal

FORTRAN-SO

f'ORTRAN

Concepts

Relational expressions

IF (NUMB .GT.

IF

(PNAME .EO. 'GEHRIG') PRINT 20, PNAME, AVG

2.2.4.3.2 tional tion

If

the

expressions.

is evaluated as

where operator

2.2.4.3.3 tional expression tion

If

two length character precedes dix E).

Interpretation

expression

set

up

by

the

operands

EXP1

(EXP1

'0'

set

up

character

of

are

The

operator

if

- EXP2)

is the same type as

in

both

Interpretation

by the

the

longer by

expression

the

value

is

TRUE

operator,

of

relational expression

it were written

operator

expressions.

is

TRUE

operator,

operands

of

are

commonly

99)

STOP

of

if

and

different

EXP2

of

if

and

adding

EXP 1 is

EXP2

used in

Arithmetic

the values

is

FALSE

types, type conversion

the

Relational Expressions.

of

the

otherwise.

0

(EXPI -EXP2)

Character

the values

is

have

different

considered

in

the

Relational Expressions. A

of

the

FALSE

blanks

ASCII

otherwise.

lengths, the

on

the right

to

be less

collating sequence,

IF

statement

operands

and

'operator'

operands

than

(Chapter

satisfy the relational condi-

is

similar

satisfy the relational condi-

shorter

of

EXP2

to

is the same relational

is

the

character

if

and

4).

An

arithmetic

that

of

arithmetic

character

'extended'

the value

vice versa (Appen-

string.

of

to

EXP

rela-

the

The

1

2.2.4.4 Logical Expressions. A logical expression

tion

and

returns a TRUE single logical ment

reference, logical

tion

of

2.2.4.4.1

2.2.4.4.2 involving

The

logical expression has

logical

Logical

Operator

.NOT. .AND.

.OR.

.EQV.

.NEQV.

Value

.NOT.

operand

operands

Operators.

and

is

or

FALSE

(logical

function joined

Meaning

Logical Logical Logical inclusive Logical equivalence Logical nonequivalence

Length

as follows:

of

OP1

TRUE

result

constant,

reference,

by logical

The

logical

negation

conjunction

Logical Expressions.

.NOT.OP1

FALSE

FALSE TRUE

the

opposite

of

type logical. This expression can be a

logical variable reference, logical

or

relational expression)

operators

disjunction

value as its

and

are:

operand.

performs

parentheses.

The

value

a logical

or a combina-

of

a logical

computa-

array

ele-

operand

2-9

FORTRAN Concepts

The true. Otherwise, execution stops.

10

following example passes

FLAG = FLAG

+ 1

control

to

line

10

if

the logical variable

DONE

is

not

FORTRAN-SO

DONE = (FLAG IF

(.NOT.

GOTO

DONE)

10

.GT.

99)

THEN

ELSE

STOP

ENDIF

The

value when two logical

OP1

TRUE TRUE

TRUE

FALSE FALSE

If

both

operands

This example achieves the same effect as the last example.

IF

(FLAG

are

.GT. 0 .AND.

operands

OP2

are

combined

OP1

TRUE

FALSE

TRUE

FALSE

true, the logical expression

FLAG

.LE.

FALSE FALSE FALSE

99)

GO

by

.AND.

.AND. OP2

is

true.

TO

10

STOP

The

value when two logical

operands

are

combined

by . OR.

is

as follows:

is

as follows:

2-10

If

either

operand

The

following

IF

The

value when two logical

If

both

statement

(FLAG

operands

.EO.

OP1

TRUE TRUE

FALSE FALSE

is

true,

50

OP1

OP2

TRUE

FALSE

TRUE

FALSE

the logical expression

branches

.OR.

on

FLAG

operands

OP2

either

.LE.

are

10)

combined

OP1.0R.OP2

TRUE TRUE TRUE

FALSE

is

true.

of

two conditions.

GO

TO

OP1

.EQV. OP2

TRUE TRUE TRUE TRUE

FALSE FALSE

are

logically the same, the logical expression

FALSE TRUE FALSE

FALSE FALSE

TRUE

250

by .EQV.

is

as follows:

is

true.

FORTRAN-80

FORTRAN Concepts

The following statement returns whenever the two logical operands are logically equivalent.

IF

(FLAG1

The

value when two logical

.EQV.

FLAG2)

operands

RETURN

are combined by

.NEQV.

is

as follows:

OP1

TRUE TRUE

FALSE FALSE

If

both

operands are logically different, the logical expression

The

following statement continues execution

IF

(FLAG1

.NEQV.

OP2

TRUE

FALSE

TRUE

FALSE FALSE

FLAG2)

OP1

FALSE

TRUE

if

the two operands are

CONTINUE

.NEQV. OP2

is

true.

not

equivalent.

2.2.4.6

Expressions are generally evaluated left are evaluated before When

Precedence

two operators have equal precedence, the leftmost

of

Operators.

other

operators

to

that

right.

Operators

immediately precede

with higher precedence

is

evaluated first.

or

follow them.

2-11

.FORTRAN

Concepts

FORTRAN-SO

Parentheses can be used to override pression enclosed in parentheses nermost are evaluated first.

15/3 + 18/9 = 5 + 2 = 7 15/(3 + 18/9) = 15/(3 + 2) = 15/5 = 3

The

following lists the precedence

• Parenthesized expressions

• Exponentiation: **

• Multiplication/Division: *, /

•

Addition/Subtraction:

• Relational

•

Logical/Boolean

•

Logical/Boolean

• Logical/

•

Logical/Boolean

Thus,

the expression

Operators:

Boolean.

.NOT. .AND.

0 R.

.EQV.,

is

of

+, - (unary

.LT.,.LE.,.EQ.,.NE.,

.NEQV.

o .OR. A + 8 .GE. C

is

interpreted as

though

it

were written

normal

evaluated first.

operators

rules

in descending order:

and

binary)

of

precedence. The

If

parentheses are nested, the in-

.GT.,.GE.

part

of

an

ex-

o .OR. ((A +

One

exception to the left-to-right rule

occur together.

A**8**C

In this case, the exponentiation

A *

*(8

* *C)

8)

.GE. C)

is

the case where two

is

interpreted right-to-Ieft as

2.2.5 Scope of Symbols

A symbolic must be alphabetic. Symbolic names may be the entire local symbol units

The following symbolic names are global to in a local context.

• Main

•

Subroutine

• External function names

• BLOCK

name

consists

program)

name

can, therefore, represent different entities in different

or

in different statement functions.

program

name

names

DATA

of

one

or

they may be local to a

subprogram

names

to

six alphanumeric characters, the first

global

(that

program

the entire

or

more

exponentiations

though

is, they may have a scope

unit

or

program

it were written

of

statement function. A

program

and

cannot

be used

which

of

2-12

The

following symbolic names are local to the

•

Array

names

• Variable names

program

unit in which they appear:

FORTRAN-80

•

Statement

function names

• Intrinsic function names

•

Dummy

Variables that

statement

procedure

appearing

only.

names

as

dummy

arguments

FORTRAN

in a statement function have a scope

Concepts

of

Common block intrinsic function in a function as a function. pearance

block names

name

in a

program

are

generally global,

unit

may

also be the

subprogram.

common

block

If a name

name, is

used for

however,

both a common

identifies only the local entity except in

if

the

2.3 Notational Conventions

This

manual

statements

• Special characters from the uppercase words (keywords) noted.

• Lowercase letters items must be substituted in actual statements.

• Brackets [ ]

• Ellipses (

.times in succession.

• Blanks are used to improve readability,

an

example

As

CALL

uses the following

and

concepts:

and

are

used to indicate

...

) indicate

of

the

notation

sub

[ ([arg[,arg]

notational

FORTRAN

are

to

be

words indicate nonspecific arguments for which specific

optional

that

the preceding

used, the description

...

] ) ]

but

do

admit

exceptions. A

name

of

any local entity

An

intrinsic function

program

unit does not reference

block

and

COMMON

name

a local entity, its ap-

and

SAVE statements.

conventions to describe

character

set, uppercase letters,

common

other

than

an

may

be used

that

FORTRAN-80

and

written as shown, except where otherwise

items.

optional

but

items can

appear

have no significance.

one

or

more

means the following forms

CALL

sub

CALL

sub

( )

CALL

sub

(arg)

CALL

sub

(arg, arg)

CALL

sub

(arg, arg, arg)

etc.

When

for

the actual

'sub'

and

CALL

each

statement

'arg.'

ROUT1(A, 3.72, 4.12)

of

the

CALL

statement

is

written, specific entities would be substituted

are

permissible:

2-13

CHAPTER 3

DEFINING VARIABLES,

ARRAYS,

This chapter describes the statements used to specify the types with symbolic names, how to assign values memory, are:

• Type statements:

•

• Assignment statements: arithmetic, logical,

•

3.1

The type statements REAL, confirm names beginning with the letters I-N refer

tion itself can be confirmed Type statements can also specify

and

how

to

assign values

REAL,

Array

definition statement:

ASSIGN, Memory

DATA

definition statements:

Type Statements

or

override the type implied by a symbolic name (i.e., the convention

to

a block

INTEGER,

DIMENSION

COMMON , EQUIVALENCE,

INTEGER,

or

changed by the

data

length

to

these symbols, .how to structure

of

memory. The statements described

LOGICAL,

to

integers-Section

IMPLICIT

pr

array dimension

CHARACTER,

and

CHARACTER

statement.

AND

and

lengths associated

character assignment,

BLOCK

2.2.2.1). The conven-

information.

MEMORY

IMPLICIT

DATA

are used to

that

Specifying the symbolic

function in a type statement establishes the type the program unit. The type in a

program

statement.

unit.

name

Program

of

a variable, array, external function,

of a name

names

cannot

and

be specified explicitly

subroutine names

3.1.1 REAL Type Statement

The

REAL

type statement has the format:

name

REAL

where

'name'

tion,

or

dummy

Examples:

[,name]

is

the symbolic name

procedure.

REAL TEMP REAL NUMB1, NUMB2, NUMB3

C REAL OVERRIDES

...

of

a real variable, array, array declarator, func-

IMPLICITTYPING

3.1.2 INTEGER Type Statement

of

that

name

for all appearances in

cannot

OF

NUMB=INTEGER

or

more

appear

statement

than once

in a type

The

INTEGER

INTEG

type statement has the format:

ER.fm!@ti£;l)

name

[,name]

...

3-1

Defining Variables, Arrays, And Memory

FORTRAN-80

where 'name'

is

one

of

the forms

v[*/en} ary[(d)][* len]

and

v

ary

ary(

d)

lenj.i~~~~l~9:!~~i*~~f¢~;0;(<trl·.ii)t~g~rV(l~i&l?1~·9t

is

an integer variable, function, or dummy procedure name

is

an array name

is

an array declarator

.·~rp~y¢l~w~#t,

••••.

~t$

• va,luernustuei{A, Qp4. .

•.•.•......•....•...............

:

.•••••....••..••.•.•...•....••....

..•.•..•..•.••

.•......•••....•...•...•..•••••.....••

..•.•

..

•...•

..•

..•.•

••.•.....•..•..

.

••...•.••....•...•..••.....•..•••.••......••..••.•.•.....••....•.•..••.•....••...•...•.

:

...•..•........•..•

'.. . •..••.•.•...

:

•..•...........•..•..•......•..•.•

~

.•. ~ .. : ...•..•.

~.;.::

:

•...•........

:

.•..•...•....•.•...•....•

~

... ! ..•..•.•.•.

::

....•.•• : ...••.••..

:

..•...

:.~

.••...

:::

.... : •...•.

:~

•......•.••.••.

;:.~

•...

:"

..

" . :

.•.

:...............

..••.

......

...•.•.•..•......•.....•..........•........•.

.

.....••.•.•.•..............•.••.......

::

••..•.•......•..........•

:

...•........•...•.....................•.•.•

:. " •.......•.

::......

.'

.. ' ..••.........

:..............

.•.•.

.•...•.....•.•••.••.••..••....•..•..........

....•..•...

. .

..•............•.•.....•.........•.

.•.•...........

... . ..•..•.•

.•... . .....

:.:.

•......

.....•.•..

.......•

..•....

........ • .'. : ..• : ....................•..

:". • .......

":

"'

•....

X~~~~P~f.~·

~.~~.~

•.• · ••.

sp~cifi9~tiqp;.·i~~;:~i~t.7~¥X~11Q~.i119·

:~l'r~~i/t~~

def:~1l111efl.gthqf··(l~··.int~$~rvad~l?1~Ofar.r;:\y

Examples:

•••

~p~~~

••

ite~.·i~···

..

·~J.1estttJe~.el1ttlp~n~yil1~

.•

J~l1~tl)·~~p~j~~

f~E~.~~.~~

•.

··.·.~.~ffi~.~~~~rl~.fP~~~~~.n~··.·~e~~e~~?rf;l.·.·~·fir~~B:.~

.•

~2·itGB'li§!er

•.

·.tQ

•.

·•

~~p~~~rfty.·.~1~mel1t?J:f119.1e~~te:i~~R~fi~i~~~~9T$.~~rent

INTEGER TOTALS

•••

..

i~s

•.

8'Y11\1~~~~~~pe9ifi~~tio~.~

1711gt~f:()~it~~t·i~~rn

..

e1~rqentis.a;~SUlTIeg(s~ySeC!~Pfl.f,4~~).

··~BRl~~~i~?

•...

·.~

.. · ••

·ltl1gt~

.?111~\

••

~A~/.~~

INTEGER*1 DIGITS(10} INTEGER*4

TOTALS, SUBS*2, DIGITS(10}*1

3.1.3 LOGICAL Type Statement

The LOGICAL type statement has the format:

LOGICAL [*len[,]) name [,name]

where 'name' is

one

of

the forms

v[*/en]

a

ry[

(d)][

* len]

and

v is a logical variable, function, or dummy procedure name

ary

ary(d)

leit

".:

,"

.:

...•.

:::",

,"

.. " ..... ; ......

is

an array name

is an array declarator

~~~~~l~~~~~~ij~~~¢sgf#lg$!~#~y~~~~~~$qp~!t~y.~~~m*p.~.lt$¥~h!~

mu&tgf;lJ~,-·911::h

~.::.

'.:~

...

<.,

...

-.:.:'

:".".:

:::

:::,-.::".:,.\:.::.

:.:

..

:.::

.•.. ~ ..

:.:.:.::.:::

..

::

.....•• : ..•.•...... : ..•.......•............•...•...•.....•

'

•..........•...•.

::

.•.•........•..•......•..•.•.......

:

•.•. : .•..............•......................

...

..

:

•..•.•.....•.....................••.••.........•.

::

...•.....................•..•.•..•...•............•..•...............•....•....••......................................•....•.•............

::

............•..............•.

. . .

:

....

;:

......

::

•...•.•..•........

:

...........•..............•....

:

..••....•..

::

..

:::

..... : ..

::

•.... : ......••....•.•.••.....

:.:::.:

..•...

:.: ...

:.

..

>:.:

..

: .

"."

.....•... - '

..

3-2

~~~~<itel'tli~l

s~:vifi~~t19rr.iPlnae~j~t~ly()n9W:il1g~n

ao'a~;

.the,l~ngthi~ppHe$

default.· length . ofalOgical

.••.•

tb!~'~f~~~nt[lQ~.··

..

t().·.eacl1··a,rray.el~~ent

v~riable

•.

~~~i!!$i~s.~~P

.

i~enJ.i~t~el~ng~~fQ~~'a~r~t~tni~I1~>'.~9r~rt

Of

arrayelenlent

..

I~pj~~~~~~~~lR~~~~~;~.~~~w~~

...••

If

~.~.

·~~ll~t~issp~~~f~e~yttl~current

..

is

!l$$91l1eO,f$ee$ectiPfl.ll,2.(j).

FORTRAN-SO

Examples:

LOGICAL *2 FLAG

LOGICAL* 1 FLAGS(10)

Defining Variablest Arrayst

And

Memory

LOGICAL

*4 FLAG1, FLAG2, SWITCH(5)*1

3.1.4 CHARACTER Type

The

CHARACTER

CHARACTER [* len[,]] name [,name]

where

'name'

v[* len] ary[(d)][* len]

and

v

ary

ary(d)

len

type statement has the format:

is

one

of

the forms:

is

a variable name

is

an

array

name

is

an

array

declarator

is

the length

character

(number

array

Statement

of

element.

...

characters)

of

a character variable

or

The length specification immediately following the keyword to

each item in the statement not having its own length specification. A length specification immediately following an item array, the length applies length

of

a character (one byte)

Examples:

CHARACTER*15 CHARACTER*12 NAMES(50), CITIES(50), STATES(50)*5 CHARACTER LETTER

3.1.5 IMPLICIT

An

IMPLICIT trinsic function names) PLICIT

specification in a

statement can also be overridden by statement containing the same symbolic name.

statement defines the type and length for symbolic names (except in-

types can be overridden, however, by type statements

to

each array element.

STRING

Statement

that

begin with the letter(s) specified by

FUNCTION

is

assumed.

statement. The length specified in

an

is

the length for

If

no

INTEGER,

length

LOGICAL,

CHARACTER

that

item only.

is

specified, the

IMPLICIT.

or

by

an

or

CHARACTER

standard

explicit type

IMPLICIT

applies

For

an

IM-

3-3

Defining Variables, Arrays,

The

IMPLICIT

And

Memory

statement has the format:

FORTRAN-SO

IMPLICIT

typ

(let

[,/et] ... ) [,typ

(let

[,let].

..

)].

..

where

typ

let

is

INTEGERf*bml ,

CHARACTER[*

is

a single letter

REAL,

len]

or

LOGICALt~lejz],

a range

of

letters in alphabetical

or

order

(e.g., C, I-M, N-Z)

The

IMPLICIT

must

precede all other specification statements in that program unit. The

unit can have more specified

more

statement applies only to the program unit in which it appears

than

one

than

IMPLICIT

once.

statement,

but

the same letter

cannot

and

program

be

Example:

IMPLICIT REAL(A-B, D-H), CHARACTER IMPLICIT INTEGER (I-N), LOGICAL

(C)

(O-l)

Unless these implicit definitions are overridden, the following symbols would have the types indicated.

AVG CNAME FPNUM INUM PFLAG

(REAL) (CHARACTER) (REAL) (INTEGER) (LOGICAL)

3.2 Array Definition

An

array

is

defined by assigning a symbolic name to the array

dimensions.

CHARACTER

LOGICAL

Arrays can also be defined by the DIMENSION

In any case, a symbol can be used only once in a

array declarator

not

be defined in a statement. The name

One

statement.

elsewhere:

way to

do

this

is

with type statements:

TICTAC(3,3)

TABLE(2,3,3)

COMMON

(section 3.2.1). The symbol

DIMENSION

array

name could,

statement as well as in the

of

course,

'TICTAC'

appear

and

specifying its

statement (section 3.4.2)

program

unit as

an

array name

in the example above could

CHARACTER

as a reference

or

array element

and

by the

in

an

type

3-4

TICTAC(3,2) = 'X'

FORTRAN-80

By

'array

element

name'

we

mean

an

theses as shown in the example above. identifies the entire array, with array

name

not

qualified by a subscript identifies the first element

An

array

name

is

local

to

the

program

one

exception. In

3.2.1 DIMENSION Statement

The

format

of

the

DIMENSION

01

M ENSION ary(d) [,ary(d)].

statement is:

..

Defining Variables, Arrays, And Memory

array

name qualified by a subscript in paren-

An

array

name

not

qualified by a sUbscript

an

EQUIVALENCE

of

the array.

unit in which it

is

declared.

statement, an

where each

ary(d

'ary(d)'

[,d]

....

)

is

an

array

declarator

of

the form

and

ary

d

is

the symbolic

is

a dimension declarator

name

of

the

array

Dimension declarators are discussed in more detail below (section 3.2.3). In general, they indicate the upper

bound)

number

of

each dimension. The maximum

of

dimensions in the

array

number

and

the

number

of

dimensions

of

elements (or

is

seven.

Examples:

LOGICAL INTEGER ARRAY DIMENSION TABLE(2,3), ARRAY(3,3,3)

TABLE

3.2.2 Kinds of Array Declarators

An

array

declarator

justable,

or

In a constant

justable

array

assumed-size

array

('ary(d)'

array

in the

declarator.

declarator, each

declarator contains one

DIMENSION

of

the dimension bounds

or

more variables as bounds:

format)

is

either a constant, ad-

is

a constant.

An

ad-

ARRAY(3,MI00LE,THIRD)

An

assumed-size

bound

An

of

the last dimension

ARRA Y(3,3, *)

array SUBROUTINE and

dummy

declarator, whereas assumed size. Like actual DIMENSION MON

statements. A variable appear program.

in the

The latter requirements can be avoided using the asterisk feature for the

array

declarator

is

an

asterisk.

name

may be used as a

subprogram.

array

declarators.

dummy

or

type statements, but unlike actuals, they

subprogram's

Thus, a program

An

array

declarators,

name

used as a dimension

dummy

last dimension, thereby gaining some

is

either constant

dummy

actual

array

or

adjustable,

argument

in a

can have actual

but

the upper

FUNCTION

array

declarators

declarator must be a constant

declarators may be constant, adjustable,

argument

program

dummy

list

efficiency.

declarators are permitted in

or

bound

in a

cannot

common

appear

of

an

array

block in the sub-

in COM-

must also

or

array

or

3-5

Defining Variables, Arrays,

3.2.3 Properties of Arrays

The

examples following the description

types

of

the array names

rays. They also specify (by default) the

And

Memory

of

the

and,

by implication, the types

lengths

DIMENSION

of

the elements.

statement specify the

of

the elements in the ar-

FORTRAN-SO

The remaining properties

of

the array are determined from the dimension

declarator. These are:

•

The

number

• The size

• The total number The

number

of

dimensions in the array

of

each dimension

of

array elements

of

dimensions equals the

number

of

declarator. Thus, the array

TABLE(4,4)

has two dimensions. The size

is assumed an

The size

of

a dimension

also the same as the

to

have the value one.

is

the same as the value

'upper

assumed-size array declarator.

of

an

array can generally be

dimension

The

upper

computed

'd'

bound.'

bound

dimensions specified by the array declarator. Thus

ARRA

Y(3,3,3)

would have

27

elements. The number

of

elements in

can be determined as follows:

dimension declarators in the array

in the array declarator format; it

The lower dimension

may be

as the

an

product

an

assumed-size

asterisk in the case

of

bound

the sizes

dummy

of

the

array

is

of

•

If

the actual

is

size

•

If

the actual argument

an

array

Array

elements are stored sequentially.

DIMENSION TABLE(3,1) TABLE(2,3)

'2.9'

would be assigned to the

is

'TABLE,'

(1,1

)(2,1

The

total number

multiplied by the

that

of

argument

of

the actual array;

size

'n,'

TABLE(3,3)

corresponding

is

an

array element name with a subscript value

the size

of

the

= 2.9 = 7.3

third

and

'7.3'

would be assigned to the eighth location.

)2.9(1,2)(2,2)(3,2)(1,3)7.3(3,3)

of

bytes in an array

number

of

bytes occupied by each element.

to

dummy

For

example, in the following sequence

storage location in the block whose low address

is

3.2.4 Referencing Array Elements

Array

elements are referenced by qualifying

form

ary

(s [,s]

...

)

the

dummy

array

the

number

an

array name with a subscript in the

is

n + l-p.

array

is

an array name, the

of

elements in the array

of

'p'

in

3-6

FORTRAN-80

where

'ary'

is

equal the

number

Each subscript dimension

bound

subscript must

Examples:

ARRA

Y(2,6) = A

ARRAY(I + ARRAY(M,M

an

array

name

and's'

of

dimensions in the

is

an integer expression in the range 1

of a dummy

not

exceed the size

J,

3)

= 8

is

a subscript. The

array

declarator.

array

is

an

asterisk, the value

of

the corresponding actual array.

+ N,M-N) = A + SQRT(8)

Defining Variables, Arrays,

number

:5

s:5 upper-bound.

of

the corresponding

subscripts must

If

the

upper

And

Memory

Figure elements

(1

:5

n:5 7),

3-1

can

is

'd'

pression.

Dimension

n Declarator

(d1)

1

(d

2

3

n (d1, ... ,dn)

1 ,d2)

(d1,d2,d3)

be used

to

calculate which element in the storage sequence

being referenced. In this figure,

is

the value

of

the upper dimension

Subscript Referenced

(s1

)

(s1,s2) 1 + (81-1)

(s1,s2,s3) 1 + (s2-1)

(s1, ... ,sn) 1

'n'

is

bound,

the

number

and's'

Element

s1

+ (s2-1)*d1

+ (s2-1 + (s3-1) *

+(s1-1) +(s2-1)*d1 + (s3-1)*d1 +

...

+(sn-1)*dn-1

+

dn-2*

of

dimensions

is

the subscript ex-

)*d2

d2

* d 1

*d2

...

d1

array

Fig. 3-1 Subscript Reference

3.3 Assignment Statements

The

type statements correlate a type

statements described in this section assign values elements.

Assignment statements are used variables

The

and

ASSIGN

array

elements. These statements have

statement

is

used

to

variable symbol name.

The

DATA

statement

is

used

to

initialize variables, arrays,

specific values.

and

length with a symbolic name.

assign arithmetic, logical,

assign a numerical statement label

to

variables, arrays,

or

character values

no

keyword.

and

to

array

The

or

array

to

an

integer-

elements to

3-7

Defining Variables,

Arrays,

And

Memory

FORTRAN-80

3.3.1

The formula.

Arithmetic Assignment Statement

arithmetic assignment

Its

v =

format

exp

is:

where

The'

v

exp

='

in

is

the

I

is

an

FORTRAN

name

arithmetic expression

has the sense 'is assigned the value'

Thus

I =

1+1

is

a perfectly correct

of

Execution 'exp'

according

of

'exp'

an

to the type

FORTRAN

arithmetic assignment

to

the rules listed in

of

'v,'

and value, as shown in Figure 3-2. for converting a real

number

tively.

statement

of

a variable

statement.

Chapter

definition

IFIX

and

to

an

integer

closely resembles a conventional arithmetic

or

array

statement

2 (see Figures 2-2

and

assignment

FLOAT

and

an

element

causes evaluation

in this figure are intrinsic functons

integer

of

type integer

rather

and

of

'v'

to

a real

or

real

than

'is equal

of

the expression

to.'

2-3), conversion

with the resulting

number,

respec-

TYPE

OF

'v'

INTEGER REAL INTEGER REAL

Example:

1=3

C = I + SQRT(25.0) C =

AS A RESULT OF THESE

C

If

the value

may

happen

of

tion

If

the length

of

'exp'

when the length

the integer value (see section

of

'v'

preserving its value.

Example:

TYPE

INTEGER REAL REAL INTEGER

f'i~.

C**2

is

too

large to be assigned

of

'v'

is

longer, the length

OF

'exp'

3-2 Result

is

of

CALCULATIONS

too

short

F.l).

of

'exp'

RESULT

exp exp IFIX (exp) FLOAT

'v

= exp'

to

'v'

the result

to

contain the processor representa-

is

converted to the length

(exp)

C = 64.0

is

undefined. This

of

'v'

while

3-8

INTEGER* 1 M(1000) INTEGER*4 N M(200) = N

C LENGTH

= 65

OF

N

IN'

IS CONVERTED TO

ONE

BYTE

FORTRAN-SO

Defining Variables, Arrays,

And

Memory

3.3.2 Logical

The logical assignment statement assigns the

or

v =

v

array

exp

variable

where

exp

Examples:

C FLAG C OTH ERWISE

3.3.3 Character

The character assignment statement assigns a character constant, variable name, array

element

Assignment

element. It has the

is

the

name

is

a logical expression

LOGICAL FLAG, TABLE(3,3) FLAG

TABLE(1,3) TABLE(1,2)

= (INT1 .NE. 1 .AND. INT2 .EQ.

IS

.TRUE.IF BOTH CONDITIONS ARE TRUE

IS

.FALSE.

Assignment

name

to a character variable

Statement

format

of

a logical variable

= .FALSE. = FLAG

Statement

or

value.

or

array

TRUE.

logical

element. Its

or

array

1)

.FALSE.

element

AND

format

to

is:

a logical

or

v =

char

where

v

char

None

of

the character positions being defined in two sides than

'char,'

'char'

is

truncated

Examples:

CHARACTER*10 NAMES(4), MGR MGR

NAMES(1)

NAMES(2) = 'OTT' NAMES(3) = 'RUTH' NAMES(4)

3.3.4 ASSIG N

The

ASSIGN name. The symbolic mat

identifier in context, it must first be redefined as statement.

is

the

name

is a character element name.

of

the assignment

the latter

= 'STENGEL'

is

on

the right until it fits into

= 'GEHRIG'

= MGR

Statement

statement

name

an

input/output

of

a character variable

constant, character variable name,

may

have different lengths, however.

padded

is

on the right with blank characters.

the only way

can then be referenced in a

statement.

or

character

'v'

can

be referenced in

'v.'

to

assign a statement label to a symbolic

GO

To

use the symbolic

an

integer value in

array

element

or

character

If

TO

statement

name

an

arithmetic assignment

'char.'

'v'

is

or

in any

array

The

is

longer

shorter,

as a for-

other

3-9

Defining Variables, Arrays,

The

ASSIGN

And

Memory

statement

has

the

FORTRAN-80

format:

ASSIGN

where

stl

name

The

statement statement must

not

An integer variable defined with a same

or a different

Examples:

stlTO

in the

be

declared as length

name

is a statement

is

an

integer variable

label

must

same

program

statement

label (1-5 digits)

be the label

unit

INTEGER * 1.

label value, as well as with

ASSIGN 1010 TO LOOP1 GOTO

LOOP1

GO TO LOOP1(1000, 1010, 1020) IF(.NOT.DONE)

THEN ASSIGN 20 TO WRFORM ELSE ASSIGN 25 TO WRFORM ENDIF WRITE (6,WRFORM) PNAME, AVG

20 FORMAT 25 FORMAT

..

.

..

.

name

of

an

as the

ASSIGN

statement

executable

statement.

label value

statement

The

may

be redefined with

an

integer value.

or

a

FORMA

variable

T

'name'

the

3.3.5 DATA Statement

The

DATA clements. or

'common,'

statements

The

DATA statements

The

DATA

where 'clist'

has

[r*]c[,[r*]c] ...

where

c

r

Dummy

'nlist'

the

statement

arguments

memory

are

part

statement

and

before

statement

nlist

felistl [[,]

is

a list

form:

is a constant

is a 'repeat'

'r*c'

gives the initial values

can be initialized by

of

a

BLOCK

must

any

statement

has the

nlist

of

variable

is

equivalent

and

functions

DATA

appear

format:

felistl] ...

(including a

character

function

names,

to

Or'

in a

and

of

variables, arrays,

cannot

subprogram

array

Hollerith

is a nonzero,

successive repetitions

be initialized by

DATA

(section 3.4.3).

program

or

executable

names,

constant)

statements

unit

and

array

unsigned, integer

DATA.

after

the specification

statements.

element

of

the

constant

if

and

Shared,

the

DATA

names,

constant;

'c.'

array

and

3-10

FORTRAN-80

Defining Variables, Arrays,

And

Memory

Items in 'Nlist'

to-one.

constant for each element

is specified must be

The constant in 'elist,' except

and

type

DATA

If

lists must agree in

'elist'

must

have the same

'nlist'

contains

of

a name specified in 'nlist' must agree with the type

an

array

of

that

an

integer constant.

that

an

constant. Given a length

corresponding initial Hollerith

'n'

is

less are equal. Note itializes

that

A variable

'g'

than'

g,'

that

entire item.

or

array element

of

a variable

the constant

initialization

cannot

constant

two symbols are associated, only Examples:

DATA

I,J,K/10,20,301

CHARACTER*10

DATA

NAMES/'GEHRIG',

number,

type,

number

and

length.

of

items, as the lists correspond one-

name without a subscript, 'elist' must have one

array

(but

see section F.2.9.2). Any subscript

of

the corresponding

item

of

any type can be initialized

or

array

element in

in 'elist' must be less

is

padded on the right with blanks until the lengths of

a character in a variable

be initialized

one

may be initialized.

'nlist,'

more

then the length In'

than

once in a

or

to

or

equal to

array

a Hollerith

element in-

program.

NAMES(3)

'OTT', 'RUTH'I

of

'g.'

that

its

If

INTEGER*1 ZEROS

DATA ZEROS 110*0/1,J,K

(10)

110,20,301

&NAMES(2) I'OTT'I

C AMPERSAND USED TO CONTINUE STATEMENT

LOGICAL TABLE(3) DATA

TABLE

I.TRUE., .TRUE., .FALSE.!

3.4 Memory Definition

The

DATA also ineludes three statements for establishing memory areas areas. These are the

EQUIVALENCE associating a variable which it appears. The

ferent

program BLOCK memory.

3.4.1 EQUIVALENCE Statement

The

EQUIVALENCE All entities listed in the memory (even

and

assignment statements assign values to specific items.

program

units, allowing

(for example, a

DATA

subprogram,

if

they are

EQUIVALENCE,

is

used

to

associate two

name

with

COMMON

COMMON,

an

array element. Its scope

statement can be used

common

data

which can assign initial values to items in

statement allows items in a

EQUIVALENCE

of

unequal lengths).

and

BLOCK

or

more

items in memory, such as

to

use

of

data

and

memory

base

or

table). BLOCK

program

statement share the same start address in

FORTRAN

and

initializing these

is

DATA

the

statements.

program

associate items in

through

DATA

an

defines a

common

unit

to

share memory.

unit in

dif-

entire

format

The

EQUIVALENCE

of

the

EQUIVALENCE

(nlist)

statement is:

[,(nlist)] ...

3-11

Defining Variables, Arrays,

where

'nlist'

may

ter

only be subscripted by integer qualified by a subscript Function

Equivalenced items ed.

The

EQUIV

Example:

And

is

names

Memory

a list

of

and

the names

may

ALENCE

variable names,

is

the same as a reference

of

have different

statement

array

dummy

data

does

names

or

constants.

The

to

arguments

types,

although

no

type conversion.

array

element names.

use

of

an

the first element

may

not

be listed.

this

is

not

array

name

of

the

recommend-

The

array.

FORTRAN-80

latun-

INTEGER* REAL

1,

ARRAY(3), LlST(6)

FPNUMB

EQUIVALENCE (ARRAY,LlST), (LlST(4),FPNUMB)

The

The than

resulting

ARRAY LIST FPNUMB

EQUIV

once in a

memory

ALENCE

memory

allocation would be:

I A(1) I A(2) I A(3) I

L(1) I L(2) I L(3) I L(4) I

I LOW I I HIGH I

statement

sequence,

must not cause the same storage item

nor

can

it

L(S)

result

quence.

Examples:

C

THE

FOLLOWING

EXAMPLES ARE

INVALID DIMENSION ARRAY(3) EQUIVALENCE (ARRAY(1),FPNUM B),(ARRAY(2),FPNUMB)

C

FPNUMB

OCCURS MORE

THAN

ONCE REAL TABLE(2), ROOTS(3) EQU IV

ALENCE

(T

ABLE(1), ROOTS(1 )),(TABLE(2),

&ROOTS(3))

C MEMORY SEQUENCE IS SPLIT

I L(6) I

in

the splitting

IN

MEMORY

to

occur

more

of a memory

SEQUENCE

se-

3-12

3.4.2 COMMON Statement

Common program. named

The

where

memory

The

or

unnamed

format

of

COMMON

cb

nlist

blocks allow

COMMON

(that

the

COMMON

[/[eb]!]

is

the

name

is

a list Function listed.

data

statement

is,

'blank').

statement

nlist

[[,]/ [eb]!

of a common

of

variable names,

names

and

defines

is:

nlist]

the

memory

common

...

block.

array

names

to be shared

blocks

names,

of

dummy

throughout

that

may

or

array

arguments

an

entire

be either

declarators.

cannot

be

FORTRAN-80

Defining Variables, Arrays,

And

Memory

The items in 'nlist' following a to be in statement refers to the blank omitted in the above present, however, block.

The statements, either in the same items in different ing the same name also have the same starting address in memory. The same of linked

3.4.2.1

A COMMON within the

An done by adding EQUIV same block are considered

that

block (or in blank common).

same

common

all declarations

program.

Common

common

EQUIVALENCE

block memory sequence consists

statement(s) for

COMMON

memory

ALENCE

program

unit

format,

if

blank

block

program

of

blank

Block Memory Sequence.

statement(s).

statement may cause a

statement must

to

be associated. Names associated with a

to

be

common

the slashes may be omitted also. The slashes must be

common

name

program

units are associated in memory. All

common.

that

beyond the highest location in the

part

of

block name (or omitted name) are declared

If

a

block.

is

specified as other

(or blank) can

unit

or

Only one blank

of

common

not

that

block, in the

common

cause two different

common

If

in

the

block.

the first

appear

other

program

common

memory

block

block name

common

than

the first

in

other

units. This

common

can exist in the final

of

all items listed in the

order

of

their appearance

to

be extended. This

common

name

is

omitted, the

block name

COMMON

blocks hav-

block. blocks in the in a

common

is

how

true

is

An

3.4.2.2 Named

Named and blank

• Within a same size. Blank

• Executing a common

•

DA

TA

named

Examples:

C PREVIOUS STATEMENT EXTENDS C

'BLOCK1'

C

BLANK ARRAY ELEMENTS

C

3.4.3 BLOCK DATA

BLOCK

named

DATA The in a SA VE,

DATA

common

statement, which

last statement must be the

BLOCK

EQUIVALENCE,

and

Blank

common

program,

common

RETURN

blocks

to

statements in a BLOCK

common

blocks.

COMMON

LOGICAL LOGICS(3,3),Z

COMMON COMMON

IS

EQUIVALENCE (Z,LOGICS(1,

COMMON

Common

blocks are treated differently in several respects.

all

common

block-s

or

END

become undefined. This

IBLOCK1/A,B,ROOTS X,Y,Z IBLOCK1/C

NOW

EXTENDED BY LENGTH EXTENDED

Subprograms

subprograms

blocks. The first statement

DATA

are used

mayor

subprogram

DATA,

may

END

and

Blocks.

blocks having the same

may

be different sizes.

statement sometimes causes items in named

cannot

DATA

statement. The only

are

subprogram

/I

BLANK

1))

AGAIN-BY

to

initialize variables

of

such a

not

name the BLOCK

IMPLICIT,

the type statements.

name

must also be

occur with blank

can

only initialize items in

TABLES(3,3)

COMMON OF

'C'

8 LOGICAL

and

array

subprogram

other

DIMENSION,

is DATA statements permitted

common.

elements in

the BLOCK

subprogram.

COMMON,

the

3-13

Defining Variables, Arrays,

3.4.4 BLOCK DATA Statement

The

format

BLOCK DATA [name]

where

'name'

Since

'name'

main

program,

unnamed

Only an item in a or

type

statement

mon

BLOCK

If

a

named

they

are

not itialized in a single block not

be specified in

Examples:

And

Memory

of

the

BLOCK

is

the

is

global,

common

BLOCK

DATA

J

common

in a

can

be initialized in a

common

all initialized.

more

symbolic

it

must

block,

subprogram

BLOCK

block

data

than

DATA

block

name

not

or

may

statement

of

be

another

DATA

BLOCK

is

initialized, all items in

More

than

subprogram,

one

block

BLOCK DATA BLK1 LOGICAL

FLAGS(3) INTEGER ZEROS(10),RESULTS COMMON

/BLOCK1/ FLAGS,ZEROS,RESULT

DATA FLAGS/.TRUE., .TRUE., .FALSE.!

&ZEROS/1 0* 0/

END

is:

the

BLOCK

the

same

BLOCK

is

permitted

appear

in a

subprogram.

DATA

one

named

but

data

subprogram.

DATA

as

the

name

DATA

per

executable

DIMENSION,

Only

subprogram.

the

common

the

same

subprogram.

of

an

external

subprogram.

an

item in a named

block

must

block

named

common

procedure,

Only

one

program.

EQUIVALENCE,

com-

be listed, even

may

have items in-

block

may

FORTRAN-SO

if

3-14

CHAPTER 4

PROGRAM

FORTRAN execution. These are statements tions), regulate execution loops tion

(PAUSE,

4.1

The

statements in this cases only when a stated condition as conditional native set

The

statements

4.1.1

includes

STOP,

16

statements,

END).

or

that

transfer control (GO

(DO,

Transferring Program Oontrol

group

pass

control

is

meL

and

unconditional branching. Some statements also allow

of

operations

that

Unconditional Computed Assigned Arithmetic Logical

GO

IF

Unconditional GO

transfer

GO

TO

IF

to

TO

be

performed

program

TO

Statement

control are:

EXECUTION CONTROLS

statement variations, for controlling

TO,

IF,

and

CONTINUE),

to

another

These alternatives are usually referred

if.the stated conditions are not met.

and

part

Block

ELSE ELSE END

terminate

of

the

program,

IF

program

their varia-

execu-

in some

to

an

alter-

The

unconditional

ecuted.

where as the

Example:

It

GO TO

'stl'

GO

GO TO 1010

has the

is

TO

GO

TO

statement transfers control

format

stl

the statement label

statement.

to

the next statement

of

an

executable statement in the same

4.1.2 Computed GO TO Statement

The

computed

based on the value

GOTO

is

GO TO

where

exp

stl

The same statement label may expression has a value in the range labels in the list), control passes to the statement pointed side this range, execution continues with the statement following the

GO

TO

statement branches

of

a controlling integer expression. The

(stl

[,stl] ... )[,]exp

is

an

integer expression

is

the statement label

program

unit as the

appear

of

computed

more

1::5

exp::5

to

one

of

several executable statements

format

an

executable statement in the same

GO

TO.

than

once in the statement.

n (where

'n'

is

to

the

by

number 'exp.'

program

of

the

If

GO

to

be ex-

computed

the integer

of

statement

'exp'

is

TO.

unit

out-

4-1

Program Execution Controls

Examples:

C IF K = C 1020

C NOTE

4.1.3 Assigned GO TO Statement

GO TO (1010,1020,1030) K

2,

FOR EXAMPLE, CONTROL PASSES TO STATEMENT

INTEGER*1 SWITCH SWITCH GO TO

THAT

GO TO

= K/J

(10,500,500,10,10)

'J' MUST BE .LE.

(10,

500, 600, 500) K * L + 1

'K'

SWITCH

IN THIS EXAMPLE

FORTRAN-SO

The assigned GO

TO variable name. Before the assigned defined with the value gram

unit.

The

format

GO TO name

where

name

stl

The

same statement label may thesized list must be one

Examples:

GO

TO

is

similar

to

of

the assigned

is

an

is

the statement label

program

of

statement labels

of

the labels in the list.

ASSIGN

GOTOSTART

ASSIGN 999 TO DONE

GO TO DON

statement

the

computed

of

a statement label by an ASSIGN statement in the same pro-

n,]

(stl

integer variable name

is

GO

TO

[,stl]

...

)]

unit as the assigned

appear

is

10TO

START

E (500, 600, 999)

used with the ASSIGN statement. The assigned GO

TO,

but in this case the control

GO

TO

is

executed, the variable name must be

statement

more

present, the statement label assigned

of

is

an

executable statement

GO

TO.

than

once in the statement.

is

an integer

III

the same

If

the paren-.

to

'name'

4-2

4.1.4 Arithmetic I F Statement

The

arithmetic IF statement behaves similarly transferred expression. The

to

one

format

of

three possible statements based on the value

of

the arithmetic IF

IF(exp)s1,s2,s3

where

exp

s

The

same statement label may

If

the value equals zero, control passes trol passes to the third statement.

of

is

an

integer

is

the statement label

program

'exp'

unit as the arithmetic IF.

is

less

than

to

or

real expression

appear

the second statement;

more

zero, control passes to the first statement listed;

of

to

a computed

is

an

executable statement in the same

than

once in the statement.

if

'exp'

is

greater

GO

TO.

of

than

a controlling

Control

if

zero, con-

is

it

FORTRAN-SO

Examples:

IF(A + 8)1010,1020,1030

Program

Execution Controls

SWITCH = IF (SWITCH) 100,200,300

A**2 -8**2

4.1.5 Logical I F Statement

We have given several examples effect, if the logical expression evaluated next.

If

the

logical expression

following the logical The

format

where

of

IF (exp)

exp

stmt

A function reference in the controlling logical expression parameters in the statement

Examples:

IF

statement.

the logical

stmt

is

is ELSE,

IF (SWITCH .EO. IF (SWITCH

IF

a logical expression

any executable statement except DO, a block IF,

END

of

logical

is

FALSE, execution continues with the statement

statement

IF,

END,

'stmt.'

1)

.NE.

1)

IF

statements already in this manual. In

is

TRUE,

is

or

other

GO TO FINISH WRITE (6,20) TOTALS

a specified statement

logical

IF

is

permitted

is

executed

ELSE

to

affect

IF,

LOGICAL DONE DONE IF (DONE) PAUSE IF (.NOT. DONE) GO TO START

=

(A * *2

.GT.

8**2)

4.1.6 IF, ELSE IF, and ELSE Blocks

The block IF statement

ELSE

IF

and

ELSE

Together with

'ELSE IF,

These blocks may be block, which may contain For

ment

An to, but nesting level as the block

An ment

has the same nesting level as the

IF

or

ELSE, respectively.

example, there may be no executable statements at all between a block

and

IF block consists

not

ELSE

up to,

other

blocks,'

its corresponding

including, the next ELSE IF, ELSE,

IF

block consists

but

not

is

used with the

statements

executable

or

'ELSE

'nested.'

of

all the executable statements

including, the nest

to

control

FORTRAN

blocks,'

another

END

IF

statement. of

the first statements

For

example,

IF

block, etc. These blocks may also be empty.

IF

statement.

all the executable statements

ELSE

IF

END

IF

statement

program

statement.

execution.

statements, they can form

of

which must be IF, ELSE

an

IF block may contain

after

the block

or

END

IF

statement on the same

after

IF,

ELSE,

or

END

and,

optionally, the

'IF

another

IF

statement up

the

ELSE

IF

statement

blocks,'

IF

state-

IF

state-

that

4-3

Program Execution Controls

An

ELSE

up

to,

but

ELSE

as the

For

each block IF statement, there must be a corresponding

same

program

block consists

not

including, the next

statement.

unit.

IF

...

ELSE

IF

...

IF

...

:

END

IF

_______________

ELSE

IF

...

END

IF

of

all the executable statements after the

END

IF statement

that

has the same nesting level

END

IIF

BLOCK

=:J

ELSE IF BLOCK

--,

IF

BLOCK ELSE BLOCK

_

IF.

. . ]

ENDIF

________

IF

BLOCK

IF

BLOCK

ELSE

statement

IF

statement in the

FORTRAN-SO

END

IF

______________

4.1.7 Block I F Statement

The

format

where

If

the value

block.

END FALSE, level as the block

Control section F.2.9.3).

If

the last statement in the IF block does statement, control passes level as the block

An

example

of

the block IF

IF

(exp)

THEN

'exp'

is

a logical expression.

of

'exp'

If

the

IF

IF statement

control passes

cannot

if

shown following the description

is

TRUE,

happens

on

the. same nesting level as the block

IF

statement.

execution continues with the first statement

to

be empty

to

the next

be transferred into

to

the next

IF

statement.

and

'exp'

is

TRUE,

ELSE

IF, ELSE,

an

IF block from outside the

not

END

IF statement

or

END

transfer control

of

the

END

_

of

the

control passes to the next

IF

statement.

IF

on the same nesting

IF

to

another

that

has the same nesting

If

'exp'

block (but see

executable

IF statement.

IF

is

4-4

4.1.8 ELSE I F Statement

The ELSE

where

IF

ELSE IF

'exp'

is

statement has the

(exp)

THEN

format

a logical expression.

FORTRAN-80

If

'exp'

is

IF

block.

TRUE,

If

normal execution continues with the first statement

'exp'

is

FALSE,

control passes

statement having the same level as the

ELSE

to

the

IF

statement.

IF, ELSE,

Program Execution Controls

of

the

ELSE

or

END

IF

Control section F.2.9.3). The statement label, referenced by

If

cannot

'be transferred into

another

statement.

the last statement in the ELSE

an

ELSE

if

IF

block does

IF

any,

ecutable statement, control passes to the next nesting level as the

An

example

ELSE

IF

statement.

is

shown following the description

4.1.9 ELSE Statement

Executing the is

ELSE

The

ELSE block ELSE statement. This ELSE

IF

Control section F.2.9.3). The statement label, referenced by

or

another

cannot

another

ELSE

statement has

is

terminated by

END

IF

ELSE

statement

be transferred into

statement.

no

effect; execution simply continues. Its

an

END

statement must

of

the same nesting level.

an

ELSE block from outside the block

if

any,

block from outside the block (but see

of

the

ELSE

IF

END

of

the

statement

not

pass control

IF

statement having the same

END

IF

statement.

to

another

cannot

format

IF

statement

appear

of

the same level as the

before the appearance

of

(but

of

the ELSE statement

cannot

be

ex-

an

see

be

An

example

is

shown following the description

4.1.10 END IF Statement

Executing the END

IF

block

IF

statement

The

format

ENDIF

Example:

END

IF statement has no effect; normal execution continues. The

statement acts as a

terminator

statement must have a corresponding

of

the same nesting level as the block

of

the

END

IF statement

IF (FLAG .EO.

3)

WRITE (6,20) NAME(3), PAY(3)

20

FORMAT

ELSE IF (FLAG .EO.

WRITE (6,40) FEDTAX

40

FORMAT

ELSE

WRITE(6,60) FICA

60 FORMAT

ENDIF

for IF, ELSE IF,

is

THEN

...

2)

THEN

...

of

the

END

IF

statement.

IF

statement.

and

ELSE blocks. Each

statement,

that

is, an

END

4-5

Program Execution Controls

4.2 Loop Control Statements

FORTRAN-80

Frequently, a series reading a series

ly). Rather

one

can create a loop until all entries have been read ment.

The

CONTINUE

statement, though it

than

of

operations must be repeated several times (for example,

of

entries from

repeat the statements

that

statement described in this section

is

not

an

input device

to

perform

causes the same statements

and

processed. This

limited

to

this use.

4.2.1 Operation of a DO Loop

The first statement labeled statement whose label define the range are executed a specific

The

DO

statement includes three values: an initial loop index value, a loop termination value, and decremented. Each time the loop ed to the initial index value, value until the loop termination value with the statement following the last statement

This sequence describes the most also terminate operation as the result loop, a transfer program,

or

program termination for any other reason.

of a DO

of

the

an

amount

of

control outside the loop, execution

loop

is

the

DO

statement itself. The last statement

is

specified in the DO statement. These two statements

DO

loop. The statements making up the body

number

of

times, as defined in the DO statement.

by which the initial value

is

performed, a

is

increased

common

or

is

reached.

operation

of a RETURN

and

extracting

these operations for each entry,

to

is

the function

is

'DO

variable,' previously initializ-

decreased by the incrementl decrement

Program

of

the DO loop.

of a DO

statement executed within the

information

be performed over

of

ordinarily used with the DO

is

to

be incremented or

execution then continues

loop. A DO loop can

of a STOP

statement in the

selective-

and

the DO state-

of

the

DO

over

is

loop

a

Program

control

cannot

be transferred into a

DO

4.2.2 DO Statement

The

format

where

In this format, and

'e3' ment iterations are performed (but see the discussion trols in section F.2.5).

The last statement TO, RETURN, logical ELSE

of

the

DO

statement

DO

sfl

[,] var = e1,

stl

var

e

is

of

one

assigned

IF

IF, ELSE,

is

the last statement in the is is

'eI'

the loop incrementl decrement

is

assumed. The values

GO

STOP,

statement, it can contain any executable statement except a

e2

the statement label

the name an

integer expression.

is

the initial loop index value,

of

a DO loop, statement

TO,

END,

or

END

IF,

is

[,e3]

of

an

executable statement and

DO

loop.

of

an integer variable, called the

amount.

of

'eI'

and

'stl,'

arithmetic IF, block IF, ELSE IF, ELSE,

DO

END,

statement.

or

another

If

the last statement

logical

loop.

'DO

'e2'

is

the loop termination value,

If

'e3'

is

not

'e2'

can be specified such

of

the

0077/0066

must

not

be

an

IF

statement.

is

the label

variable.'

specified,

unconditional

of

the

an

compiler con-

END

DO

loop

DO,

block IF,

incre-

that

of

no

GO

IF,

is

a

4-6

FORTRAN-80

DO

loops

may

be nested,

DO

statement the second can

share the same last statement.

I f a

DO

of

the

DO

appears

DO

statement

statement lies within

loop

must

within the range

be entirely within

that

is, a

DO

must

be within the range

an

IF

block,

loop

of

that

can

another

ELSE

block.

contain

DO

of

the

IF

block,

another

loop,

outer

or

ELSE

Program Execution Controls

DO

loop, etc.

the

loop

specified by

DO

loop.

block, the range

DO

If

loops

a

I f a block

statement must also be within the range Examples:

The

IF

statement

first example demonstrates the looping process.

is

within the range

of

N=O

DO 100 I

J = I

DO L=K

100 N = N + 1 110 CONTINUE

When

looping

of

blanks

use

The following example players (one team).

C LOOP C

is

completed~

to

isolate the nested

Compare

CHARACTER*12 PNAME DO READ 10, PNAME, AB, HITS

10 FORMAT

AVG = HITS/AB PRINT 20,

20

FORMAT (A12,

30 CONTINUE

CAN NOT TERM I NATE WITH STATEM ENT 20

BECAUSE

'FORMAT'

==

1,10

100 K = 1,5

I = 11, J = 10, K = 6, L = 5,

is

another

this

30

1=1,25

(A12, 2(2X, F3.0))

PNAME, AVG

IS

DO

loop

program

5X,

F4.3)

NONEXECUTABLE

of a DO

the

in this example.

to Figure 1-1...

loop, its corresponding

DO

loop.

to

compute

END

and

N = 50.

batting averages for

Note

IF

the

25

4.2.3 CONTINUE Statement

The

format

of

the

CONTINUE

statement

is

CONTINUE

as shown in the preceding example.

statement

This next executable statement.

has

no

effect

on

program

execution, which simply continues with the

4.3 Program Termination Statements

FORTRAN The

PAUSE statement program program.

provides three statements for halting

statement

terminates

unit. It terminates a

halts execution,

program

execution.

main

but

program

or

terminating

allows execution

The

END

statement

and

acts as a

program

to

resume.

marks

RETURN

execution.

The

STOP

the end

from

a sub-

of

a

4-7

Program Execution Controls

FORTRAN-SO

4.3.1

The

where time the console terminal.

Program program control, however, and might be initiated, for example, by rupt quence

Examples:

PAUSE Statement

format

such as a key being pressed.

of

the

PAUSE

PAUSE [msg]

'msg'

is

a string

PAUSE

execution must be resumable following the pause. Resumption

is

continued with the statement following the

PAUSE PAUSE PAUSE

is

executed

1234 'BREAK

statement

of

not

12'

is

more

than

five digits,

and

program execution ceases,

If

execution

is

or

is

a character constant.

'msg'

resumed, the normal execution

PAUSE

statement.

4.3.2 STOP Statement

The

format

of

the

STOP [msg]

STOP

statement

is

At

is

displayed on the

is

not

under

an

external inter-

the

se-

where

'msg'

is

a string

STOP

terminates execution

displayed on the console terminal. Examples:

STOP STOP

22

STOP

'CHECK

4.3.3

The

The gram, it terminates the program.

RETURN

gram unit

An must that

EN

0 Statement

format

END

of

the

END

statement

statement and returns

must

END

statement

not

be continued. No other statement in a

appears

to

be

of

not

more

of

the executable program. When this happens

SUM'

END

statement

marks

the end

an

END

statement.

is

written only in columns 7

an

END

statement.

than

five digits,

is

of

a program unit.

If

executed in a

to

the main program. The last line

or

is

a character constant.

If

subprogram,

through

program

unit can have an initial line

executed in a main pro-

72

'msg'

it has the effect

of

every pro-

of

an initial line and

of

is

a

4-8

CHAPTER 5

FUNCTIONS AND SUBROUTINES

Functions logical structures, conserve storage, avoid the tedium error functions.

The term is

referenced. A subroutine

alter the values All functions

• Intrinsic,

• Single-statement, user-defined functions (statement functions);

• User-defined function by their initial

• Subroutine subprograms identified by their initial

External functions cedures.' These may be the assembly language procedures)

When a procedure the place called. The use of

this chapter.

and

subroutines reduce coding, break programs into readily-visible

in repetitive coding, and eliminate the coding

'function'

program,

of

'actual

refers

to

a statement

is a subprogram

of

variables outside itself.

and

subroutines are called

or

predefined,

FUNCTION

and

or

they may be procedures created elsewhere (e.g.,

is arguments' to be substituted when the procedure

of

dummy

FORTRAN

subprograms

statement;

subroutines are referred to collectively as 'external pro-

FUNCTION

and

defined, it usually includes

and

actual arguments will become clearer in the remainder

or

subprogram

that

'procedures.'

functions;

(external functions), which are identified

or

SUBROUTINE

linked to the

and

increased probability

of

commonly-used mathematical

that

returns a value when it

does not return a value, but may

These include:

SUBROUTINE

subprograms

program

'dummy

where

appropriate.

arguments'

statement.

defined within

PL/M

used to hold

is

referenced

and

of

or

5.1

Intrinsic and Statement Functions

5.1.1

FORTRAN operations such as square tions, etc. The complete list dix tion definitions, type that PLICIT

An

The reference (for example, the actual value arguments with the specifications in Appendix B and may be any expression type.

Arguments for which the result numerical range tions on the range notes in Appendix

Intrinsic Functions

provides a

B.

This appendix shows the names

have

more

than

statement has no effect

intrinsic function

A = 33 + SQRT(8)

resulting value depends on the value

that

constitute the

of

number

of

arguments,

one argument, all arguments must be

is

referenced by specifying it in an expression.

the processor cause results

of

arguments

B.

of

predefined functions for performing

root

calculation, type conversion, trigonometric calcula-

of

available intrinsic functions can be

of

and

on

the types

argument

is

not mathematically defined

and

results for intrinsic functions are listed in the

the various intrinsic functions, their func-

type

of

results.

of

intrinsic functions.

of

the actual argument(s) used in the

of

'B'

in the expression above). The actual

list must agree in type, number,

not

For

those intrinsic functions

of

the same type. The IM-

or

defined in this manual. Restric-

common

found

in Appen-

and

order

of

the specified

which exceed the

5-1

Functions

And

Subroutines

I f

the

name

of

an

intrinsic

TION

or

SUBROUTINE the intrinsic its intrinsic quality. normal

If

the

procedure

function

(by

default

name

of

reference, the

an

intrinsic

or

function

subprogram,

within the scope

The

data

by a type

function

name

5.1.2 I NTRI NSIC Statement

The

INTRINSIC function pearance unit an

The

and of

is

allowed. A symbolic

EXTERNAL

format

statement

allows

a symbolic

statement

of

the

INTRINSIC

that

name

appears

type associated with the symbolic

statement).

is

must

first be specified in

confirms

to

be used as

in all

of

name

may

in the

same

statement

in the

the

name

of

the

program

to be used as

that

a symbolic

an

the

INTRINSIC

not

appear

program

unit.

is

dummy

is

considered

unit

an

actual

an

name

actual

statements

in

an

argument

to

have

and

the

name

argument

INTRINSIC

represents

argument.

in a given

INTRINSIC

list

of a FUNC-

no

name

is

specified as

in

an

statement.

an

Only

statement

relation itself loses

to

external

intrinsic

one

ap-

program

and

FORTRAN-SO

INTRINSIC

where

'func'

The

names therefore, type conversion ing a largest AMINO,

Examples:

cannot

MINI).

INTRINSIC INTRINSIC

tunc

[,tunc]

is

an

intrinsic function

of

certain

or

smallest value (MAXO,

intrinsic

appear

(lNT,

SQRT

EXP, LOG, LOG10

in

IFIX,

INTRINSIC

FLOAT,

...

functions

5.1.3 Statement Functions

Straightforward

= ax

f(x)

are

defined

keyword;

tunc

where

mathematical

2

+

bx

in

FORTRAN

the

format

is

([dum [,dum]

+ c

...

])

functions

using

= exp

name.

cannot

statements.

REAL,

AMAXI,

like

statement

be used as

ICHAR)

AMAXO,

functions.

actual These and

the functions for choos-

MAXI,

These

arguments

are

the functions for

MINO,

functions

and,

AMINI,

have

no

5-2

June

dum

exp

As

an

example,

F(X) =

Substituting would

result in the value

is

the symbolic

is a dummy the

function

is

an

expression

the

mathematical

A*(X**2)+

the

actual

argument is

referenced

8*X

+ C

argument

'9A

+ 3B +

name

function

'3'

of

the

statement

to

be replaced by

above

for

the

C.'

would

dummy

function

an

actual

be

written

argument

'X'

argument

in this

function

when

FORTRAN-80

The

statement

The

type

The

dummy

statement function.

function only,

The

type

used outside the The

name

the same type in

identify a variable

unit, but they have no

function

of

the value

argument

and

each

of a dummy

statement

of a dummy

other

of

name

and

returned

is

as shown in Figure 3-2.

list indicates the

The

names

of

dummy

name

may

argument

name

function.

argument

can

statement-function the same type other

relationship.

the expression

order,

arguments have a scope

appear

only once in the

is

the same as it would be

be used

statements.

or a common

5.1.3.1 Referencing Statement Functions

A statement function quired actual arguments).

is

referenced by specifying its symbolic

'exp'

may

number

to

and

identify

other

The

type

name

block within the same

Functions And Subroutines

be

of

different types.

of

arguments for the

of

the statement

dummy

argument

if

the

arguments

name

can also be used to

name

(with any re-

list.

were

of

program

DATA A,

FSUM(X) = A *(X* *2) + TOTAL

This

operation

function

B,

C

110.0,

= 33 + FSUM(3.0)

substitutes

definition.

At

10.0,

the

3.81

B*X

the

value '3.0'

end

+ C

of

the

operation,

for

every

occurrence

the

value

'156.8.'

5.1.3.2 Statement Function Limitations

All statement functions must follow all specification statements executable statements.

name

of

The symbolic in

any

specification statement, except in a type statement

mon

block in the same

an

EXTERNAL

A statement function

a statement function

program

statement

or

cannot

as

be

unit.

an

of

The

actual

type type character, the length specification tant.

Each

operand

in the statement function expression

ing:

• A

constant:

FPROD(C) = C*3.1412

• A variable reference:

FPROD(C)

•

An

array

element reference:

FSUM(C) = C +

•

An

intrinsic function reference:

FHVP(A,B)

• A reference

•

An

external function reference:

FSU.B(C) = 3*C

• A

dummy

FSUB(C) =

to a statement

FTOTAL(C)

procedure

=

C*3.1412+

2ITABLE(1,3)

= SQRT(A * *2 +

function:

=

C/3 + FUNC(3.8)

- EXFUN(3.

reference:

3*C

- EXFUN(X, Y,Z)

is

local

name

is

argument.

character.

of

that

argument

AV12

B*

*2)

0,

3.

and

cannot

or

as the

also prohibited from being used in

If

it has a

must

'exp'

must

0,2. 5)

of

'X' in the

of

'TOTAL' is

and

must precede all

be a symbolic

name

of

dummy

be one

be

an

argument

integer cons-

of

the follow-

name

a com-

of

5-3

Functions And Subroutines

A

statement

The

statement function FUNCTION reference

cause a ed.

is

dummy

5.2 External Procedures

This

section desi:ribes external

FUNCTION

to

external

function

defined later in the

subprogram

to

an

and

procedures:

may

function

external

argument

may

function

of

SUBROUTINE

RETURN,

be referenced only in

not

program

must

the

statement

reference

not

in

the

procedures,

subprograms.

SAVE,

another

unit.

Furthermore, a statement

reference

expression

function

EXTERNAL,

the

program

statement

the

name

of a statement

to

become

specifically

It

also describes

and

unit

where it

function

of

the

subprogram.

function

undefined

procedures

statements

the

subroutine

is

defined.

if

that

other

function

must

or

redefin-

defined by

FORTRAN-80

in a

A

not

5.2.1

The TION

program

The

where

FUNCTION Statement

FUNCTION

statement

must be

format

of

the

[typ] FUNCTION tunc

typ

June

dum

The

FUNCTION

must

become defined

of

this variable when a

An

external

addition

If

'typ'

conventions

function

to the

function

is

not

specified, the type

described in sections 2.2.2.1

statement

must always be the first

terminated

FUNCTION

is

either

1,2,

or

is

the symbolic

name.

is a dummy procedure

subprogram

or

in a

introduces a FUNCTION

with

an

END

statement

([dum

[,dum]

INTEGER

[*Ien],

4.

name

of

argument

name.

name,

appearing

redefined each time the

RETURN

subprogram

value

or

returned.

and

length

END may use

statement

statement.

is

...

])

REAL,

the

subprogram

and

is

either a variable,

as a variable within the

is

executed

dummy

of

'func'

and

2.2.2.2.

subprogram.

of

the

subprogram

or

LOGICAL

and

subprogram

is

the value

arguments

are

determined

[*lenJ, len being

is

an

external

array,

is

executed.

of

to

return

by the

The

FUNC-

and

the sub-

function

or

dummy

subprogram,

The

value

the

function.

values in

default

5-4

5.2.1.1

An with all necessary

The with their subroutine type.

•

• A

Referencing External Functions.

external

actual

Actual

An An An An

function

arguments

corresponding

name

arguments

expression array

name intrinsic function external

dummy

procedure

is

referenced by specifying its

actual

arguments).

in the reference

as

an

actual

must

procedure

dummy

argument; also be

name

must

agree in

arguments.

subroutine

one

of

the following:

order,

One

names

name

in

number,

exception

do

an

not

have

expression

type,

and

is

the use

an

associated

(along

length

of

a

FORTRAN-SO Functions And Subroutines

An actual a

dummy

reference.

5.2.1.2

A

FUNCTION subprogram. SUBROUTINE,

The other

program. name may

specified in the

be type character. The symbolic

program DATA,

A function specified by a subprogram procedure however, either directly

compiler control in section F.2.3).

When program.

gram before it TRAN

argument

argument list within the subprogram containing the external function

FUNCTION

name

of

name in the

Within the

appear

unit

or

COMMON

or

in the main

an

external function reference

External functions created outside the

compiler

may also be a

dummy

Subprogram

statement

The

subprogram

BLOCK

the

FUNCTION

subprogram,

is

FUNCTION

name

and

cannot

is

executed. External procedure linkage

operator's

may

be used only as the first statement

itself can consist

DATA,

subprogram,

a type statement,

of a dummy

statement (except as a

program.

or

PROGRAM

subprogram

except for its use as a variable in the

the only nonexecutable statement in which the

statement. A

argument

be used in an

A function

indirectly (but see the description

manual.

argument as long as the

Limitations

of

any

other

statement except a

statement.

is

global

and

even this

FUNCTION

in a function

EQUIVALENCE,

common

can

be referenced within any

is

executed, the function must be

and

is

not

subprogram

block name).

subprogram

program

cannot

be the same as

permitted

SAVE,

must.not reference itself,

of

the

must be linked is

described in the

dummy

of a FUNCTION

body

is

part

of

any

of

the sub-

if

the type

name

is

local to the

INTRINSIC,

other

REENTRANT

part

to

may

not

external

of

the

the pro-

FOR-

is

Example:

C THE FOLLOWING EXAMPLE TOTALS THE C VALUES IN AN ARRAY

FUNCTION

DIMENSION ARRAY(I)

TOTAL

DO

100 CONTINUE

RETURN

END

OF

LENGTH I

TOTAL(ARRAY,

= 0.0

100 K =

TOTAL

1,

I

= TOTAL + ARRAY(K)

5.2.2 Subroutines

A subroutine

be called from anywhere in a

dummy

its operations, then returns control

SUBROUTINE

The program.

statement

or a FUNCTION,

subprogram,

return

control

is

used primarily

arguments specified in the subroutine definition. The subroutine performs

statement

The

subprogram

and

can contain any statement except

BLOCK

the

RETURN

to

the point

to

avoid

redundant

program,

must

(being a DATA,

statement can be specified

of

call.

with actual arguments substituted for the

to

the point

be the fitst statement program

or

PROGRAM

I)

blocks

of

call.

unit) must be terminated by

anotqer

of

code. A subroutine can

of a SUBROUTINE

SUBROUTINE

statement.

to

At

some point in the

tell the subroutine when

sub-

an

END

statement,

to

5-5

Functions

And

Subroutines

The

subroutine

is

called by

the

CALL

PROGRAM

FORTRAN-SO

statement.

CALL

.~RETURN

CA.~END

END

Subroutines, time a subroutine program described in the

5.2.3 SUBROUTINE Statement

The

SUBROUTINE

SU

where

sub

being external

program

definition

containing a CALL

must be

FORTRAN

subprogram.

BROUTI N E

is

procedures,

part

of

the

or

by being linked

compiler

statement

Its

sub[([dum[,dum

the symbolic

name

to

calling

to

operator's

is

used only as the first

format

] ... ])]

of

..

SUBROUTINE

can be defined

the

subroutine

program,

the

program.

is

the

subroutine.

either by

manual.

outside

External

the

program.

is

executed, however, the

SUBROUTINE

procedure

statement

By

sub-

linkage

of

the

is

a

dum

If

no

dummy

'SUBROUTINE

The

symbolic

any

other

name

The

symbolic used in ment,

An

example

except as a

name

an

EQUIVALENCE,

is

5.2.4 RETU

The

RETURN

only in a

have statement

The

FUNCTION

one

or

terminating

format

RETURN

more

of

is a dummy procedure

arguments

subO'

name

of

in the

of a dummy

common

shown following

RN

Statement

statement

RETURN

the

RETURN

name.

are

is

acceptable.

the

subroutine

program

block

returns

or

SUBROUTINE

such a

statement

argument

specified, either

unit.

argument

SAVE,

name.

the

control

statements,

program

and

is

a global

is

INTRINSIC,

RETURN

to

or

unit has the

is

simply

is

either a variable,

of

the

forms

name

and

local to the

statement.

the

calling

subprogram.

they

DATA,

may

same

program

have

effect as a

array,

'SUBROUTINE

must

not

be the

unit

and

or

COMMON

unit. It

These

none

may

subprograms

at all.

RETURN.

or

dummy

sub'

same

cannot

appear

The

or

as

be

state-

may

END

5-6

FORTRAN-SO Functions

And

Subroutines

When a

RETURN

is

executed in a

FUNCTION

tion must be available to the referencing

Whenever

an

external procedure

RETURN

is

executed, the association between the

and

the current actual arguments

tion 5.2.5, the SA VE statement). Example:

C TH

REE N UM

C TO 11F THEIR

BERS ARE ADDED AN D TH E FLAG

TOTAL

IS

POSITIVE

SUBROUTINE POSTOT(A,B,C)

IF

((A

+ B +

C) POSTOT = 1 ELSE POSTOT

= 0

ENDIF

RETURN END

5.2.5 SAVE Statement

The SAVE statement can be used to FUNCTION RETURN

The

format

or

SUBROUTINE

or

END

is

executed.

of

the SAVE statement

subprogram

is

program

.GE.

make

subprogram,

the value

of

the func-

unit.

0)

THEN

sure

that

dummy

is

terminated (but see sec-

'POSTor

common

arguments

SET TO

variables within a

do not become undefined when a

of

SAVE

where that

block. A specific SA VE statement. Entities in blank executing

'cb'

is

RETURN

lebl

[,Iebl] ...

a named

common

or

END.

block. Naming the

bloj:k

cannot

common

be listed

never become undefined as the result

common

Example:

SAVE IBLOCK1/,/BLOCK21

5.2.6 EXTERNAL Statement

The

EXTERNAL

dummy The

procedure

format

EXTERNAL

where

'proc'

If

an intrinsic function can no longer be used stead becomes the

A symbolic name can be specified only once in all the program

unit.

statement confirms

and

allows

of

the

EXTERNAL

proe

is

the

name

to

name

that

name

statement

[,proe]

name of

...

of

an

external

is

specified in

specify

an

external procedure.

that

a symbolic

to

be used as

is

or

dummy

procedure.

an

EXTERNAL

intrinsic function in the

block saves all items in

more

than

name

represents

an

actual

argument.

statement,

program

EXTERNAL

once in a single

of

an

external

that

or

name

unit, but in-

statements in a

Example:

EXTERNAL HYP,

POSTOT, SIN

5-7

Functions

And

Subroutines

5.2.7 CALL Statement

The

CALL enced within must REENTRANT

The

format

where

statement

any

other

not

reference itself either directly

compiler

of

the

CALL

sub

[([arg [,arg].

is

used

external

control

statement

to

reference a procedure

or

in section F.2.3).

is

..

])]

subroutine. A subroutine

or

in the

main

program. A subprogram

indirectly

(but

see the description

can be refer-

of

the

FORTRAN-SO

sub

arg

The

actual and

(if applicable) length with the

referenced

ment, the type

is

the symbolic

is

an

arguments

subroutine.

conformity

actual

in the

If

associated type.

Each actual

•

An

•

An

argument

expression array

name

must

• An intrinsic function

• An external

• A

dummy

An

actual

name

appears

procedure

argument

in a dummy

in a

CALL.

Examples:

THE

C

C C

FOLLOWING

THE

SUBROUTINE

EXAMPLE

(SECTION

name

argument.

CALL

the

name

rule does

be

one

of

name

CALL

statement

argument

STATEMENTS

DEFINED

5.2.5)

of a subroutine

or

dummy

statement must agree in

corresponding

of a subroutine

not

apply since subroutines

dummy

is

specified as an actual argu-

the following:

may

be a

dummy

list within the

COULD

IN

THE

subprogram

BE

'RETURN

procedure.

order,

number,

argument

do

argument

containing the

USED

TO

CALL

STATEMENT'

list

not

name

type,

of

have

if

the

an

that

5-8

CALL CALL

POSTOT(3.2,-2.7,0.08) POSTOT(X,5.2**I,-Y)

5.3 Arguments And Common Blocks Revisited

Arguments cedures ment

external procedures common

FORTRAN

and

argument subroutine of

others

and

function

common

statements referencing the procedures.

or

intrinsic function by

and

blocks

other

are

program

the means

an

argument

units by

of

communicating

Data

can be passed

list.

Data

argument

can be passed between

lists

blocks.

has a

number

lists. We

of

very specific rules governing the use

touched

on

the most

important

references were discussed. This section repeats those rules plus a

arguments

and

common

blocks.

between pro-

or

can be shared in

of

common

when function

to

a state-

number

blocks

and

FORTRAN-80

Functions And Subroutines

5.3.1 Common Blocks

Common to share the same referencing

The

subprograms memory given data the same type.

An gram

blocks reduce storage requirements by allowing two

memory.

data.

variables

common

referenced

integer variable

unit containing its

and

that

location

block

rather

and

arrays in a

contain

ASSIGNed

This sharing

common

a declaration

than

by

may

be different in the different

the

common

a statement label

ASSIGN

may

block

name,

statement.

so the names

block name used to reference the

be limited by the rules for defining

can

of

that

5.3.2 Dummy And Actual Arguments

A

dummy tual referenced.

Dummy SUBROUTINE

argument

the tion, the

argument

arguments

dummy

is

used in the

is

used in the corresponding

are used by statement functions,

subprograms

is

a single value,

argument

must be a variable.

argument

to

specify the types

array,

or

list when defining a procedure.

argument

procedure.

or

more

subprograms

be defined

common

of

subprograms.

can

only be referenced in the pro-

FUNCTION

of

actual arguments

In

the case

and

referenced in all

block. 'Association

variables

list when the procedure

and

arrays

However, the

data

must be

subprograms,

and

of

a statement func-

and

is

by

of

An

ac-

is

and

whether

a

Dummy

(that

prohibited.

SA VE,

dummy ment function

Actual

arguments

function sions (if the associated ternal procedure). defined later in the

argument

is, variable,

Dummy

INTRINSIC,

name

also

name

arguments

for a

or

subroutine

names

can

be used anywhere

array,

or

dummy

argument

or

COMMON

cannot

particular

An

program

be the same as a

declared in the same

list the entities to be associated with corresponding

procedure

references,

dummy

actual

procedure)

names

argument

unit containing the reference.

cannot

statements (except as

program

reference. Actual arguments

or

expressions, including parenthesized expres-

argument

cannot

5.3.3 Association Of Arguments

When a procedure associated; the first actual argument, when the actual

If

an sion, the actual is

referenced.

actual

etc.

argument

reference

Therefore,

argument

is

associated with a

argument

is

executed, the actual

argument

the

arguments

is a subroutine

must be defined with

replaces all occurrences

must agree in

name),

dummy

an

actual

name

of

and

type can be used, unless explicitly

appear

FUNCTION,

is

not defined during execution

in

EQUIVALENCE,

common

SUBROUTINE,

unit.

be the

name

of

a statement function

and

dummy

of

order,

number,

and

length (where applicable).

appearing in

an

integer value when the

an

adjustable

may

the first

the same 'class'

DATA,

block names). A

or

state-

dummy

be constants,

of

the ex-

arguments

type (except

are

dummy

dimen-

procedure

If

an tion. tion.

actual

If

the actual

argument

is

an

is

an

expression, the expression

array

element name, its subscript

is

evaluated before associa-

is

evaluated before associa-

5-9

Functions

And

Subroutines

NOTE

The subscript value remains constant as long as the arguments are associated, even

if

the subscript contains variables redefined during the

association.

FORTRAN-80

Argument association can be carried

through

more

than

one level

of

procedure reference. A valid association exists at the last level, however, only if the association is

maintained when a cedure reference

5.3.3.1 Agreement

If

a

dummy

and

length.

through

RETURN

argument

If

the

all intermediate levels. The association normally terminates

or

END

is

to

the next.

Of

dummy

executed. The association

Argument Lengths

is

type character, its actual argument must be the same type

is

an

array name, the length requirement applies

is

not retained from one pro-

to

each

element in the array.

If

the dllmmy

lepgth~lfaninteger

as

shq""n

integer

c~ssordetermines

argumentis

c()nstant

in~ection

variable

2.2.Z.2.That

length;if

its length implicitly.

type integer. the arguments must again

is

used as

the base

is,

an

of

the integer

an

actual

integer

argument, its length is deterInined

without

abas~suffix

constantisstated

agreeiu

type

and

has the default

explicitly, tbe pro-

5.3.3.2 Variables As Dummy Arguments

A variable ment, name, its Otherwise, if the actual argument dummy

dummy

or

expression.

dummy

argument

argument can be associated with an actual variable, array ele-

If

the actual

argument can be defined

must

not be redefined within the subprogram.

argument

is a constant,

is

a variable name

or

redefined within the subprogram.

function reference,

or

array

or

expression, its

element

5-10

5.3.3.3 Arrays As Dummy Arguments

An array ment name. The declarator can be declarator.

If

not

If

ray element with a subscript with the subscript value (p than ment.

dummy

argument

number

differept from the dimensions in its associated

the actual argument

be greater

than

the size

the actual argument

the size

of

the actual array plus one minus the subscript value

can be associated with an actual and

size

of

dimensions in an actual

is

an array name, the size

of

the actual argument array.

is

an array element name with a subscript

'q'

becomes associated with the actual array element

+

q-l).

The size

of

5.3.3.4 Procedures As Dummy Arguments

I f a

dummy argument a

dummy actual argument subroutine referenced as a function.

argument

is

used as if it were an external function, the associated actual

must be an intrinsic function, external function, argument appears in a type statement

argument

must be the

is

referenced as a subroutine, the actual argument must be the

or

dummy procedure

name

of

a function

and

must

not

of

the

dummy

the

dummy

array

or

and

an

EXTERNAL

or

dummy procedure.

appear

in a type statement

array

name

or

array ele-

argument

dummy

argument array must

'p,'

the

dummy

must not be greater

of

the array ele-

dummy

procedure.

statement, the

If

the

dummy

name

array array

ar-

If

of

or

be

a

FORTRAN-SO

NOTE

Functions

And

Subroutines

In a given associated cedure statement, enough with a

A

dummy association. the

result

of

dummy

as a the

subprogram. A dummy

procedure

as a number

5.3.3.5

If reference the

and

Argument Association Limitations

dummy

to

subprogram.

program

with a

name

and

to

determine

subroutine

argument

Therefore,

all

actual

external

name

type

with

arguments

that

subprogram, For

FUNCTION F(A,B,C)

could

not

be

referenced

F(X,

Y,

X)

function

appears

an

actual

or a function.

associated

the

arguments

function

in a

those

in

the

example,

by

unit,

det~rmining

or a subroutine

only

in a

dummy

argument

whether

type

argument function

specified

with

of

the

associated

name

same

neither

an

external

the

an dummy

loses its

associated

reference

for

subprogram

dummy

whether a dummy

may

not

argument

list,

examining

symbolic

intrinsic

the

function

name

argument

with it.

property

with

must

intrinsic

are

can

function

An

have function.

be

beginning with

be possible.

list,

an

the

subprogram

should

has

must

agree

intrinsic

as

an

intrinsic

an

intrinsic

arguments

associated

redefined

procedure

If a pro-

EXTERNAL

is

be

associated

no

automatic

with

the

function

as

during

name

function

that

the

result

execution

is

not

type

used

within

and

used

agree in

of

a

of

because'

If a subprogram item in a

mon

item

A'

and

common

can

be

'C'

would

reference

block

redefined

become

causes a

in

the

within

associated.

dummy

same

subprogram,

the

subprogram.

argument

neither

to

become

the

associated

dummy

nor

with

the

an

com-

5-11

The

FORTRAN

and

external units

ternal units

the

format

input/output

or

within the processor itself. These statements can specify the ex-

to

be used, the variables whose values are being entered

of

I/O

data.

(I/O)

statements transfer

data

CHAPTER

INPUT/OUTPUT

between a processor

or

output,

and

61

The first

BACKSPACE, used for connecting of

The external unit

by the output

group

of

I/O

statements are the file-handling statements

REWIND,

and

a file.

to

be used

READ,

is

WRITE,

formatted, these statements are used with the

ENDFILE).

disconnecting files, positioning files,

and

the list

PRINT

of

data-transfer

(OPEN,

As their names imply, these statements are

and

marking the end

variables to be

input

or

output

statements. When the input

FORMAT

statement.

CLOSE,

is

supplied

or

~~f:~~~~t~~~~~r.~~~~I~~.9:~~~~.~I~Btg1~<l~~t~.*P~~~2t:E~?P:S>~~i~~Intetf:9RjR~N

?~~~;~~~f~rp/~Ptq~$gt;l~~!pp.qr~lif:.~~~'~~Jl9.~~ffJf!J;Jff

s.ectionF.2,2. . .

Before any units are reviewed. They can be described only in general here, however, since many of

the specifics related to them are dependent on the processor

ment.

6.1

6.1.1

A record generally the same as the sum (in bytes) is

stated specifically in a record length specifier (section 6.2.1.8). Records are

classified as formatted

of

these statements are discussed, the properties

Records, Files, And Units

Record Properties

is

simply a sequence

or

of

values

unformatted.

or

of

the items written into the record, unless it

characters. The length

••

jq~ril1s~<:.

of

su

broutine~

records, files,

or

operating environ-

of

a record

·in

and

is

A formatted record can be any sequence cessor (except escape, carriage record can be a sequence

data.

These two kinds

formatted

I/O

statements, respectively.

of

return,

of

values containing

records can only be read

or

6.1.2 File Properties

The main properties

•

That

is

may exist;

•

That

it may have a name;

•

That

it has a position;

•

That

it

may be external

•

That

it may be accessed sequentially

•

That

it may have records

•

That

it

may have formatted

of

a file are:

or

internal

of

specific length or

unformatted

to

the processor;

or

of

characters representable in the pro-

line feed characters). An

both

character

or

written by formatted

directly;

(for

direct access files);

contents.

unformatted

and

non character

and

un-

6-1

Input/Output

6.1.2.1 File Existence

At

the time an executable program files are said to operating system

A file may exist without containing any data; an example would be a newly-created file having only a name.

exist,

and

or

environment in which the

is

running, a certain set

the particular files

program

of

files

is

available. These

that

exist are determined by the

is

running.

FORTRAN-SO

FORTRAN

All CLOSE, ist

and

are in the process

6.1.2.2

FORTRAN dependent. names:

File Name

MYFILE MYFILE.SRC :LP: :F1:PROG.OBJ

I/O

WRITE,

has no

For

example,

statements can refer

PRINT,

standard

and

ENDFILE)

of

being created.

file-naming convention. Such conventions are system

in

the ISIS-II environment all

1-6

character file name

file name plus device name, in this case line printer device name plus file name plus extension

to

existing files. Some statements

can also refer

1-3

character extension

to

files

that

of

the following are valid file

6.1.2.3 File Position

Once a file has been connected to a unit, it has a position. The 'initial is

the position preceding the first record. Its 'terminal the last record. record.'

Executing certain can cause the file position to become indeterminate.

If

the file

I/O

is

positioned within a record,

statements affects the position

point'

is

the position

that

record

of

the file. Some circumstances

do

not

point' is

the

(OPEN,

yet ex-

of

a file

just

after

'current

6-2

6.1.2.4 External

An

external file character variable, character array, you to transfer

An internal file CLOSE, BACKSPACE, sequential- access, formatted matting.'

An

internal file has the following properties:

• Each record

• The length

array element, it

array element. I f the file

an

array element in the array

• A record may be read only if it has been defined. A variable

record

ment statement).

the record, the characters are left-adjusted in the record the record

•

An

is

internal file

And

Internal Files

is

a file

that

can be connected to an external unit.

or

character array element. Internal files allow

data

within processor memory.

cannot

is

of

defined by writing the record (or by making it the target in

is

filled with blanks.

be specified by one

REWIND,

I/O

statements

a character array element. the file depends on its kind.

is

a single record whose length

is

a character array, every record has the same length as

and

If

the number

is

always positioned

of

ENDFILE).

the file has as many records as array elements.

characters written

at

its initial point before a

the file-handling statements

It

can only be read

that

do not specify 'list-directed for-

If

the file

is

the length

is

a character variable

less

and

An

internal file

(OPEN,

or

written by

of

the variable

or

array element

an

than

the length

the remainder

data

transfer.

is

or or

assign-

of of

a

FORTRAN-SO

6.1.2.5 File Access

An

external file can be accessed sequentially

be accessed sequentially.

Some external files

operating to

a unit.

environment.

may

be allowed The

access

more

method

or

directly.

than

one access

is

determined when the file

An

method,

internal file can only

depending on the

is

connected

Input/Output

6.1.2.5.1

Sequential Access File. A file connected for sequential access has the

following characteristics:

•

The

records are a totally ordered set, having the

written.

•

The

records

If

the file

the same as the

6.1.2.5.2

are

either all

may

also be connected for direct access, the

order

of

the sequential write.

Direct Access Files. A file connected for direct access has the following

characteristics:

• The

• Its records are either all

read

order

or

written in

of

its records

any

order.

formatted

• Its records can be read List-directed

• All records

• Each record

written. A record

formatting

of

the file

of

are

the file has a unique record

cannot

ten.

6.1.3 U nit Properties

A

'unit'

is

a logical way executable CLOSE

program.

statement

A unit has the refers

to

a file. All

connected to a file.

property

of

referring

All

FORTRAN

can

also refer to nonexistent units.

of

110

statements except

formatted

is

the

or

cannot

or

all

unformatted.

order

of

their record numbers. Its records can be

or

all

unformatted.

written only by direct-access

be used.

the same length.

number,

be deleted

being connected

or

its

number

to

a file. Like files, units can exist

110

statements can refer to existing units. The

or

OPEN

and

order

in which they were

order

of

a direct-access read

110

specified when the record

changed,

but

is

can be rewrit-

or

disconnected.

CLOSE

If

connected,

must

reference a unit

is

statements.

is

not for

an

it

is

Typically, a file CLOSE be

statement. Depending

preconnected, meaning they

connected by the

having to be connected by the nected the first time it

environment the console

nected. See the discussion A unit

may, the same

must

not be connected to

of

course, be disconnected by the

or

a different unit by the

The

only way

is

referenced by

output

of

to

refer

to

statement. Consequently,

it

nected once

has been disconnected.

OPEN

on

the

operating

can

be referenced by

OPEN

the

statement. A preconnected file becomes con-

an

110

device

and

UNIT

run-time

more

than

CLOSE

OPEN

statement.

NOTE

a disconnected file

an

unnamed

statement

and

disconnected by the

environment, some units may also

110

statements without first

statement.

console

control

one file

statement

file

may

For

example, in the ISIS-II

input

device

in section F.3.

at

a time,

and

is

by naming it in

not

be able

are

always precon-

and

vice versa. A file

then reconnected

an

OPEN

to

be recon-

to

6-3

Input/Output

6.2 File-Handling Statements

6.2.1 OPEN Statement

The

OPEN preconnected file, create a file the

file/unit

The

format

OPEN (open-list)

statement can be used to connect

and

connection.

of

the

OPEN

statement

connect it

is

to

an

existing file to a unit, create a

a unit,

or

change certain specifiers in

FORTRAN-SO

where 'open-list'

[UNIT IOSTAT = ERR = FilE STATUS = ACCESS = aee FORM = RECl BLANK = b/nk CARRIAGE::::;

The unit specifier specifiers are

is

being connected for direct access. Some specifiers have default values.

file

The following sections 6.2.1.1 through 6.2.1.10 describe each

specifiers in detail.

6.2.1.1

The

Unit

format

[UNIT

is

a list

=]

unit

stname

st/

= fname

stat

fmat

= ree/en

car

'unit'

optional

Specifier

of

the unit specifier

=]

unit

except

of

specifiers separated by commas.

U nit specifier

II

0 status speci fier

Error

specifier

File

name

specifier

File status specifier

must be present

that

the record length

is

Access

Formatting Record length specifier Blank specifier

Carriage

method

specifier

control specifier

and

the unit specified

(REeL)

The

specifier

must

list

of

specifiers is:

exist. All

be specified

of

the 'open-list'

other if

the

6-4

where

'unit'

identifies

specifier must be the first item in

An external unit identifier

0:$unit:$255.

Examples:

OPEN OPEN

6.2.1.2

The

format

IOSTAT =

where

(UNIT=3) (4)

1/0

Status

of

'stname'

the

an

Specifier

input/

output

stname

is

an

integer variable

external unit.

must

If

the

optional

'open-list.'

be

an

integer expression whose value

status specifier

or

integer

is

array

'UNIT

element

='

is

omitted, the unit

name.

is

in the range

FORTRAN-80

Input/Output

Executing defined with a zero value positive integer value

Example:

6.2.1.3

The

where 'st!' OPEN

If

the processor discovers

the following steps occur:

1.

2.

3.

4. Execution continues with the statement named by the

OPEN

Error

format

ERR =

statement.

The

OPEN

The

position

If

the

error

condition;

an

OPEN

(4,

statement containing this specifier causes

if

an

error

IOSTAT = ERRFLG)

Specifier

of

the

error

specifier

stl

is

the label

operation terminates;

OPEN

of

an

of

the file specified by

statement has an

no

error

condition exists,

condition does exist.

is

executable statement in the same

error

condition while executing the

OPEN

IOSTAT

specifier,

or

becomes indeterminate;

'stname'

with a processor-dependent

program

OPEN

is

ERR

specifier.

set

to

become

unit as the

statement,

reflect the

Example:

OPEN

6.2.1.4

The

format

FILE = {name

where constant

program

the processor-determined file.

Example:

OPEN(UNIT=3,

6.2.1.5

The

format

STATUS = stat

(4,

IOSTAT=

File

Name

of

the file name specifier

'fname'

is

or

variable. The file

is

executing.

File

Status

of

the file status specifier

Specifier

the

name

Specifier

FILE=

ERRFLG,ERR= 1010)

is

of

the file

name

If

'MVPROG.SRC')

to

be connected, expressed as a character-type

must be valid for the particular system in which

FILE

is

omitted, the unit becomes connected to a

is

where

'stat'

is 'UNKNOWN'. 'UNKNOWN'.

a character expression evaluating

If

the STATUS specifier

to

'OLD',

is

omitted, the default value

'NEW',

'SCRATCH',

or

is

6-5

Input/Output

If

'OLD'

file

The specified with duration unit.

If

'UNKNOWN'

Example:

or

must

exist already; a

'SCRATCH'

of

OPEN

(3,

'NEW'

option

an

unnamed

program

is

specified, the file

FllE=

is

specified, the

'NEW'

must

file, the file

execution

FILE

specifier

file

not

or

cannot

until a

exist already.

be specified with a

is

connected

CLOSE

status

is

processor dependent.

must

statement

'MYPROG.SRC', STATUS = 'NEW')

be present also.

named

to

the specified

is

issued for the same

file.

An

When

unit

'OLD'

it

is

for the

FORTRAN-SO

6.2.1.6

The

Access

format

Method

of

the access

Specifier

method

specifier

is

ACCESS = ace

where

'acc'

is a character (see section 6.1.2.5). TIAL'.

If

the file already exists, the specified access

For

a new file, the processor creates the file with the specified access

access

method

sent in Example:

'open-list.'

OPEN

& ACCESS = 'SEQUENTIAL')

6.2.1.7

The

formatting

formatted

FORM =

where

FORMATTED'.

FORMATTED' TED'

'fmat'

when the file

is

'DIRECT',

(3,

FllE='MYPROG',

Formatting

specifier states whether a file

input/

output.

(mat

is a character

If

when the file

is

expression evaulating

If

the

ACCESS

the record length specifier (6.2.1.8)

specifier

to

is

omitted,

method

'SEQUENTIAL'

must

STATUS='NEW',

Specifier

is

being connected for

Its

format

the

FORM

being connected for sequential access.

is

expression evaluating

specifier

is

being connected for direct access,

is

omitted,

to

the

'FORMATTED'

or

'DIRECT'

the

default

be allowable for

default

is

'SEQUEN-

method.

must

also be pre-

formatted

value is

and

'FORMAT-

that

or

If

or

'UN'UN-

file.

the

un-

6-6

If

the file already exists, the specified

new file, the processor creates the file with the specified

Example:

OPEN

& ACCESS = 'SEQUENTIAL', FORM = 'FORMATTED')

6.2.1.8

The

record

nected for direct access. Its

RECl

where

(3,

FllE='MYPROG',

Record

Length

length specifier identifies the length

Specifier

format

= ree/en

'reelen'

is

a positive integer expression.

formatting

must be legal for

STATUS='NEW',

of

is

each

formatting.

record

in a file being con-

that

file.

For

a

FORTRAN-80

If

the file

If

the file

If

created. In the case

is

being connected for formatted

is

being connected for

unformatted

110,

the file already exists, the length specified must be

of

a new file, the processor creates a file with the specified length

for each record. The

REeL

specifier must be included in the

OPEN

connected for direct access.

Example:

'reclen'

is

the

the length

that

number

is

used when the file was

statement when a file

of

characters.

measured in bytes.

is

being

Input/Output

OPEN

& FORM = 'FORMATTED',

6.2.1.9

The

format

BLANK

where specifier

If

'NULL'

cept

that

(3,

FILE='CARDS',

Blank Specifier

of

the blank specifier

=

blnk

'blnk'

is

one

of

the character constants

is

omitted, the default value

is

specified, all blanks in numeric

a field

of

all blanks has the value zero.

STATUS='NEW',

RECL=80)

is

cept leading blanks have the value zero. The

specifier

is

permitted only for files being connected for formatted

Example:

OPEN

&

FORM='FORMATTED',

(UNIT=3,

FILE='TOTALS',

BLANK='ZERO')

ACCESS='DIREcr,

'NULL'

'NULL'.

formatted

If

'ZERO'

STATUS='NEW',

or

'ZERO'.

input

is

specified, all blanks ex-

If

the BLANK

fields are ignored, ex-

input/output.

6-7

Input/Output

FORTRAN-SO

6.2.1.11 Opening A Connected

The

OPEN

file.

That

not

included in the

I f the file to be connected depends

nected,

statement can be specified for a unit already connected to

existing file

is

OPEN

assumed

'open-list.'

is

on

whether

or

if

it was preconnected

or

not the file was preconnected.

Unit

to

be the value

the same as the connected file, the effect

and

ment, only the BLANK specifier (and

If

existing attributes.

the file was preconnected,

the file, the propetties specified in the

connection. Subsequent

OPEN

statements can change only the BLANKS

attributes as described above.

I f the file connected file deleted status

I f a file

if

it was previously opened with status

is

'OLD',

is

already connected to a unit, no

different unit

be connected

is

closed 'NEW',

can

be executed.

is

not the same as the file already connected, the currently

and

the new file

or

'UNKNOWN'.

to

6.2.2 CLOSE Statement

The

CLOSE

IS

statement

is

used

to

disconnect a particular file

of

'fname'

if

the

FILE

If

the file was not precon-

has already been referenced by

RECL

is

for sequential files) may differ from

but

no

I/O

has been

OPEN

statement become the properties

opened with this unit.

'SCRATCH',

OPEN

statement connecting

The

but

from

is

a unit. Its

connected file

not

an

existing

specifier

of

OPEN

an

I/O

state-

performed

and

RECL

deleted

that

file

format

if

to

is

on

of

is

its

a

CLOSE (close-list)

where 'close-list'

[UNIT IOSTAT = ERR = STATUS =

The

unit specifier must be present; all

=]

stl

is

a list

unit

stname

stat

of

specifiers separated by commas.

U nit specifier I/O

status specifier

Error

specifier

File disposition specifier

other

specifiers are

The

optional

list

specified only once.

The

IOSTAT

and

6.2.1. 3, respectively.

and

ERR

specifiers have the same interpretation as in sections 6.2.1.2

The

UNIT

and

STATUS

specifiers are described in the

following sections.

6.2.2.1 Unit Specifier

The

unit specifier has the same interpretation as in section 6.2.1.1. Execution CLOSE as its corresponding CLOSE

Once a unit has been disconnected by the to been disconnected, it the file still exists.

statement containing this specifier need

OPEN

statement, however.

has no effect.

the same file

or

a different file within the same

may

be reconnected

CLOSE

to

the same

not

occur in the same

If

the specified file does not exist,

statement,

program.

or

it

may

Similarly, once a file

a different unit, so long as

of

specifiers is:

and

can be

of

program

unit

be reconnected

the

has

6-8

Example:

CLOSE

(3,

IOSTAT=

ERRFLG,

ERR=

1020)

FORTRAN-80

Input/Output

6.2.2.2

The

where specifier status be specified for a file opened with

If statement

If

File Disposition Specifier

format

STATUS =

'KEEP'

'DELETE'

of

the file disposition specifier

stat

'stat'

is

a character expression evaluating

is

omitted, the default value

of

'SCRATCH',

is

specified for

is

executed. Otherwise,

is

specified, the file ceases to exist

is

to

'KEEP'

is

'DELETE'

and

'KEEP'

otherwise. Under no circumstances can

'SCRATCH'

an

existing file, the file continues to exist

'KEEP'

has no effect.

for a file

status.

after

the

CLOSE

ecuted. Following

normal

are deleted; all others

program

termination, all connected units are closed. Scratch units

are

closed with disposition

'KEEP'.

Example:

CLOSE

(3,

ERR = 1020, STATUS

='

KEEP')

6.2.3 BACKSPACE Statement

The

BACKSPACE

positioned before the preceding record. Its possible formats are

statement causes the file connected to the specified unit to be

or

'DELETE'.

that

previously

after

statement

the

If

this

had

'KEEP'

CLOSE

is

ex-

a

BACKSPACE BACKSPACE (arg-list)

where

'unit'

is

by commas. The list

[UNIT

IOSTAT = ERR =

=]

unit

an external unit specifier

of

arguments is:

unit

stname

stl

and

'arg-list'

is

a list

External unit specifier

110

status specifier

Error

specifier

of

arguments separated

The argument list must include an external unit specifier (section 6.2.1.1) contain

an

110

status specifier (section 6.2.1.2)

and

an

error

specifier (section

6.2.1.3). The file being backspaced must be connected for sequential access.

If

the file has no preceding record, the

end-of-file condition has occurred, the file

BACKSPACE

is

positioned such

statement has no effect.

that

the last record

the file becomes the preceding record.

Backspacing over a record written using list-directed formatting Backspacing a nonexistent file

or

unit

is

prohibited also.

is

not

Examples:

BACKSPACE 3

(3,

BACKSPACE

ERR = 1020)

6.2.4 REWI N D Statement

The

REWIND

tioned

at

possible formats for the

statement causes the file connected

to

the specified unit

to

its initial point. The file must be connected for sequential access. The

REWIND

statement are

and

may

If

the

of

allowed.

be reposi-

REWIND REWI N D(arg-list)

unit

6-9

Input/Output

FORTRAN-80

where described

but has

Examples:

'unit'

for

If

the specified file

does

not

occurred,

is

an

external

the

BACKSPACE

is

already

exist,

the

file

can

unit

positioned

REWIND

still be

statement

rewound.

REWIND 3 REWIND

(3,

IOSTAT=

ERRFLG,

6.2.5 ENDFILE Statement

When

the

ENDFILE

becomes

on

must

The

the

the file

be

connected

possible

last

without

formats

ENDFILE ENDFILE (arg-list)

where described

before

f the file

I

'unit'

for

can

the

end-of-file

is

the

be

statement

record

of

the

first issuing a

for

sequential

of

the

unit

an

external

BACKSPACE

connected

record

is

file.

access

ENDFILE

unit specifier

statement

for direct access

can

be

specifier

and

(section

at

its initial

has

ERR=

executed,

No

data-transfer

BACKSPACE

when

statment

and

(6.2.3).

read

during

'arg-list'

6.2.3).

point,

no

effect.

1030)

the

record

I/O

or

REWIND

ENDFILE

are

'arg-list'

also,

only

subsequent

is

a list

or

if

the

If

an

end-of-file

preceding

statement

statement.

is

issued.

is

a list

those direct access

of

arguments

file is

the

can

of

arguments

records

connected

condition

ENDFILE

be

executed

The

file

appearing

operations.

as

I f a file

ENDFI Examples:

is

preconnected

LE

statement

to

causes

a unit

the

ENDFILE 3 ENDFILE

6.3 Data-Transfer

Once

a file

statement, statements. connected

has

or

Note

to

(3,

ERR = 1040)

been

data

that

a line

printer,

1/0

connected

can

be

the

keyword

nor

6.3.1 READ Statement

The

READ

where

statement

READ

(ctl-list)

[in-list]

READ f [,in-list]

etl-Iist

in-list

reads

data

is a list

is

but

does

file

to

be

created.

Statements

to a unit,

written

does

from

a list

data

into 'PRINT' the

keyword

a specified

of

control

of

the

data

not

yet exist, specifying

in

the

file

can

the file using

does

not

imply

'WRITE'

unit.

Its

information

to

be

read

be

read

the

WRITE

that

imply

possible

specifiers

the

using

an

that

it is

formats

unit

the

or

output

not.

are

in

an

READ

PRINT

file

is

6-10

is a

f

format

specifier in

identifier,

'ctl-list.'

and

is

the

same

as

the

'FMT

=

f'

FORTRAN-SO

Input/Output

The list

of

control

[UNIT [FMT REC

=]

= recno

IOSTAT =

information

unit

f

stname

specifiers is:

ERR = stl END =

6.3.1.1

The control READ

stl

Control

Information

information

list must contain a unit specifier.

statement shown above

List

is

used (that is,

the default unit.

The

'ctl-list'

The following sections 6.3.1.1.1

may

contain, at most, one

through

detail.

6.3.1.1.1

OPEN 'unit'

Unit Specifier.

The

unit specifier has the same interpretation as for the

statement (section 6.2.1.1). In addition, for

may be

an

asterisk.

If

this

processor-determined external unit. The unit specifier may also point

of

a character variable, character array, case, 'ctl-list' must contain a contain a record

number

format

specifier.

U nit specifier Format

Record I/O

Error

specifier

number

specifier

status specifier

specifier

End-of-file specifier

If

the second form

if

of

each

no unit

of

the

is

specified), the unit read

other

specifiers.

6.3.1.1.6 describe control list specifiers in

data-transfer

is

the case, the specifier identifies a particular

to

an

internal file (that is,

or

character

identifier (other

'unit'

array

than

I/O

statements,

may be the name

element name). In this

asterisk)

and

of

may

the

is

not

If

the optional

'UNIT

list.'

Example:

READ

(2)

PNAME, AB, HITS

6.3.1.1.2

ment The

where

• The label

•

Format

is

aformatted

format

[FMT

'f'

is

of

this specifier

=]f

one

of a FORMAT

An

integer variable ASSIGNed the label

• A character

containing a

•

An

asterisk (*) specifying list-directed formatting (section 6.4.4);

•

An

integer, real,

data.

If

the optional 'ctl-list' 'UNIT

=.'

'FMT

and

the first item must be the unit specifier without the optional characters

='

Specifier.

I/O

statement; otherwise, it

of

the following:

array

name, character variable name,

format

specification;

or

logical

='

is

omitted, the unit specifier must be the first item in 'ctl-

If

'ctl-list' contains a

format

is

an

unformatted

specifier, the

I/O

READ

statement.

is

statement in the same

array

containing a

is

omitted, the

format

program

of a FORMAT

format

unit as the READ;

statement;

or

character expression

specification as Hollerith

specifier must be the second item in

state-

6-11

Input/Output

FORTRAN-SO

I f the asterisk

If

the unit specifier

not

be

an

asterisk.

Examples:

25 FORMAT

25

6.3.1.1.3 list' if

Record

and

only if the file to be read

REC = recno

'recno'

where pression

is

the

option

is

selected, 'ctl-list'

is

an

internal file, the

READ (2,25) PNAME,

...

READ 25, PNAM FpRMAT

...

ASSIGN 25 TO

E,

AS, HITS

INFMT READ (2,INFMT) PNAME FORMAT

READ

Number

is

an

integer expression whose value

number

...

(2,

*) PNAM E

Specifier.

of

the record to be read.

The

is

connected for direct access.

must

not

format

AS, HITS

record

number

include a record

number

specifier must be present,

is

positive.

specifier

is

included in 'ctl-

It

has the

The

value

specifier.

but

can-

format

of

this ex-

Examples:

READ READ READ

6.3.1.1.4 essentially interpreted as of

data-transfer

negative value

6.3.1.1.5 OPEN

6.3.1.1.6

END =

where

(2,

REC =

(2,

REC =

(2,

REC = K +

Input/Output

at

Error

20) K)

1)

Status

it

was for the

110

statements, however, the variable

end-oF-file.

Specifier.

The

error

statement (section 6.2.1.3).

End-Of-File Specifier.

stl

'stl'

is

the label

of

an

executable statement in the same

Specifier.

The

OPEN

specifier has the same interpretation as for the

The

format

statement containing this specifier.

When

the end-of-file

statement

terminates,

is

detected during a read

'stname'

is

assigned a negative value (6.3.1.1.4),

tion continues with the statement specified by

110

status specifier,

10STAT,

statement (section 6.2.1.2). In the case

'stname'

of

the end-of-file specifier

operation,

is

program

execution

also assigned a

is

unit as the

of

the

READ

and

execu-

END.

is

6-12

If

END

is

Example:

READ

specified, the file must be connected for sequential access.

(2,25,IOSTAT=STFLG,ERR=

1200,END=860)

A,S,C

FORTRAN-SO

Input/Output

6.3.1.2

The read. name. ordering sequence. The the input list.

Input

List

list 'in-list' in the

An

item in

If

an

array

input

name

READ

statement identifies the items whose values are

list must be a variable name,

is

listed, the entire

name

of

an

assumed-size

6.3.1.3 Implied-DO List

An

implied-DO list embedded in the be used for input list read without specifying each individual implied-DO list

(in-list, var =

where

'var,

el, (section 4.2.2) may also contain additional

For

READ

Example:

statements, the

CREAD

20

array

elements.

is

el,

e2, e3)

e2, and e3' have the same interpretation as for the

and

'in-list'

ONLY

DIMENSION TABLE(60) READ FORMAT

is

a list

implied-DO lists.

DO

THE ODD ELEMENTS IN ARRAY 'TABLE'

(2,

20) ...

READ

For

of

input

variable

(TABLE(N), N =

array

name,

or

array

element

array

is

read in

dummy

statement allows a range

example,

array

items as described above. The list 'in-list'

'var'

half

element to be read. The

must not

appear

1,59,2)

normal

array

must

the items in

as

an

array

element

not

appear

of

subscripts

an

array

format

DO

statement

item in 'in-list.'

to

be

in

to

can be

of

the

6.3.2 WRITE

The

WRITE

statement

where

The control and 'out-list' ment, including the

Like input list items, ray element name. An pression involving

Examples:

is

WRITE

ctl-list

out-list

following subsections) except

is

120 FORMAT

C WRITE DOUBLE COLUMN PRINTOUT

C EACH ARRAY

120 FORMAT

Statement

statement

(ctl-list)

is

information

defined in the same

WRITE (6,120) PNAME, AVG

WRITE (6,120,IOSTAT= ERRFLG, ERR=2000)

&PNAME +

DIMENSION PNAME(25), AVG(25)

WRITE (6,120)(PNAME(K), AVG(K),

outputs

[out-list]

a list

of

is

a list

of

implied-DO

an

output

operators

data

control the

list

...

(1X,

information

data

to be written

is

the

same as for the that

manner

option

list item

or

enclosed in parentheses.

I,

AVG + I

A,

5X, F4.3)

to

a specified unit.

specifiers

no

END

specifer

as the 'in-list'

(sections 6.3.1.2

may

be a variable name,

may

also be

The

format

READ

OF

statement (section 6.3.1.1

is

allowed. The

portion

and

an

expression, including

FIRST ITEMS

K=1,

of

6.3.1.3).

10)

of

the

array

the

output

READ

name,

OF

WRITE

list

state-

or

ar-

an

ex-

6-13

Input/Output

6.3.3 PRINT Statement

The

PRINT

format

PRI

where

NT f

statement

[,out-list]

outputs

formatted

data

to

the default write unit.

It

has the

FORTRAN-SO

f

out-list

that

Note printer

The

6.3.1.1.2). The list 'out-list'

6.3.2).

Examples:

or

format

120 FORMAT

is a format

is

a list

the keyword other

print device.

specifier

PRINT 120, PNAME, AVG

ASSIG N 120 TO OUTFMT

PRINT OUTFMT,

identifier

of

the

data

to be written

'PRINT'

'f'

has the same meaning as for the

does not necessarily imply the default unit

is

defined for

...

PRINT

PNAME, AVG

exactly as it

READ

is

statement (section

for

WRITE

6.4 Formatted And Unformatted Data Transfer

In the description formatted specifier are MATTED' perty

or

is

confirmed by the presence

of

the

OPEN

unformatted

'UNFORMATTED'

if the access method

I/O

statement

(section 6.2.1.7). The defaults for the formatting

if

the file

is

sequential. The formatted

or

is

absence

we

saw

that

a file can be connected for

connected for direct access

or

unformatted

of

the

format

specifier

is

(section

and

a line

'FOR-

pro-

6-14

[FMT

=]

f

in READ, statement

WRITE,

is

normally used with a

or

PRINT

statements. in the case

FORMAT

6.4.1 Unformatted Data Transfer

The unit specified in a an

external unit.

the connected file and items in the

The

number record. The type responding

On

output,

do

not

of

input

if the file

fill the record, the remainder

data-transfer

Data

is

transferred without editing between the current record

items in an input list must

of

each value in the record must agree with the type

list item. The item

is

connected for direct access

statement involving

110

list. Exactly one record

and

its value must also agree in length.

of

the record

of

statement.

unformatted

not

exceed the number

and

the values in the

is

undefined.

formatted

is

read

of

110,

the

data

must be

or

written.

values in the

of

the cor-

output

I/O

of

list

FORTRAN-80

Input/Output

6.4.2 Formatted Data Transfer

During the specifications can be given:

•

• As values

• As Hollerith values assigned I f the

array referenced tions. may the or

6.4.2.1

If

the record ing

When

formatted

110

list.

In

FORMAT

expressions;

format

name,

must

The

follow the right parenthesis

format

logical

specification itself.

array.

data

transfer,

The

editing

statements;

of

specifier (section 6.3.1.1.2) in a

character

contain

format

is

character

variable

a valid

specification

Printing Formatted Records

a

formatted

characters

this convention

is

not

record

printed.

of

the record

is

printed

It

is

specified,

directed by some kind

The

is

used instead to indicate vertical spacing.

are

data

is

transferred

arrays,

to

format is that

on

then printed beginning

character

integer,

name,

described below (section 6.4.3).

same applies

some external listing device, the first

the

real,

or

other

specification in its leftmost

ends the

first

character

with editing between the file

of

variables,

or

logical

formatted

character

format

specification with no effect

to

Hollerith

is

format

at

interpreted

specification.

arrays.

110

statement

expression, the value

data

the left margin.

Format

or

other

in an integer, real,

as follows:

character

is a character

character

Character

character

The

remain-

and

posi-

data

on

of

Character

Blank

o

I

+

This-interpretation is requested by (section 6.2.1.10).

preted

as vertical spacing

6.4.2.2

The classified as either repeatable

Both the

right.

is control repeated, say five times, it

If

the left parenthesis tion ing the first pass time

Format Control

edit descriptors

format

One

item in the

no

corresponding

communicates

a

format

contains

specification list ends

format

If

'car'

is specified as

information

that

make

up

or

nonrepeatable.

specification list

110

list

corresponds

110

list item for

directly with the

corresponds

matching

nested parentheses). Repeat specifications have the same effect as

through

reversion occurs.

the

format

Vertical Spacing

One

line

Two

lines

Ski p

to

advance

the

CARRIAGE

'FORTRAN"

and

is

not

the

format

and

its

corresponding

to

nonrepeatable

110

record.

to

five consecutive

before

rightmost

the

I/O

right parenthesis

specification list. A new

specifier

printed.

specification list (section 6.4.3)

each

repeatable

list, it reverts to its beginning

of

the

OPEN

the

first

character

110

list

are

edit descriptor.

edit descriptors,

If

a repeatable edit descriptor

I/O

list items.

if

the

format

record

statement

scanned

and

specifica-

is

begun each

is

inter-

left

There

format

(or

dur-

are

to

is

to

6-15

Input/Output

6.4.3 FORMAT Statement

The

form

of

the

where

FORMAT

stl

FORMAT ([flist])

statement

FORTRAN-SO

is

stl

JUst

Each

item in

descriptor,

nonzero,

with a 'flist'

can

F5.3 5F5.3

X

3(2X,I5)

Note

that used only must

have

is a 1-5

digit

is a format

'flist'

must

or a parenthesized

unsigned integer

also be prefixed by a

Repeatable Repeatable Nonrepeatable Entire

the

FORMAT

if

the

at

least

110

one

statement

list

is

repeatable

6.4.3. t Edit Descriptors

6.4.3.1.1

of

a letter indicating the type

the

data

lw

Fw.d

Ew.d

Ew.dEe

Lw

A

Aw

Bw

Zw

Repeatable

field

and

Edit

how

it

is

Integer Real Real Real

Logical Variable-length Fixed-length

Binary

Hexadecimal

statement

label

specification list whose items

be a

repeatable

'flist.'

constant

repeat

An

specification.

edit

descriptor

called a

descriptor, a nonrepeatable

'repeat

descriptor descriptor

prefixed with

descriptor

'flist'

prefixed with

with

also

empty.

edit

Descriptors.

of

data

to

be divided.

no

Conversely,

descriptor.

Repeatable

involved

The

repeat

'flist'

specified,

if

and a number

repeatable

descriptor number number number

descriptor descriptor descriptor

descriptor

alphanumeric

descriptor

are

separated

by

commas

edit

is

repeated

specification.'

repeat

specification

'FORMAT ( ),'

the

110

list

by prefixing it

The

specification'

'3'

is

not

empty,

entire

5'

can be

'flist'

edit descriptors generally consist

indicating

the

size

of

edit descriptors are:

descriptor

6-16

where

I, F,

LandA

BandZ

w

d

e

The

I,

F,

and

E serve the

notation.

E,

and

E edit descriptors

same

function

indicate

the

is a nonzero, of

the

entire edited field

is

an

unsigned integer

digits

that

should

is a nonzero,

of

the

exponent

are

on

input;

type

of

data

being edited

number

base

of

unsigned integer

constant

follow

the

unsigned integer

field.

used

to

specify

E allows

data

constant

decimal

constant

110

output

being

edited

representing the width

representing

the

point.

representing

of

integer

real

and

numbers

number

the

real

of

width

data.

in scientific

F

FORTRAN-80

Input/Output

Certain

•

• A decimal

•

The B and Z cdit respectively.

'-1' is

general

On

input, to the setting BLANK

an

For

for the processor

On

output, left.

On

output,

exponent

remarks

leading

sp~cifier

point

E descriptor.

values

has

The

'Ff').

Integer Editing

An

I/O

list item

input

list item ed with integer one

digit.

Examples:

is

data.

20 FORMAT

apply

to

all

blanks

of

the

in the

in

to

are

if

the

more descriptors

internal

matched

defined with integer

The

are

not

non

repeatable

OPEN

input

data

The

input

approximate

right-justified.

number

than

representation

integer

of

'e'

digits, the entire field

specify

with

an

constant

overrides the decimal

'Iw'

PRINT 20, INTNUM

(15)

three

of

these numeric editing descriptors.

significant.

descriptors BN

statement

field

may

the

data's

If

necessary,

characters

data

110

of

the

edit descriptor

data;

the

read

Other

blanks

and

(section 6.2.1.9).

also have

value.

the

exceeds the field width

is

filled with asterisks.

in

binary

data

is

output

must

output

list item

or

written always consists

BZ

point

more

field

and

hexadecimal

(e.g.,

be

must

are

and

location specified by

digits

is

blank-filled

of

treated

the

type integer.

already be defin-

according

the value

than

are

'w,'

notation,

'Z'

format

of

the

needed

on

the

or

The

at least

an

of

READ (2,20) INTNM1, INTNM2,

20 FORMAT (215,14)

<F'

Descriptor Editing

An

I/O

list item

put

to

this descriptor

are

interpreted as the fractional

On

input, nent consisting ly signed integer

Output modified edit descriptor, section 6.4.3.1.2). Leading zeros field would be

Examples:

edited by the

matched

a string

of

a signed integer

constant.

by

an

established scale

blank

with

an

contains

of

digits by the basic

'F'

otherwise.

no

part

descriptor

factor

'Fw.d'

decimal

descriptor

point,

of

the

input

'F'

descriptor

constant

or

is

rounded

(see the description

the letter

READ (2,20) REALN M

20

FORMAT (F5.3) DIMENSION TABLE

PRINT

20,

TABLE

(10)

20 FORMAT 5(F5.3, 2X, F5.3)

CTHE

TABLE

WILL

BE PRINTED OUT IN TWO

INTNM3

must

the

rightmost

value.

can

'E'

to

'd'

fractional digits,

are

not

generated unless the

have a real value.

'd'

digits

of

be followed by

followed by

of

the

an

and

'P'

nonrepeatable

COLUMNS

If

the in-

the string

an

expo-

optional-

may

be

output

6-17

Input/Output

'E'

Descriptor Editing

An

I/O

list item matched with an

variable. The exponent On

output,

the

format

zero is:

'Ew.d'

'e'

has no effect on input

of

the

output

field for a scale factor (section 6.4.3.1.2)

or

'Ew.dEe'

data.

FORTRAN-SO

descriptor must be a real

of

[±]

[0].x1

where

± signifies a plus

exp

Edit Descriptor Absolute value of

Ew.d

Ew.dEe

The sign in the exponent plus. The

'Ew.d'

Decimal normalization

-d<k:$O,

the number

xdex

x2···

P

are the after

rounding

is

a decimal exponent having one

('y'

is

lexpl:$

99<

lexpl:$(10**e)-1

is

always present.

descriptor should

is

controlled by the scale factor

output

will have exactly

or

'd'

most significant digits

a digit):

99

lexpl:$

999

If

not

be used if exp exceeds 999.

significant digits following the decimal point. exactly cant digits

k significant digits

to

the right

of

to

the left

of

the decimal point

the decimal point.

Other

minus

'exp'

the exponent

Iklleading

If

0<

k<

d + 2, the

values

of

the

data's

of

the following forms

Form of

± ±

E ±

is

zero, it

'k'

'exp'

OY1

Y2

Y1

Y2 Y3

Y1

Y2··

is

prefixed by a

(section 6.4.3.1.2).

zeros and

number

and

'd

- k +

k are not legal.

value

·Ye

'd

- Ikl'

will have

1,

signifi-

If

Examples:

30

110

READ (2,30) FORMAT

WRITE (6,110) RLOUT FORMAT

RLNUMB

(E4.2)

(E6.5E6)

Logical Editing

An

I/O

item matched with

The input field includes

an

'Lw'

descriptor must be

an

optional period followed by a

of

type logical.

'T'

(for

TRUE) FALSE). These characters may be followed by additional characters. the logical

The value

constants'

output

of

the internal

.TRUE.'

field consists

data.

and'

of

the letters

.FALSE.'

'T'

or

are acceptable inputs.

'F,'

based on the

TRUE

Examples:

50

80

DIMENSION

READ

WRITE (6,80) TRUTH(1) FORMAT

(3,50) TRUTH(1), TRUTH(4)

FORMAT

TRUTH(4)

(2L6)

(L

1)

or

For

or

'F'

(for

example,

FALSE

6-18

FORTRAN-80

A /phanumeric Editing

An

110

item

matched defined with Hollerith 'w'

characters. Otherwise, the

110

list item.

On

input,

if

the

character

cated

on the string filling

is

the right.

is

blank-filled

done

on

If

the left.

with

an

'A'

or

'Aw'

descriptor

data.

If

the field width

number

string

is

the specified width exceeds the length

on

the right.

of

longer

The

characters

than

same

must

'w'

is

specified, the field consists

in

the

the specified width, the string

is

true

on

output,

have type

field

is

of

the

except

character

the length

character

that

or

of

the

is

trun-

string,

blank

Input/Output

be of

NumlJt!f Bast!

A:nil1~lltit7ttiirp~t¢~~9~H~

()rht:x~d~cirnaldjg~ts;re~.l)ectiv7~Y'

either a sjgu

I··f··.O.n~."

•.

()(

number~~se tobe~ur7

exampletif~Fisa,~ingle-byteinn~ger

Editing

.~.

'Bw'

or':fw~?~~pr~~~Rt~tis~cpn~~sts

..

·10

,.p~rticula~,$tlpljaninput

ora

letter indicatingthebas¢.

••.•

tne~~

.•..•

4~s;ct.iptors

in~ic~ted.Leadiflk1Ter()$ar7

th~rear~asrn~ny.

•..

·i$

•.•••

sP~.~i.fi.eq.·

digit~·pres~:n!~~~t7n.~~pedto

.·fo~

....

9.HtJrM.t#

sup~Jie~~~9:utptltaI1Qacceptedon

••. • ••

;.~.'.·'·~nat~Pt¢rs.·

represent'

wn¢se

valuei~'4,'

itisOlltput

onlyofbi~ary

·fi~ld

cannot contain

•••

are

out

the

as

'bbb04

a 'Z5'editdesctiptor.

The

flum ber· hase

6.4.3.1.2

'hlh2···hn'

nHhl

nX

/

Nonrepeatable Edit Descriptors.

h2

...

kP

edit

descriptors

h

n

can

be

specif\edfof tlataof any type.

The

nonrepeatable

Literal string descriptor

Hollerith string descriptor

Record position

Record

Scale

termination

factor

control

descriptor

edit descriptors are:

pu

t·

•.

io

·t~e

input,

data . For

'under

BN

BZ

$

where

apostrophe

of

kind

Apostrophe

The enclosed in the

editing

h

n

k

apostrophe

character

Blank descriptor

Alternate

C),

H,

X,

slash

(I),

P,

and

is

any

character

is a nonzero, is

an

optionally-signed integer

Editing

edit descriptor

apostrophes

field, show it as two consecutive

to

representable

unsigned, integer

can

be used only for

be written

out

record

BN, BZ,

on

constant

literally.

apostrophes.

termination

and

the

the processor

output.

To

indicate

desc:riptor

dollar

sign ($) indicate the

representing a scale factor

It

causes the

an

apostrophe

characters

within

6-19

Input/Output

The

width

of

the field

is

the length

of

the character string.

Example:

WRITE (7,100) ITSTNO

100

'H'

Descriptor Editing

The

Hollerith field descriptor

apostrophe

FORMAT

editing. Like

('THIS IS THE TEST NUMBER', 2X,

is

an

alternate

apostrophe

editing, it

way

is descriptor causes the In' characters following the embedded blanks).

on is

an

likely

Because the Hollerith field descriptor relies duce the correct

output,

apostrophe

editing

method. Example:

WRITE (7,100) ITSTNO

100

FORMAT

(1

H1, 19HTHIS IS TEST NUMBER, 2X,

C FIRST H DESCRIPTOR CAUSES SKIP TO NEW PAGE

12)

to

perform

used only for

'H'

to

be written

the same

output.

accurate character

to

be less

error

operation

out

(including

count

prone

12)

The

to pro-

than

FORTRAN-SO

as

'nH'

this

'X'

Descriptor Editing

The'

nX'

descriptor indicates

is

the

character

is

to insert In' blanks

"n' positions from the

Example:

WRITE (7,100) ITSTNO

100

FORMAT

C FIRST X DESCRIPTOR CAUSES SINGLE SPACING BY

C INSERTING A CTHE

RECORD

Slash Editing

The

slash

(I)

edit descriptor acts as

mainder

of

On edit descriptor to provide

The before

of

the

current

a record, the entire record

output,·

comma

the current record

may

blank

lines

that

normally separates

or

after

a slash.

also be used

that

into

the

the next

output

current

record.

character

record position. On

(1X, 'THIS IS TEST NUMBER',

record

on

printed

BLANK

AS THE FIRST CHARACTER

an

end-of-record indicator.

is

skipped

is

skipped.

is

terminated

to

write

or,

if

the file

and

an

empty record, which

output.

format

specification list items

transferred

2X,12)

is

positioned

a new record

to

or

from a record

output,

the effect

OF

On

input, the re-

at

the beginning

is

begun.

is

a convenient way

is

The

not

required

slash

6-20

Example:

WRITE (7,100)

100

CTHIS

FORMAT

SLASH CAUSES

WRITE (7,150)

150 FORMAT

(1

H1, ' PLAYER AVERAGE'!)

BLANK

LINE

FOLLOWING

PNAME, AVG

(1X,

A12,

4X,

F4.3)

HEADING

FORTRAN-80

Input/Output

Scale Factor (P)

A scale factor factor. needed between the

A scale factor been changed by the'

On input, the scale factor has no effect field. number true part reduced by

As multiplier printed in scientific Changing the scale factor to one changes the multiplier range the scale factor generally not desirable for scale factor, the scale factor should probably be respecified as occurrence

The following table and the use the field descriptor used

It

1 PE8.6E2

kP'

descriptor

Otherwise, the effect

equals the internally represented

of

output

of

the quantity

we

saw in the description

Editing

is

established by

is

used with the

'P'

descriptor and

of

zero

is

the'

kP'

is

issued again

with

'F'

to

'k.'

is

useful for very large

of

'F'

editing.

of

number

of

the scale factor would be most applicable. The column headings show

the'

kP'

edit descriptor,

'F'

and

'E'

descriptors

an

immediately following

assumed at the beginning

edit descriptor, the new scale factor remains in effect until

or

until the end

if

there

is

of

the scale factor

number

editing. On

be produced

'F'

to

produce each representation.

output

of

'E'

notation

edited numbers. Following specification

representations illustrates instances where

with

is

multiplied by

descriptor editing, the

with a scale factor

or

very small

where'

to

edit real numbers. No

of

an

110

of

the

110

statement.

an exponent in the

is

that

the externally-represented

multiplied by

'E'

editing, the basic real

'10**k'

to

'E'

edited numbers,

k'

represents the scale

comma

'F'

or

'E.'

statement. Once it has

'F'

or

'E'

'10**k.'

and

output

of

zero

1.0 - 10.0. Changing

'OP' before the next

The same

constant

the exponent

range

is

0.1 to 1.0.

of

a nonzero

'E'

edited

is

of

but

is

editing

is

a

Real

Number

4.32

7255000.0

0.0065

Clearly, clearly,

'7255000'

'BN'

These two edit descriptors are used for They can be used

If

zero. Until the

specifier in the blanks.

Example:

Dollyt

'F'

or

and

'BZ'

'BN'

is

specified, blanks are ignored except

If

'BZ'

50

Sigfjl$ijitrrrg·

descriptor editing

'0.0065.'

Editing

is

specified, blanks are regarded as zeros.

or

FS.2

4.32

******

0.01

is

preferable for simple numbers like

is

inadequate for very large

to

specify the interpretation

'BN'

descriptor

OPEN

statement (section 6.2.1.9) determines the interpretation

READ (2,50) INTNUM, FPNUM FORMAT

(BN,

is

specified (or,

15,

5X, F7.4)

E10.5

0.43200E +

O.

72550E + 07

0.65000E-02

input

only

of

blanks,

that

if

1 PE10.4

01

and

a field

neither

4.3200E+00

7.2550E+06

6.5000E-03

'4.32.'

or

very small numbers like

affect only numeric editing.

other

than

leading blanks.

of

all blanks

is

specified), the BLANK

is

Just

as

treated as

of

Th¢

..•

~oll.~f

•••.

$~~·~

••••••

~.~.).···.·e~l~

c()nsoletYFl1l~l\l~LJtlea~~s.~·~l\tYF~inalfnlrs~r~tr~~~~§~~~~n~m,tn~~i~tel~fCl)~19\Ving

the

I/Odataju~t

.••.••

(J~~pti~t?f

pr()ces$ed;·rathe:rthant>egil1nill,gan1tw~ine~Ifthe·f{)r.i)lat·c()l1tTlpl

.. · •••

~.s·

....

·.~.~~~.

·.p~i~~t~l,~

•••.

f~r~~~~r~~ti,~~

.•

ilq

...

t~r?Mgp.·L.a

6-21

Input/Output

FORTRAN-SO

scanner

trol

donar

encounters a dollarsign

terminates

sign .edit

without

descriptor

at

positioning

has

no

effect

the

end

on

Example:

PAINT 25, PNAME

25 FORMAT (A20, $)

6.4.4 List·Directed Formatting

List-directed

specifier

free terminal. No

the

A list-formatted file consists

either a

of a data-transfer

form

input

programmer.

constant,

I

formatting

and

output,

FORMAT

statement

a null value, or a

is

indicated by specifying

110

statement's

which

is

especially helpful

is

required as all necessary

of

a string

constant

specification in the form:

r*c r*

The

first form

cond

is characters separator

is

equivalent

between value

in a record.

to

'r'

successive

to

'r'

null values. A null value

The

separators

null value

or

is

not

of a format

file

to

the beginning

direct~accessfil¢s.

control

specification list,

of

the

asterisk (*) in the

list (section 6.3.1.1.2).

if

the

110

device

formatting

of

values

and

value

separators.

or null value prefixed by a repeat

appearances

no

characters

produced

of

the

constant

can

also be specified as

preceding the first value

by list-directed

output.

format record.

It

is

a console

is

done

Each

'c;'

con;.

The

format

allows

for

value

the se-

no

is

Value

separators following the last put.

When

put

statement

tional items, they

Any sequence character

constant.

6.4.4.1 List-Directed

Input

forms acceptable list-directed may not be

An input list item unless a decimal

can

be a

constant

encountered

after

assignment

are

effectively assigned null values.

of

blanks

An

formatting

embedded

of

type real

point

comma,

in a record.

during list-directed

is

treated as a single

end-of-record

Input

to

format

with a few exceptions. Blanks

in

constants

appears

must not include either slashes An

input

list item

of

characters enclosed in

the end

of

one mally has the effect the

character

and

are

truncated

A null value has

previous value

was

encountered.

of

record

of a blank

constant

on

no

effect

or

remains

type

character apostrophes.

to

the beginning

in this case.

the right

on a corresponding

undefined,

slash,

or

one

or

more

blanks

The

slash

separator

input,

it terminates execution

of

the previous value.

blank

If

except when

has the same effect as a

specifications for a given type

are

(except

is

assumed

within the field. An

or

commas

in list-directed

Character

if

their length exceeds the width

character

to

have

in its input

consists, in its

Character

of

the next.

formatting, a blank

constants

input

depending

constants).

no

fractional

input

form.

input constants

Though

are

list item.

on

its

is

not

the

110

blank.

never

list item

form,

can be

the

transferred

of

status

between

constants

produced

of

list

contains

it

appears

are

acceptable for

treated

part

as zeros,

in

its

input

of

type logical

of a nonempty

continued

end

of

a record nor-

is

not inserted

left-justified

the

input

list item.

The

item retains its

before

the null value

or

on

out-

the in-

addi-

within a

and

form

string

from

into

6-22

FORTRAN-80

Input/Output

6.4.4.2

The responding as the end and

Integer reasonable value editing, depending of constants

Output

List-Directed Output

form

of

the values

output

of

a record does not fall within a

blanks are not embedded within constants.

'1

P'

output

is

constants

used. Logical

are

output

records are single spaced.

produced

by list-directed

list items. The processor separates records as necessary, so long

are

produced

of

'w.'

Real constants are

on

the

magnitude

constants

simply as

are

character

output

constant

with the effect

produced

of

the value.

'T'

for

TRUE

strings without

agrees in type with their cor-

(except a

of

with the effect

Where

and

'F'

surrounding

character

'Iw'

formatting,

of

reasonable, a scale factor

for

FALSE.

apostrophes.

constant)

for

some

'F'

or

'E'

Character

6-23

CHAPTER 7

PROGRAMMING GUIDELINES

This chapter techniques. mer.

For

ming, a

The first section ming in any language. The second section lists suggestions for programming in FORTRAN

7.1

Program Development

The

recommended method. initially,

final code.

going

to

The

first step in this

7.1.1

Before considering any actual programming, pletely the problems involved.

plugs

on the new plugs by the possibility put

a trace

easier,

is

by

no

means intended as an exhaustive discussion

It

is

intended simply as a guideline, primarily for the novice program-

those who wish

number

the next level

of

specifically.

Essentially, this means defining a and

then working down

At

each level, debugging

to

go

more

deeply into the science

references are listed in the bibliography

this chapter deals with general guidelines applicable to program-

approach

of

approach

to

program

through

refinement.

is a thorough

is

Problem Definition

For

example, a person

his car. This

of

and

limit the chance

is

fine,

but

hand

first

and

stripping threads.

'anti-seize'

compound of

breaking a plug.

he will be better

using a plug wrench only for final tightening reduces

And

of

FORTRAN

at

the end

development

program

a series

performed

definition

one

prepared

he will be better advised still

on

the threads first

is

the so-called

in the broadest possible terms

of

increasingly detailed steps

to

whatever extent possible before

of

the programming task.

must

understand

may

have

if

he knows

to

of

programming

program-

of

the chapter.

'top-down'

clearly

to

change the spark

make

and

that

inserting

if

he

is

later removal

com-

told

to

Similarly, knowing

it to print out only employee names (federal, social security, state, city,

for the

show the hours worked, splitting

tion pay? Should it show accrued sick used to display all this

While the he at least needs what form,

company's

progammer

and

that a program

stock plan, pension plan, credit union,

information?

need

not

to

know in detail

exactly what

must print

ad

infinitum)? Should it show

out

overtime, shift differential, holiday,

or

state the specific algorithms

what

output

the

out

payroll statistics

and

net pay? Is it

vacation leave?

input

the

program

is expected to produce.

to

And

program

7.1.2 Program Documentation

Every

program

At

the earliest stage preliminary functional specification should be reviewed by will gaps be filled in, of

programming, for coworkers to familiarize themselves with each

and

for developing a feeling

should have

of

program

one's

but

this review creates

good

documentionfrom

development,

of

programming

of

teamwork within a

the beginning!

documentation

the

program.

peers for constructive criticism. Not only

an

environment for exchanging theories

programming

Ideally, this specification

is

insufficient.

show taxes withheld

other

deductions

or

whatever? Should it

or

what

format

to

be used

will be receiving

would normally be a

group.

at

other's

should be

this stage,

and

projects,

Is

vaca-

in

7-1

Programming Guidelines

FORTRAN-80

As the development become correspondingly detailed.

program

final

ingful mnemonic names for symbols, lines to improve cessor

memory

These comments could include

of a program

becomes

The

ultimate

more

detailed, the

documentation

code, which should include numerous useful

and

make

good use

program

size, each

readability.

program

more

If

the

program

is

not

unit could be prefaced with a

than

a description

of

documentation

comments, of

severely limited by pro-

the tion; for example, it might say who coded this unit originally latest changes.

7.1.3 Refining The Problem Definition

Once the problem level in the series being increasingly detailed.

For

the sake

defined as follows:

• The

input employee. This record contains all rates, deductions, etc.

• The

program field), the hours worked (HRS), the gross pay PAY).

After the last record has been printed, the

• showing the total (gross) pay disbursed.

is

defined in detail, a series

of

example, let's assume the payroll statistics

will be a

formatted

file on diskette containing a record for each

of

refinements

information

is

to

print

out

(on the line printer) only the employee's

(GRPA

program

number

of

employees, the total hours worked,

printout

related to hours worked, pay

Y),

is

should

is,

of

course, the

have mean-

blanks in statement

comment

program

and

who

is

begun, with each

unit's

made

block.

func-

task has been

name

(EMP

and

net pay (NET-

to

print a summary

and

the

total

the

The

first two levels

Print

out payroll statistics

Initialize variables.

Open

files.

Read

input

record.

If

(last record) then

Print

totals

Close files

Else

Print

individual statistics Update Go

Note

in level two words here beginning

totals

to

'read

and

there

to

take shape,

FORTRAN.code.

of

program

development might look like:

Level One

on

line printer.

Level

Two

record'

that

statement

we are still using essentially English sentences with a few

that

begin

to

but

at

look like

this

point

FORTRAN.

we're

more

The

program

concerned with logic

structure

than

is

7-2

FORTRAN-SO

Programming Guidelines

The next pass statements for English sentences.

C INITIALIZE VARIABLES NEEDED FOR TOTALS

C OPEN INPUT/OUTPUT FILES,

C PRINT TOTALS IF NO MORE RECORDS

C OTHERWISE PRINT

is a more

DATA TOTEMP, TOTHRS, TOTPAY / 3*0.0/

OPEN

OPEN 10 20

40

60 FORMAT (flist)

READ (unit, FORMAT

I

F (no

WRITE (unit,

FORMAT (flist) CLOSE CLOSE

ELSE WRITE

TOTEMP TOTH TOTPAY = TOTPAY

formal description

(input (output

more

file)

20)

(flist)

records) TH

40)

(input (output

EMPLOYEE DATA

(unit, 60) EMP, HRS, GRPAY, NETPAY

= TOTEMP + 1

RS

= TOTH

of

Level Three

READ EMPLOYEE RECORD

file)

EMP, HRS, GRPAY, NETPAY

EN

TOTEM

file)

RS

+ H

+GRPAY

P,

TOTH

RS

level two, introducing

RS,

TOTPA Y

AND

UPDATE TOTALS

FORTRAN

C READ NEXT EMPLOYEE RECORD

number

A

complete the

accuracy

only

at

GOTO ENDIF END

of

details remain to be specified,

and

can be debugged to the extent

of

the

end

concern ourselves with such details as

10

program's

logic, then

7.1.4 Final Coding

At

the level debugging the program

•

Take programming can execution paths. These

• Avoid tricky programming.

•

Concentrate produced as a last step.

•

Concentrate Syntax details initial

• Again, use

blanks

of

detailed code, a

program,

later.

advantage

be interspersed

effort

to

improve

of

language.

initially

especially

can

in this

area

comment

program

number

or

to

simplify the task

the built-in debugging aids available in the particular

For

example,

throughout

can

be removed in the final version

Code

on

making the

on

getting the statement syntax correct the first time.

be particularly annoying

can save a lot

lines frequently. Use meaningful labels for variables. Use

listing readability.

but

each

that

it

is

110

interfaces

of

steps can be

of

another

PAUSE,

initial

conservatively!

FORTRAN

program

of

work. Beautiful

to

grief in the long

of

these levels

complete. Thus

and

basic calculations,

format

STOP,

FORTRAN

specification.

taken

to

programmer

and

code to trace

of

run.

simplify writing

WRITE

the

programmers,

is

in some sense

we

can confirm

updating the

statements

program

program.

printouts

and

can be

but

7-3

Programming Guidelines

This

is

checklist. Rereading such a list frequently, like rereading

good

way to refresh

7.2 FORTRAN Coding

just a 'starter'

list. Certainly, any experienced

or

reconfirm

programming

knowledge.

programmer

programming

could

manuals,

add

to

FORTRAN-SO

this

is

a

Section 7.1 lists some general TRAN

specifically,

other

programming

points should be kept in mind.

7.2.1 Functions And Subroutines

The

first point has already been

functions all, they reduce the chance And, grammed

understand.

and

of

error.

most

importantly,

and

subroutines in a

amount

They also save processor

debugged

they break a

and

made

in

program

of

coding

that

also clarify its logical structure, making it easier to

wherever it makes sense

to

program

7.2.2 GO TO Statement

The

GO

TO

statement should be used only when necessary. around ning. No painter would worry by simply shouting whether gram

at

will within a

an

alternative solution

has been overlooked.

program

'go

to

can be a strong

about

painting himself

exit.' When the

that

would improve the logical structure

considerations. When coding in

Chapter

be done, saving time

memory

GO

5,

but

is

worth repeating. Use

to

do

and

by allowing

into

units

that

temptation

into

a corner

TO

seems necessary, consider first

shorter

can

be separately pro-

The

ability to

to

neglect logical plan-

if

he could escape

so. First

reducing the

FOR-

of

programs.

jump

of

the pro-

7-4

7.2.3 Crossing Unit Lines

The

ability

to

divide a

It

also has some potential pitfalls. procedures, gram aware memory!

and

unit!

Take

of

the interaction

program

variables whose values have been

advantage

into

subprograms

Be

careful when using global variables, external

of

the capability provided by

among

all the

program

is a major

computed

common

units

benefit

outside the

memory,

that

reference

of

7.2.4 Computing Variables And Constants

Complex calculations can frequently be simplified by breaking them steps

and

computing

variables are used several times

can

tion time

ly

more

variables

be reduced by using intermediate variables

readable. Like most

can

be abused

intermediate variables. This

after

their value has been

programming

and

requires good judgement.

tools, however, the use

is

particularly

computed.

and

the

Program

program

of

FORTRAN.

current

into

true

intermediate

but

common

several

if

such

execu-

is

general-

pro-

be

FORTRAN-SO Programming Guidelines

When calculating a value for use in a the loop iteration

is

entered. Otherwise, the

of

the loop. Consider the following

C EXAMPLE

OF

RECOMPUTED CONSTANT INTEGER*1 DO

25

R =

R

1,60

X = (2217)*(r* *2)

WRITE (4)

25

C SAME

25

In the first case, the value

CONTINUE

EXAMPLE WITH PRECOMPUTED

INTEGER*1 R PI

=

DO

X = WRITE CONTINUE

X,

2217

25 R = PI*(R**2)

'22/7'

1,60

(4)

X,

would have to be

R

7.2.5 Reminders

We've

already mentioned the use to improve debugging aids available in the and

WRITE

other

debugging tools DEBUG cludes

command

ICE-80

program

or

and

readability

PRINT

statements). In addition, the

that

in Intel's ISIS-II

ICE-85 for the 8080

might be available in his system environment, such as the

of

FORTRAN

DO

loop, be sure the value

program

comments,

and

or

could

compute

short

Examples:

CONSTANT

computed

good

mnemonic names,

understandability. We've also mentioned

language (such as the

programmer

Intel's in-circuit

and

8085 microprocessors.

emulator

is

computed

the

value again for each

60 times!

PAUSE,

should explore

family, which in-

and

before

blanks

STOP,

7.3 References

The

following list suggests material for further reading. Some the books by Ledgard Bear in mind and

FORTRAN

in the new version. Dahl,

Academic Press, New York, 1972. Dijkstra,

0.1.,

Edgar

that

Dijkstra,

the A CM, Vol. 15, No.

Henderson,

BIT

12, 1972. Hilburn,

Software,

Knuth,

P.,

1.L.,

and

Applications, Prentice-Hall, Inc., 1976.

Donald

Algorithms, Addison-Wesley, Reading, Mass., 1968.

Ledgard, Rochelle

Henry

Park,

N.l.,

and

McCracken

all

of

this material was written before

'limitations'

W.,

'GO

and

Snowdon,

and

lulich,

E.,

The

F.,

Programming Pro verbs

1975. .

discussed in these works may

E.W.

10,

and

TO

Statement Considered

Oct. 1972.

R.,

P.M.,

Art

of

Computer

or

the article by Ogdin,

Hoare,

'An

C.A.R.,

Experiment in Structured

Microcomputers/Microprocessors: Hardware,

Programming, Vol.

for

FORTRAN

of

the material, like

is

essentially tutorial.

FORTRAN

no

77

~as

longer be a problem

Structured Programming,

Harmful,'

Communications

Programming,'

1,

Fundamental

Programmers, Hayden,

specified

of

7-5

Programming Guidelines

FORTRAN-SO

McCracken, Daniel

A.,

A Simplified Guide to

Wiley & Sons, New York, 1974.

Mills,

Technical Report,

Harlan

B., Mathematical Foundations

FSC

72-6012, IBM Federal Systems Division, Gaithersburg,

1972.

Ogdin, Carol Wirth,

of

the

ACM,

A.,

Niklaus,

Vol. 14,

'Software

'Program

No.4,

Design

Development by Stepwise Refinement,' Communications

April 1971.

Course,'

EDN,

FORTRAN

for

June

Programming,

Structured Programming,

5, 1977.

John

Md.,

7-6

FORTRAN-80 STATEMENT SUMMARY

A.1

Statement Sequence

The

following

FORTRAN

1.

Comment

after

2.

The program.

SUBROUTINE,

Within a program

FORMAT

3.

4.

IMPLICIT

5. All specification statements must precede all precede

statements.

6.

The

last line

These rules

order

program.

lines can

an

END

statement.

PROGRAM

The

unit

statements

statement

of a program

are

summarized

of

statements

appear

statement,

first

statement

or

BLOCK

that

permits the following statements:

can

appear

must

function statements, which

unit

in Figure 1-2.

and

lines

before

DATA

precede all

or

between statements,

if

used,

must

of a subprogram

statement.

before

the

other

must

be

an

must

END

APPENDIX A

be observed when coding a

but

cannot

be the first

statement.

specification statements.

DATA

must

statement.

statement

must

be a

FUNCTION,

statements, which

precede all executable

appear

of a main

must

A.2 Statement Summary

In the following Ellipses indicate the preceding item can be repeated indefinitely (within length limits).

Category: Executable

Text Reference: Section

ASSIGN

Category: Executable

Types:

Text Reference: Sections

Arithmetic,

Format

stl

Format

summary,

3.3.4

TO

name

Character,

3.3.1,3.3.2,3.3.3

any

format

ASSIGN

Assignment

Logical

item enclosed in

Statement

Assign statement label ' stl' integer variable

Statement

square

Function

'name.'

Function

brackets

is

optional.

statement

to

v = e

Assign value

type

is character.

integer, real,

of

'e'

to

'v,'

logical,

where

'v'

or

A-l

FORTRAN-80

Statement

Summary

FORTRAN-SO

Category: Executable

Text Reference: Section 6.2.3

Formats

BACKSPACE

unit (arg-list)

Notes: 'unit'

is

unit specifier

is

'arg-list'

following argument list:

[UNIT IOSTAT =

ERR

=

=]

stl

unit

stname

Category: Nonexecutable Text Reference: Sections

3.4.3,3.4.4

BACKSP

ACE

Statement

Position file connected before preceding record.

unit specifier

110

status specifier

error specifier

BLOCK DATA Statement

Function

to

'unit'

Format

BLOCK

DATA

[name] Identify start

subprogram; program

CALL

Statement Category: Executable

Text Reference: Section 5.2.7

}'ormat

CALL

sub [([a[,a]

...

])]

Call argument(s)

CHARACTER Category: Nonexecutable, specification, type Text Reference: Section 3.1.4

'name.'

subroutine

Statement

Function

of

BLOCK

DATA

optionally name sub-

Function

'sub'

with actual

'a.'

A-2

CHARACTER

Format

[*len['llname[,name]

Specify

...

character type variable

Function

name

and

length

or

array.

for

FORTRAN-SO

FORTRAN-80 Statement Summary

Category:

Text

Notes: 'close-list'

Category:

Text Reference: Section 1.1.1

Executable

Reference:' Section

Format

CLOSE

(close-list)

can

be following fields:

[UNIT

=]

IOSTAT =

ERR

=

stl

STATUS =

Nonexecutable

Format

6.2.2

unit

stname

stat

CLOSE

Comment

Statement

Function

file

Close described by 'close-list.'

unit

specifier

110

status

error

specifier

file disposition specifier

Line

(disconnect

specifier

Function

unit)

'c'

or

'*'

in

column

character

Category:

Text Reference: Section

COMMON

Category:

Text Reference: 4.2.3

in

columns

Nonexecutable,

Format

[/[cbJ!]

Executable

Format

1;

any

ASCII

2-72

COMMON

specification

3.4.2

nlist[[,]/[cbJ!nlist]

CONTINUE

Program

Statement

Name

common

Statement

documentation

Function

and

define

block(s)

Function

'cb.'

contents

of

CONTINUE

No

effect unless this statement depends

of a DO

on

DO

loop;

variable.

terminal

is

then

action

A-3

f'ORTRAN-80

Statement

Summary

FORTRAN-SO

DATA

Category: Nonexecutable

Text Reference: Section 3.3.5

Format

DATA

nlist/

clist/[[,)

nlist/

elist/) ...

DIMENSION

Category: Nonexecutable, specification

Text Reference: Section 3.2.1

Format

DIMENSION

a(d) [,a(d))

...

DO

Statement

Assign 'nlist.'

Statement

Name dimension(s)

Statement

Function

values in 'elist'

Function

array(s)

'a'

'd.'

to items

and

in

define

Category: Executable Text Reference: Section 4.2.2

Format

DO stl [,) var =

el,

e2 [,e3)

Notes:

stl label var el

of

last (executable) statement in

DO

variable

initial loop index value e2 loop termination value e3

loop

increment/

Category: Executable, block

Text Reference: Section 4.1.9

Format

decrement

ELSE

IF

Define set

up

loop counters.

amount

Statement

Function

beginning

DO

loop

Function

of

DO

loop

and

A-4

ELSE

Continue

execution; alternate execution ELSE

IF.

path

provides

from

IF

or

FORTRAN-80

ELSE

IF

Statement

FORTRAN-80 Statement Summary

Category: Executable, Block

Text Reference: 'Section 4.1.8

Format

ELSE

IF

(exp)

THEN

Category: Executable

Text Reference: 4.3.1

Format

END

IF

END

Continue 'exp'

Statement

Terminate from

subprogram;

gram

unit.

IF

Statement

is

TRUE

main

Function

execution

Function

program;

mark

if

end

of

expression

return

pro-

Category: Executable, block IF

Text Reference: Section

Format

ENDIF

Category: Executable

Text Reference: Section 6.2.5

Formats

ENDFILE ENDFILE

Notes:

'unit'

is

unit specifier

'arg-list'

is

following

4.1.10

unit (arg-list)

argument

Mark execution.

ENDFILE

list:

Function

end

of

IF

block;

Statement

Function

Write end-of-file connected

to

'unit.'

continue

record

on

file

[UNIT lOST ERR

=]

A T =

= stl

unit

stname

unit specifier

110

status

specifier

error

specifier

A-5

FORTRAN-80

Statement

Summary

FORTRAN-80

EQUIVALENCE

Category: Nonexecutable, specification

Text Reference: Section 3.4.1

Format

EQUIV

ALENCE

(nlist) [,(nlist)]

...

EXTERNAL

Category: Nonexecutable, specification

Text

Reference: Section 5.2.6

Format

EXTERNAL

proc

[,proc

] ...

Statement

Allow entities in same storage area.

Statement

Function

Allows name

procedure

to

be used as an

argument.

Function

'nlist'

of

external!

to

share

dummy

the

actual

Category: Nonexecutable

Text Reference: Section 6.4.3

Format

stl

FORMAT

([flist])

Notes:

'flist'

includes the following repeatable

Repeatable

Iw Fw.d Ew.d Ew.dEe

Lw A Aw

Bw Zw

integer real no. real no. real &

exponent logical alphanumeric alphanumeric binary hexadecimal

FORMAT

and

Statement

Specify

format

data.

nonrepeatable

'string' nHstring nX

/

kP BN BZ

$

Function

of

formatted

edit descriptors.

Nonrepeatable

literal Hollerith record position record

termination

scale

factor blank blank alternate

termination

110

record

A-6

FORTRAN-80

Statement

Summary

Category: Nonexecutable

Text Reference: Section 5.2.1

Format

[type]

FUNCTION

func ([d[,d]

Category: Executable

Text Reference: Section

4.1.1,4.1.2,4.1.3

Formats

GOTOs GO

TO

(s[,s]

...

)[,]exp labeled

GO

TO

i [[,](s[,s]

...

)]

Notes:

FUNCTION

...

J)

GO

TO

Statements

Transfer

variable

Statement

Name

the

'func,'

define its type

parameter(s)

control

's'

name

Function

FUNCTION

'd.'

Function

to

ASSIGNED

or 'i'.

subprogram

and

dummy

statement

to

format

First Second Third

branches unconditionally.

format

branches based

branches

must be included in list.

Cagetory: Executable

Text Reference: Sections

Formats

IF

(e)

sl,

s2,

s3

IF

(e) st

IF

(e)

THEN

Notes:

In first

format

'e'

is

an

trol passes to:

sl

ife<O

s2

if

e = 0

s3ife>O

on

value

of

integer expression

unconditionally,

but

IF

Statements

4.1.4,4.1.5,4.1.6,4.1.7

Transfer statement tion(s) based sion

arithmetic

expression

statement

control or

'e'.

and s 1,

label

Function

to

perform

on

the value

s2,

s3

are

'exp.'

corresponding

specified

a

specified ac-

of

expres-

standard

to

'i'

labels; con-

In second format

introduces

expression

format,

'e'

is

TRUE.

statement

IF

block;

'st'

is

executed

statements

if

logical expression

following

IF-THEN

'e'

are

is

TRUE.

executed

if

logical

Third

A-7

FORTRAN-80

Statement

Summary

FORTRAN-SO

Category:

Nonexecutable,

Text Reference: Section

Format

IMPLICIT

Category:

Text

Reference: Section

type

(1[,

1]

Nonexecutable,

Format

INTEGER

[*

len [,]]

IMPLICIT

specification

3.1.5

...

)[,type(1[,I]

INTEGER

specification, type

3.1.2

name

[,name]

...

)]

...

Statement

...

Define names

range

Statement

Define

with length

Function

implicit typing

whose

first letter

'1,1.'

Function

'name'

to

be

'len.'

for

is

of

type integer

variable

'1'

or

in

Category:

Nonexecutable,

Text Reference: Section

.

Format

INTRINSIC

Category:

Text

Nonexecutable,

Reference: Section

fune

Format

LOGICAL

[*

len [,]]

INTRINSIC

specification

5.1.2

[,func]

...

LOGICAL

specification, type

3.1.3

name

[,name]

Statement

Function

Allow intrinsic function(s) be used as

actual

Statement

,.'

Function

...

Define

with length

'name'

'len.'

argument(s).

to

be

of

type logical

'fune'

to

A-8

FORTRAN-80

F'ORTRAN-80

Statement

Summary

Category:

Executable

Text Reference: Section 6.2.1

Format

OPEN

(open-list)

Notes:

'open-list'

consists

[UNIT

of

the

following specifiers:

=]

unit IOSTAT = stname ERR

= stl

FILE = fname STATUS = stat ACCESS

= acc

FORM = fmat RECL = reden BLANK = blnk CARRIAGE = car

OPEN

Statement

Open

the

to

unit).

unit

specifier

I/O

status

error

specifier

name

file

status

file

method

access

formatting record blank

specifier

carriage

Function

specified file

specifier

length specifier

control

specifier

(connect

file

Category:

Executable

Text Reference: Section 4.3.1

Format

PAUSE

Category:

Text Reference: Section

[msg]

Executable

6.3.3

Format

PRINT

f [,o'utlist]

PAUSE

PRINT

. -

Statement

Halt

program under 'msg'

control

1-5 digits

is

constant.

Statement

Output

preconnected

specified by

Function

of

Function

items

'f.'

execution;

external

or a character

'outlist'

in

unit

in

resume

signal;

format

to

...

-

..

-.-.--

A-9

f'ORTRAN-80

Statement

Summary

FORTRAN-SO

Category:

Text

Nonexecutable

Reference: Section

Format

PROGRAM

Category:

Text

Executable

Reference: Section 6.3.1

Formats

READ READ

(ctl-list) [in list]

f [,inlist]

Notes:

'ctl-list'

includes the following specifiers:

name

2.1.2

PROGRAM

Name be

READ

Statement

main

first

statement

Input

items in

specified

Function

program

Function

'inlist'

controls.

if

'name;'

used.

as

must

directed

by

Second

format

Category:

Text

Reference:

REAL

Category:

Text

Reference:

[UNIT [FMT

=]

= ] f

REC = recno lOST

A T =

ERR

= stl

END

= stl

is

for

preconnected

Nonexecutablc,

Section

Format

name

[,name]

Executable

Section

unit

stname

specification,

3.1.1

...

RETURN

5.2.4

units;

REAL

unit specifier format record 110 error

specifier

number

status

specificr

end-of-file

'f'

is

the

format

Statement

type

Define

'name'

Statement

specifier

specifier.

Function

to

be

of

type

real.

A-tO

Format

RETURN

Function

Return

from

SUBROUTINE

FUNCTION

subprogram.

or

Intel fortran-80 User Manual

Specifications and Main Features

Frequently Asked Questions

User Manual