Digital PDFs

DEC-08-KFXB-D

April 1969

86 pages

Original

2.5MB

Document:	8K Fortran System Prog Ref 196904
Order Number:	DEC-08-KFXB-D
Revision:
Pages:	86
Original Filename:	http://bitsavers.org/pdf/dec/pdp8/fortran/DEC-08-KFXB-D_8K_Fortran_System_Prog_Ref_196904.pdf

OCR Text

Digital Equipment Corporation
Maynard, Massachusetts

mnmoomo

PROGRAMMER'S REFERENCE MANUAL
PDP-8 FAMILY

DEC-08-KFXB-D

PDP-8 FAMILY
BK FORTRAN SYSTEM
PROGRAMMER'S REFERENCE MANUAL

For additional copies order No. DEC-08-KFXB-D from Program Library,
Digital Equipment Corporation, Maynard, Massachusetts

Price: $1 .75

DIGITAL EQUIPMENT CORPORATION o MAYNARD. MASSACHUSETTS

1st Printing June 1968
2nd Printing (Rev) August 1968
3rd Printing (Rev) Aprii i 969

CONTENTS
Page
CHAPTER 1

SK FORTRAN
1. 1

1-1

Line Format

1. 1. 1

Statement Numbers

1-3

1.1.2

Line Continuation Field

1-3

1. 1 .3

FORTRAN Statements

1-3

1.2

Comments

1-3

1.3

Character Set

1-4
CHAPTER 2
LANGUAGE ELEM EN TS

2.1

2-1

Constants

2. 1. 1

Integer Constants

2-1

2.1.2

Rea I Constants

2-1

2.2

Variables

2-1

2.2. 1

Inteqer Variables

2-2

2.2.2

Real Variables

2-2

2.2.3

Scalar Variables

2-2

2.2.4

Array Variables

2-2

2.3

Expressions

2-3

2.3. 1

Function Calls

2-4

2.3.2

Library Subprograms

2-5

CHAPTER 3
ARITHMETIC STATEMENT
CHAPTER 4
CONTROL STATEMENTS
4. l

GO TO Statement

4-1

4. 1. 1

Unconditional

4-1

4.1.2

Computed

4-1

4.2

IF Statement

4-1

4.3

DO Statement

4-2

4.4

CONTINUE Statement

4-3

4.5

PAUSE Statement

4-3

iii

CONTENTS (Cont)
Page

4.6

STOP Statement

4-3

4.7

END Statement

403
CHAPTER 5
INPUT/OUTPUT STATEMENTS

5. l

Nonexecutab le Statement

5-1

5. l. l

FORMAT Statement

5-1

5.2

Data Transmission Statements

5-5

5.2. l

READ Statement

5-6

5.2.2

WRITE Statement

5-7

5.2.3

Device Designations

5-7

CHAPTER 6
SPECIACATION STATEMENTS

6. l

COMMON Statement

6-1

6.2

DIMENSION Statement

6-1

6.3

EQUIVALENCE Statement

6-1

CHAPTER 7
SUBPROGRAM STATEMENTS

7 .1

7-1

General

7. 1. l

Dummy Identifiers

7-1

7.2

Function Subprograms

7-1

7 .2. l

FUNCTION Statement

7-1

7.2.2

Function Type

7-2

7.3

Subroutine Subprograms

7-2

7 .3. l

SUBROUTINE Statement

7-2

7.3.2

CALL Statement

7-3

7.3.3

RETURN Statement

7-3

CHAPTER 8
OPERATING INSTRUCTIONS

8. l

8-1

Loading the Compiler

8. 1. l

Loading Into Core Memory

8-1

8.1.2

Loading on the Disk

8-1

CONTENTS (Cont)
Page

8 .. 2

Operating the Compiler

8-2

8.3

Loading the SABR Assembler

8-3

8.3. 1

Loading Into Core Memory

8-3

8.3.2

Loading on the Disk

8-3

8.4

Operating the SABR Assembler

8-4

8.4. 1

Method 1

8-4

8.4.2

Method 2

8-6

8.4.3

Method 3

8-6

8.5

The Linking Loader

8-7

8.6

Loading the Linking Loader

8-9

8.7

Operating the Linking Loader

8-9

8.7. 1

Library Organization

8-10

8.8

Executing the FORTRAN Program

8-11

CHAPTER 9
DEMONSTRATION PROGRAM
APPENDIX A
DECIMAL AND OCTAL REPRESENTATIONS OF THE CHARACTER SET
APPENDIX B
STATEMENT SPECIFICATIONS
APPENDIX C
FORMAT SPECIFICATIONS
APPENDIX D
STORAGE ALLOCATION
APPENDIX E
ERROR MESSAGES
APPENDIX F
OPERATING PROCEDURES
APPENDIX G
IMPLEMENTATION NOTES
ILLUSTRATIONS
1-1

1-2

Typical FORTRAN Coding Form

CONTENTS (Cont)
Page
TABLES

2-1

Function Library

2-5

5-1

Numeric Field Codes

5-2

PREFACE
This manual describes a version of FORTRAN II designed specifically for the PDP-8/I, 8/L, 8,
8/S, and 5 computers with at least SK words of core memory and a high-speed reader and punch.

It is assumed that the reader is fa mi liar with the basic concepts of the FORTRAN language.
Severa I excel lent texts are available for a more efementary approach to FORTRAN programming.

Guide to FORTRAN Programming," by Daniel D. McCracken (published by John Wiley and Sons, Inc.)
is recommended.

CHAPTER 1
8K FORTRAN

8K FORTRAN (acronym for FORmuia TRANsiation) is used interchangeabiy to designate
both the 8K FORTRAN language and translator or compiler.
The 8K FORTRAN compiler is a computer program that enables the programmer to express
his problem using English words and mathematical statements similar to the language of mathematics and
acceptable to the computer. The compiler translates the programmer's source program into symbolic
language, and then the symbolic version of the program is translated into relocatable binary code, that
is, machine language, the language of the computer. The relocatable binary code, which is output on
paper tape, is then loaded into the computer for solution of the problem.
The 8K FORTRAN system has the fol lowing features:

Subroutines

Two levels of subscripting

Function subprograms

Input/output supervisors

Relocatable output loaded by the Linking Loader

rI o

COMMON statements

I, F, E, A, X, and H format specification

Arithmetic and trigonometric library subroutines

The 8K FORTRAN system (hereafter referred to as FORTRAN} consists of a one-pass
FORTRAN Compiler, SABR Assembler, Linking Loader, and a library of subprograms (see the appendices}.
This FORTRAN system requires a PDP-8/I, 8/L, 8, 8/S, or 5 computer* with ~t least two
fields of core memory, an AS R33 Teleprinter, and a high-speed paper tape reader and punch.
The appendices contain lists of the FORTRAN character set, statements, specifications,
operating procedures for all phases of the system, error messages, and implementation notes.

1 .1

LINE FORMAT
A line of data in FORTRAN is a string of 72 characters or less designated columns 1 through

72. Each line consists of three fields: statement number field, line continuation field, and statement
field, as shown in Figure 1-1.

*The PDP-5 requires a PDP-8 extended memory control modification.

1-1

CODER

FORTRAN

CODING FORM

C-Comr'f)en t

DATE

PAGE

PROBLEM

________

~-s~:~~el~~ -~
STATEMENT -2
NUMBER 8

FORTRAN STATEMENT

IDENTIFICATION

1 2 3 4 51617 8 91011 121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465666768697071721737475767778798
--~

-- ---· --·-·-- -- ----------··--···--·····----,---------,========

I1tl1-1+\-tP1Ri.qqRAl'1 ,V\{I1L, '1.- 1SH~T, f1N ,A,R1Rf\Y,P1F·--1N1J/\IW+E,R.+?1 11-,NIP1 1~?tS~NQJ/'!-QQ.R+P-J+R.-.·1-+

+ -l -+ I·+
+-+-+-!-+-

1--t---1---t--i-+--f'd1R 1s-1I1 $f l·P+ 1V"\ E+ }~UM B, E,R ,s, ', 1T,H ,E f'J1 1s,q ~T 1 T1H+E, ~"iUM~F1R+st-tt+++++-+---t--+-+++--++-~

~l\?+I,,t----t'f(R+l+I+E+--ll+H+-l;.-+~l4MB+Et-~1l-1~+QRP+E1R+·+- -+++++1-+++-++-+-+++1 I I I I I I I I I +-+-+-+-+---+-

-+-+-+--

+-+ -++-++-++++ -!-+ +--+ ···+ I +···I· t + -t- + I · t · + 1---1 I · + I I t - -!--+---+-++ +·+---+++++- + -+-+-+-+ +- + +--+···+· +-+· t- + ++++-t-r+-+--t-+--+
QJJY\ E, Ns, I,q N 1A (, 4-~Q)._+-+-+++++---+-+--+-++++---1----H--+-- -l-++-+-+-++-+-+-+--t-+-+
+-+++-+--+- I I I I I I
I I I I I
-1-+ ++t+ I I

I I

-t--+--+ t·-+-+---1···+·+ ···! · I +····I
+l+D+-+lf==+l+_,.-N ft

I t 11 I I I I
I

·f

I I I 11
I

I I I I +

I I t I·-+·! \··++·ll··t-t+-+--+---t·t t+t--+·t

+··+···+-·+-

ii t- t 1 + I I · I + I I I I I - I I I I I I I -

-t·

-+ -1·

+-+-··+-+--t--1 --+ +-+- +--+-!-+-!--+·· +---!··-+··-+--+··-+-·+··+-···+····+····

t·- +- I -·+-···+

+--+- t-·-1 -+ t-+··-+----+·-·+-+ ·-t- ·+·--+ t -+--+ 1·····-t---t--t-t--:rl--t--t-1-·-+-·+-l

R+.I:-jAP~_Q_µ+-1+-~)_+--/~,c1),, t\ c1,+I11), , , , , , ,_, +-t -+--+-+-++-1--t--t + + ++, t-+--+--+-+---t--+-+--+ --/ + , + •++-+--+- -+-+-+-+-+ +--+·
~RiMAT1(121E1l~+-+-+-+I I I I 1-1-++++-+--+-++++---·t-+-++-++-++-I I I I I I I I I I I I I I I I
I I I I
-+-~
Q--+~-A=+-4~-t--+--t--t--t---+-++-+-t--+-t--+--+--+--+-+--+--+---+---+------t---t---t--+--+-+--t ·t t···--t··· t····1- t·· +-····t·-t-·-t---t·-t· · t-·+· ·t I ·t- I· I ·t · -t-t--+-+-t-+-1-+---t-·-+·+-t-·
_,Jj 1-1Si=-1Jf-l-t-++++++ +++-l t +-I + I 1
+-+-+-+ --:--+--+--+-+-+-+--I -+--+· f·-·1
111+ l-+++-+l+-1 t-++-+-+t- I-+-+-+-+ +-+I.Ji.(_f.-J+J, ),-A, (,J1+-J+~~+-~_Q+-·1+-~-~t-++
++--+
1--t-+ ++ +-I·-+ -+-t-+-+ +-++-+-+-+-+-+-++-+-+-t-l-t---+---t-+-1-+

I "'l -1

I -+'-t"'-1'

·.=A (I~ ),-+---+--t--t-++-i. + ++-+. ++

(J)=A{J+l)

·-+-+'+-t-+·+-t-+-+··+-+- I

--+L..1+:+1--1+ l1=::+T1E1'4.+-+-+-+-+--+-+--+-+---1

t I

!-+-+-+-+~.....+..:·

I -1

-j

l._1_~ _:l_A__ : 1~
PG-3

+ I

I -+--+-+--+

-··+-t ··t + I

1-··t---1 t-·-t t---+- + +-+-+--+--+-t--t-·+·I t--t· .t .. ·t ... t- t·--t-·--+·-+-·+--·+-t··-+-t--t-+---+-

- -t-+-+-+-+-·-+-+-+-+-+-+----+-+-t--t-+-+--+-+-t--t-+-+--+---+-+-1-+-+-+-+-+-t--t-+-+-+-1--1
I

···I

+ I t + 1····-t--1 t---t-+-+-t +-+·+-+-+-1--++··+++-1 +-+-++--+-+- --1 ···+-·t-+-+·-+--+---·

+f.~JJ.dJl-+1--t--~+9+..lt~t-q +--+ I ·I I I I ! I I I I l I I I I I I I ·-- I I I t I I··+·--·\ t--t-+-+-t---t-+-+-+-+-t--t--t--t-+-+-t-+---t---t--++-+-

+-+-+·t--·+··+-·+-

t8~-~t+-t-+ +-I--+ +t -tt I I t t t I I 1 I I I I I I I I

+-+--+-1----+-+-

M ,4~, ,1,=,1,.,b',...4.
1.-2-++-+ ..LJ~~-j3~+-~+i.~4---~_).-!·-~I (11+)+-i\ .Q_,+l 1,)

E::~

J I = 1-=!-+---+--+-+-+-+-+-+-+--t·-+-+-+-+-+-+-t-+-t-t-·
J-1
I· ··-I I ·· ·l I I · I -t · ·t-·-+ -- '· -t-·+·· I- +-·-+- t--t--t +-+ -+-t-+ -+-+-+-+-+--t---1· --t--t--·t-+-+-·+-+--++-+---+-1--+---+--+---+-1--+---+--+

i-+--+--+-+-t----+-J

-+++

f ·t t + t

t-+-t ++++·t--+-t--t-+--+-+--t--++-++---++++++-+---

I I I I I
f

I t t + I· t -+ I

;- t + +- + + I -+---1 t-t- + t -t t--t- +--+ +--t --t--t-+ i +---+ +-+--+ +- t I

+ + t- t

+ I

-+++-+-+-++++t i : \ ;--t-t-+---+-+-++-+-++-++-+ +-+--t---t--+-+--r-++-+-+~ t -l t I-~ H f H-+-+ + + +t+ ~+~ t-+-+++ +-I I-+ f ·H I tE+Q.RM~J+l+.2-f--146+---~,l+---t-+++-+-++t--t ~--t t I I I + + I + :-+ -t t ++ +-+--+ 1--+ +-+-+---+-+-+-+++-+--+-+-+--+-+-+ +--+-t-++ --+ ~ H-- + -+-+-~N+l:>+--1--+· +++-+-1- +-1--1-+--+ ·It-· 1-1--l l- I I I l I I I I l I I I 1- I t-t t -ti -t--t t---t--1 +-+ t-t 1--t- + t-+-t -t t t----t--11---t--t--+-+ -+---+ ~ -1- t t--

;L9!~c±;o~~;1;;\i;~ ;~\-:; ~2~;;'.;:;;':;i2~2;;;;~~3~ 3~3~ ):i J~J~~;~-3s~~ :i ~2~J~~~~ ~(;~:~~~9~1's~s:~l~!;:;;;1;;;'6;6A 656t,~1;,-;-;;/;-::11f;!;.,~;~·;,,~±~;~
1

;;

DIGITAL EQUIPMENT CORPORATION

MAYNARD, MASSACHUSETTS

Figure l - l. Typical FORTRAN Coding Form

100-12164

1 •1•1

Statement Numbers
Each statement may have a positive integer as a label, which is used to reference that state-

ment elsewhere in the program. A statement number consists of from one to four digits in columns 1
through 5. Statement numbers may be assigned nonsequentially; however, no two statements can have
.he same number. Statement numbers must have a decimal value of 2047 or less.

1 . 1 .2

Line Continuation Field

If a FORTRAN statement is so large that it cannot conveniently fit into one statement field,
additional lines may be used to specify the complete statement. Any line which is not continued, or
the first line of sequence of continued iines, must have a biank or zero in coiumn 6. Continuation lines
must have a character other than blank zero {must be a digit from 1 to 9 if a TAB is used in the label
field) in column 6.

1 .1 .3

FORTRAN Statements

Any FORTRAN statement {listed below) may appear in the statement field {columns 7 through
1
72 or immediately following a tabulation {CTRL/TAB) character ). Each statement must begin on a
separate line. Except for data within a Hollerith field (see Input/Output Statements), spaces are
ignored and may be used freely for appearance purposes.
There are five types of FORTRAN statements:
a.

Arithmetic Statements define calculations to be performed.

Control Statements govern the sequence of execution of statements within a program.

Input/Output Statements direct communication between the program and input/output

Specification Statements describe the form and content of data within the program.

Subprogram Statements define the form and occurrence of subprograms and subroutines.

devices.

Each of the above statements is discussed in separate chapters of this manual.

l .2

COMMENTS
The letter C in column l of a line designates that line as a comment line. A comment has

no effect upon the compilation process but it is listed on the printed output. There is no limit to the
number of comment lines which may appear in a given program.

l A tabulation character is generated by typing CTRL/rAB, that is, holding down the CTRL key while
depressing the TAB key.

1-3

1. 3

-CHARACTER SET*
The following characters are used in the FORTRAN language.
a.

The alphabetic characters:
ABCDEFGH IJ KLMNOPQRSTUVWXYZ

The numeric characters:

0123456789
c.

The specia I characters:

(
)

<
>

blank (space)

*Appendix A lists the octal and decimal representations of the FORTRAN character set.

1-4

CHAPTER 2
LANGUAGE ELEMENTS

The rules for defining constants and variables and for forming expressions are described beiow.

2.1

CONSTANTS
Constants are self-defining numeric values appearing in source statements. Two types of

constants, integer and real, are permitted in a FORTRAN source program.

2. l. 1

Integer Constants
Integer (fixed-point) constants are represented by a digit string of from one to four decimal

digits, written with an optional sign and without a decimal point. An integer constant must fall within
the range -2047 to +2047.
Examples:
47
+47

(+sign is optional)

-2
0434
-0

2. l .2

(leading zeros are ignored)
(same as zero)

Rea I Constants

Real constants are represented by a digit string, an explicit decimal point, an optional sign,
3
and possibly an integer exponent to denote a power of ten (7. 2 x 1o is written 7. 2E+o3). A rea I
constant may consist of any number of digits but oniy the ieftmost eight digits appear in the compiled
33
program. Real constants must fall within the range .14 x 10- to 1.7 x 1038 •
Examples:
+4.50
4.50
-23.09
-3.0E14
2.2

(+ is optional)

(same as -3.0 x 10 14)

VARIABLES
A variable is a quantity whose value may change during execution of a program. Variables

are specified by name and type. The name of a variable consists of one or more alphanumeric characters, the first of which must be alphabetic. Only the first five characters are interpreted as defining
the variable name, the rest are ignored.

2-1

•
The type of variable (integer or real) is determined by the first letter of the variable name.
A first letter of I, J, K, L, M, or N indicates an integer variable, and any other first letter indicates
a real variable. Variables of either type may be either scalar or array variables. A variable is an
array variable if it first appears in a DIMENSION statement.
2.2. l

Integer Variables
Integer variables undergo arithmetic ca I cul at ions with automatic truncation of any fractiona I

part. For example, if the current value of K is 5 and the current value of J is 9, J/K would yield l as
a result.
Integer variables may be converted to real variables by the function FLOAT {see Section
2 .3 .2) or by an arithmetic statement. Integer variables must fall within the range -2048 to +2047.
2.2.2

Real Variables
A variable is a real variable when its name begins with any character other than I, J, K,

L, M, or N. Real variables may be converted to integer variables by the function IFIX (see Section
2 .3 .2) or by an arithmetic statement. Real variables undergo no truncation in arithmetic calculations.
2.2.3

Scalar Variables
A scalar variable, which may be either integer or real, represents a single quantity.

Examples:

A
G2
TOTAL

J
2.2.4

Array Variables
An array variable represents a single element of a one- or two-dimensional array of quanti-

ties. The variable is denoted by the array name followed by a subscript Iist enclosed in parentheses.
The subscript Iist may be any integer expression or two integer expressions separated by a comma. The
expressions may be arithmetic combinations of integer variables and integer constants.

Each expression

represents a subscript, and the values of the expressions determine the referenced array element. For
example, the row vector A. would be represented by the subscripted variable A(I), and the element in
I

the second column of the first row of the matrix A, wou Id be represented by A (1, 2).
Examples:

Y(l)
PORT (K)
A (3* K+2, I)

2-2

The arrays above (Y, PORT, and A) would have to appear in a DIMENSION statement prior
to their first appearance in an executable statement. The DIMENSION statement specifies the number
of elements in the array.
Arrays are stored in increasing storage locations with the fiist subsciipt var;ing most rapidly
and the last subscript varying least rapidly (see Appendix D). For example, the two-dimensional array
B (J,K) is stored in the following order:
B(l, 1),B(2, l), ••• ,B(J, 1),B(l,2),B(2,2), ••• ,B(J,2), ••• ,B(J,K)

2.3

EXPRESSIONS
An expression is a sequence of constants, variabies, and function references separated by

numeric operators and parentheses in accordance with mathematical convention and the rules given below.
Without parentheses, algebraic operations are performed in the fol lowing descending order:
**
*and/
+and -

exponentiation
unary negation
multiplication and division
addition and subtraction
equals or replacement sign

Parentheses are used to change the order of precedence. An operation enclosed in parentheses is performed before its result is used in other operations. In the case of operations of equal precedence, the
calculations are performed from left to right; this is aiso true for exponentiation.
Integers and real numbers may be raised to either integer or real powers. An expression of
the form
A**B

c
• I f,...X\ u11u
........ ,J .............
,..I I
• h • fl
( H
means AB and is real unless both A and B are integers. 1...xponentia1
........ v•Y• .cgant .. m1c ,.og ,x,,
1
e
functions are supplied as subprograms (see Appendix E).

Excluding **(exponentiation), no two numeric operators may appear in sequence unless the
second is a unary plus or minus.
The mode (or type) of an expression may be either integer or real and is determined by its
constituents. Variable modes may not be mixed in an expression with the fol lowing exceptions:
a.

A real variable may be raised to an integer power.
A**2

Mode may be altered by using the functions IFIX and FLOAT.
A*FLOAT(I)

2-3

Any numeric expression may be enclosed in parentheses and be considered a basic element.
Example:
IFIX(X+Y)/2
(ZETA)
(COS(SIN(PI* EM)+X) )
A numeric expression may consist of a single element (constant, variable, or function call).
Example:

2 .71828
Z(N)
TAN(THETA)
Compound numeric expressions may be formed using numeric operators to combine basic
elements.
Example:
X+3.
TOTAL/A
TAN(PI*EM)
Alphabetic expressions preceded by a +or a - sign are also numeric expressions.
Example:

+X
-(ALPHA* BETA)
-SQRT(-GAMMA)
As an example of a typical numeric expression using numeric operators and a function cal I,
the expression for the largest root of the general quadratic equation

-b

+~ b2-4ac
2a

wou Id be coded as
(-B+SQ RT(B**2-4. *A *C))/(2. *A)

2 .3. l

Function Calls
In addition to the basic numeric operators, function ca I ls are provided to foci I itate the eva 1-

uation of functions such as sine, cosine, and square root. A function is a subprogram which acts upon
one or more quantities (arguments) to produce a single quantity called the function value. A function
cal I may be used in place of a variable name in an arithmetic expression.

2-4

Function calls are denoted by the identifier which names the function (e.g., SIN, COS,
etc.), followed by an argument list enclosed in parentheses as shown below.
identifier (argument, argument, ... ,argument)
At least one argument, which may be an expression or an array identifier, must be present (see Section
7.2). A function call is evaluated before the expression in which it is contained.

2.3.2

Library Subprograms
The standard FORTRAN library includes built-in functions, FUNCTION subprograms, and

SUBROUTINE subprograms (see Chapter 7) = Built-in functions are open subroutines, that is, they

are incorporated into the object program each time they are referred to by the source program.
FUNCTION and SUBROUTINE subprograms are closed subroutines; their names derive from the types
of subprogram statements used to define them.

Table 2-1
Function Library
,..

Name

\..Cl I

Absolute Value

ABS or
IABS

Ix I
Ix I

Real
Integer

Float

FLOAT

Conversion from integer to real

Integer

Fix

IFIX

Conversion from real to integer

Real

Remainder

IREM

Integer

Exponentia I

EXP

Remainder of last integer
divide*
x
e

Switch Register

IRDSW

Read console switch reg.

Integer

Natural Logarithm

ALOG

Real

Trigonometric Sine**

SIN

log (x)
e
sine (x)

Trigonometric Cosine**

cos

cos(x)

Real

Tangent**

TAN

tan(x)

Real

Square Root

SQRT

(x) 1/2

Real

Arctangent**

ATAN

arctan(x)

Real

Definition

Argument

Real

*If IREM is called as IREM(I/J), the remainder of I/J will be returned. If the argument of IREM does
not contain a division, the remainder of the last integer division will be returned.
**Trigonometric functions use radians rather than degrees.

2-5

The I RD SW function cal I (Switch Register) takes the decimal equivalence of the octal
integer in the switch register as its result. For example, if the contents of the switch register is
1234 (668 in decimal) when the statement
N=IRDSW(O)
is executed, the switch register is read and its contents becomes the value of N, i.e.,

N=668
The switch register can be set in two ways:
1.

Before executing the FORTRAN program, i.e., after pressing LOAD ADD and before

pressing ST ART.
2.

During execution of the FORTRAN program, i.e., using the PAUSE statement.

CHAPTER 3
ARITHMETIC STATEMENT

One of the key features of FORTRAN is the ease with which arithmetic computations can be
coded. Computations to be performed by FORTRAN are indicated by arithmetic statements, which have
the general form
v=e
where v is a variabie name (subscripted or nonsubscripted), e is an expression, and = is a repiacement
operator. The arithmetic statement causes the FORTRAN object program to evaluate the expression e
and assign the resultant value to the variable v.

Note that = signifies replacement, not equality. Thus,

expressions of the form
A=A+B
and
A=A*B
are quite meaningful and indicate that the value of the v.i:iriable A is to be changed.
Examples:

Y = l. l *Y
P =X**2+3. *X+2.0
X(N) = EN*ZETA* (ALPHA+EM/PI)
The expression value is made to agree in type with the assignment variable before replacement occurs. For example, in the statement
META=W* (ABETA+E)

if META is an integer and the expression is real, the expression value is truncated to cm integer before
assignment to META.

3-1

CHAPTER 4
CONTROL STATEMENTS

FORTRAN compiled programs normally execute statements sequentially in the order in which
they were presented to the compiler.

However, the fol lowing control statements are available to alter

normal sequence of statement execution: GO TO, IF, DO, PAUSE, STOP, and END.
4.1

GO TO STATEMENT
The GO TO statement has two forms: unconditional and computed.

4. l. l

Unconditional
Unconditional GO TO statements are of the form:

GO TO n
where n is the number of an executable statement. Control is transferred to the statement numbered n.

4.1.2

Computed
Computed GO TO statements have the form:

where n ,n , .•• , nk are statement numbers and J is a nonsubscripted integer variable. This statement
1 2
transfers control to the statement numbered n , n , .•• , nk if J has the value 1, 2, .•. , k, respectively.
1 2
If J is zero or if it exceeds the size of the list, execution will proceed to the next executable statement. For example, in the statement
GO TO (20, 10,5), K
the variable K acts as a switch, causing a transfer to statement 20 if K=l, to statement l 0 if K=2, or to
statement 5 if K=3.
4.2

IF STATEMENT
Numerical IF statements are of the form:
IF (expression) n , n , n
1 2 3

where n , n , n are statement numbers. This statement transfers control to the statement numbered n ,
1
1 2 3
n ,n if the value of the numeric expression is less than, equal to, or greater than zero, respectively.
2 3
The expression may be simple.

4-1

Examples:
IF (ETA) 4, 7, 12
IF (KAPPA-L(l0))20, 14, 14

4.3

DO STATEMENT
The DO statement simplifies the coding of iterative procedures.

DO statements are of the

form:

where n is a statement number, i is a nonsubscripted integer variable, and m ,m ,m are integer cons1 2 3
tants or nonsubscripted integer variables. If m is not specified, it is understood to be l .

The DO statement causes the statements which follow, up to and including the statement
numbered n, to be executed repeatedly. This group of statements is called the range of the DO stateIn the example above, the integer variable i is cal led the index, the values of m , m , m are,
1 2 3
respectively, the initial, I imit, and increment values of the index.

ment.

Examples:
DO 10 I= l , 5, 2
DO l 0 f=J ' KI 5
DO 10 L=I, J , K
The index is incremented and tested before the range of the DO is executed.

If the Iimit value is less

than the initial value, the range of the DO will not be executed.
After the last execution of the range, control passes to the statement immediately following
the range. This exit from the range is called the normal exit.

Exit may also be accomplished by a

transfer from within the range.
The range of a DO statement may include other DO statements, provided the range of each
contained DO statement is entirely within the range of the containing DO statement. That is, the
ranges of two DO statements must intersect completely or not at al I. A transfer into the range of a DO
statement from outside the range is not al lowed.
Within the range of a DO statement, the index is available for use as an ordinary variable.
After a transfer from within the range, the index retains its current value and is available for use as a
1

variable'. The values of the initial, limit, and increment variables for the index and the index of the
DO loop may not be altered within the range of the DO statement.
The last statement of a DO loop must be executable, and must not be an IF, GO TO, or
DO statement.
1

The index of a DO loop should not be used as a variable after a normal exit from that DO loop until
it has been redefined.

4-2

4.4

CONTINUE STATEMENT
This is a dummy statement, used primarily as a target for transfers, particularly as the last

statement in the range of a DO statement. For example, in the sequence
DO 7 K=INIT, LIMIT

IF (X(K))22, 13 ,7

CONTINUE

a positive value of X(K) begins another execution of the range. The CONTINUE provides a target
address for the IF statement and ends the range of the DO statement.

4.5

PAUSE STATEMENT
The PAUSE statement enables the program to incorporate operator activity into the sequence

of automatic events. The PAUSE statement assumes one of two forms:
PAUSE
PAUSE n
where n is an unsigned decimal number.
Execution of the PAUSE statement causes the octa I equivalent of the decima I number n, to
be displayed in the accumulator on the user's console. Program execution may be resumed (at the next
executable statement) by depressing the CONTinue key on the console.

4.6

STOP STATEMENT
The STOP statement has the form:
STOP
The STOP statement terminates the program.

END STATEMENT
The END statement has the form:
END
The END statement informs the compiler to terminate compi lotion. The END statement must

be the last statement of the program.

4-3

CHAPTER 5
INPUT/OUTPUT STATEMENTS

Input/output (I/O) statements are used to controi the transfer of data between computer
memory and periphera I devices and to specify the format of the output data. 1/0 statements may be
divided into two categories.
a. A nonexecutable statement that enables conversion between internal data within core
memory and external data: FORMAT
b. Data transmission statements which specify transmission of data between computer
memory and I/O devices: READ and WRITE.

5.1

NONEXECUTABLE STATEMENT
The nonexecutable statement FORMAT enables the user to specify the form and arrangement

of data on the selected external device.

5 .1 . 1

FORMAT Statement
Nonexecutable FORMAT statements may be used with any appropriate input/output device.

FORMAT statements are of the form:

where n is a statement number and each S is a data field specification.
FORMAT statements may be placed anywhere in the source program. Unless the FORMAT
statement contains oniy a iphanumeric data for direct input/output transmission, it wiil be used in
conjunction with the list of a data transmission statement.
Unit records must be one of the following:
a.

A paper tape record preceded by and fol lowed by a carriage return/line feed.

A printed line with a maximum of 72 characters for a Teletype keyboard.

During transmission of data, the object program scans the designated FORMAT statement and
if a specification for a numeric field is present {see Data Transmission Statements) and the data transmission statement contains items remaining to be transmitted, transmission takes place according to the
specification. This process ceases and execution of the data transmission statement is terminated as
soon as all specified items have been transmitted; thus, the FORMAT statement may contain specifications for more items than are specified by the data transmission statement. The FORMAT statement may

5-1

also contain specifications for fewer items than are specified by the data transmission statement, in
which case, format control will revert to the rightmost left parenthesis in the FORMAT statement (see
Section 5.1.1.7).
Both numeric and alphanumeric field specifications may appear in a FORMAT statement. The
FORMAT statement also provides for handling multiple record formats, skipping characters, space insertion, and repetition. If an input list requires more characters than the input device supplies for a given
unit record, blanks are inserted.
5 .1.1.1

Numeric Fields - Numeric field specification codes and the corresponding internal and

external forms of the numbers are listed in the following table.
Table 5-1
Numeric Field Codes
Conversion
Code

Internal Form

Binary floating point

Decimal floating point
with E exponents: .324E+l0

Binary floating point

Decimal floating point with
no exponent: 283 .75

Binary integer

Dec i ma I integer: 79

Externa I Form

Conversions are specified by the form:
rEw.d
rFw.d

rlw
where r is a repetition count, E, F, and I designate the conversion type, w is an integer specifying
the fie Id width, and d is an integer specifying the number of decimal places to the right of the decima I
point. For E and F input, the position of the decimal point in the external field takes precedence c..ver
the value of d.
Example:
FORMAT (15 ,FlO .2 ,El6 .8)
could be used to output the line
bbb32bbbb-17.60bbb.59625476E+o3
on the output listing. (The letter b throughout this manual indicates the presence of a space.)
The field width should always be large enough to include the decimal point, sign, and
exponent. In all numeric field conversions, if the field width is not large enough to accommodate the
converted number, the excess digits on the left are lost; if the number is less than the field width, the
number is right-adjusted in the field.

5-2

5 .1 .1 .2

·Alphanumeric Fields - Alphanumeric data can be transmitted in a manner similar to n_umeric

data by use of the form
rAw
where A is the control character and w is the number of characters in the field. Alphanumeric characters are transmitted as the value of a variable in an input-output ii st; the variabie may be either
integer or rea I .
Although w may have any value, the number of characters transmitted is limited by the
maximum number of characters which can be stored in the space allotted for the variable. This maximum depends upon the variable type; for a real variable the maximum is six characters, for an integer
variable the maximum is two characters. If w exceeds the maximum, the leftmost characters are lost
on input and replaced with blanks on output. If, on input, w is less than the maximum, blanks are
filled in to the right of the given characters unti I the maximum is reached. If, on output, w is less
than the maximum, the leftmost w characters are transmitted to the external device.
5 .1 .1 .3

Hollerith Conversion - Alphanumeric data may be transmitted directly from the FORMAT

statement by using Hollerith conversion (H). H-conversion format is referenced by WRITE statements
only.
In H-conversion, the alphanumeric string is specified by the form
nH h ,h , .•. ,hn
1 2
where H is the controi character and n is the number of characters in the string, including blanks. For
example, the statement below can be used to print PROGRAM COMPLETE on the output listing.
FORMAT (l 7HbPROGRAMbCOMPLETE)
A Hollerith string may consist of any characters capable of representation in the processor. The space

character is a valid and significant chaiactar in a Ho! !erith string. See also Section G. 1 .4 for an
alternate method of outputting alphanumeric data .
5 .1 .1 .4

Mixed Fields - An alphanumeric format field may be placed among other fields of the format.

For example, the statement
FORMAT (15,7HbFORCE=Fl0.5)
can be used to output the line:
bbb22bFORCE=bb 17 .68901
The separating comma may be omitted after an alphanumeric format field, as shown above.

5 .1 .1 .5

Repetition of Fields - Repetition of a field specification may be specified by preceding the

control character E,F, or I by an unsigned integer giving the number of repetitions desired.
For example:

5-3

FORMAT (2El 2 .4,315)

is equivalent to
FORMAT (El2.4,El2.4, 15, 15, 15)

5. 1.1 .6

Repetition of Groups -

A group of field specifications may be repeated by enclosing the

group in parentheses and preceding the whole with the repetition number.
For example:
FORMAT (218,2(El 5. 5,2F8. 3))

is equivalent to
FORMAT (218, El 5. 5,2F8. 3, El 5. 5, 2F8. 3)

5.1 .1 .7

Multiple Record Formats - To handle a group of output records where different records have

different field specifications, a slash is used to indicate a new record. For example, the statement
FORMAT (318/15,2F8.4)

is equivalent to
FORMAT (318)

for the first record and
FORMAT (I5,2F8 .4)

for the second record.
The separating comma may be omitted when a slash is used. When n slashes appear at the
end or beginning of a format, n blank records may be written on output or ignored on input.
When n slashes appear in the middle of a format, n-1 blank records are written or n-1 records skipped.
Both the slash and the closing parentheses at the end of the format indicate the termination of a record.
If the list of an input/output statement dictates that transmission of data is to continue after the closing
parenthesis of the format is reached, the format is repeated from the last open parenthesis of level one
or zero. Thus, the statement:
FORMAT (F7. 2, (2(E15. 5, El 5.4) I 17))

level 0

_t
J
level l

level l

causes the format:
F7.2,2(El5.5,El5.4), 17

to be used on the first record, and the format:
2(El5.5,E15.4),17

5-4

J"L level 0

to be used on succeeding records.
As a further example, consider the statement:
FORMAT (F7.2/(2(E15.5,E15.4),I7))
The first record has the format:
F7.2
and successive records have the format:
2(E15.5,E15.4),I7
5.1 .1 .8

Blank or Skip Fields - Blanks may be introduced into an output record or characters skipped

on an input record by use of the specification nX. The control character is X; n is the number of
blanks or: characters skipped and must be greater than zero. For example, the statement:
FORMAT (5HbSTEPI5, 1OX2HY=F7 .3)
may be used to output the Ii ne·:
bS TE Pbbb28bbbbbbbbbb Y=bb3 • 872

5.2

DATA TRANSMISSION STATEMENTS
There are two data transmission statements, ·READ and WRITE. Data transmission statements

accomplish input/output transfer of data that may be .1 isted in a FORMAT statement. The data transmission statement contains a Iist of the quantities to be transmitted. The data appears on the external
device in the form of records.

1
a. Inout/Outout Lists - The I ist of an input/output statement specifies the order of transmission of the variable values. During input, the new values of I isted variables may be used in subscript or control expressions for variables appearing later in the I ist.
For example,
READ(2, 1000)L,A(L),B(L+1)
reads a new value of Land uses this value in the subscripts of A and B. Where 2 is the device designation code, and 1000 is a FORMAT statement number.
b. Input/Output Records - All information appearing on input is grouped into records. On
output to the printer a record is one line. The amount of information contained in each ASCII record
is specified by the FORMAT reference and the input/output list.

The implied DO in input/output lists is not implemented.

5-5

Each execution of an input or output statement initiates the transmission of a new data
record. Thus, the statement
READ{l, lOO)FIRST ,SECOND I THIRD
is not necessarily equivalent to the statements where 100 is a FORMAT reference
READ(l, lOO)FIRST
READ(l, lOO)SECOND
READ(l, lOO)THIRD
since, in the second case, at least three separate records are required, whereas, the single statement
READ (d If) FIRST ,SECOND I THIRD
may require one, two, three, or more records depending upon FORMAT f.

If an input/output statement requests less than a full record of information, the unrequested
part of the record is lost and cannot be recovered by another input/output statement without repositioning the record.

If an input/output list requires more than one ASCII record of information, successive records
are read.

5.2 .1

READ Statement
The READ statement specifies transfer of information from a selected input device to internal

memory, corresponding to a list of named variables, arrays or array elements. The READ statement
assumes the fol lowing form:
READ (d,f) list

where dis a device designation which may be an integer constant or an integer variable, f is a format
reference, and list is a list of variables.
The first form of the READ statement causes ASCII information to be read from the device
designated and stored in memory as values of the variables in the list. The data is converted to
internal form as specified by the referenced FORMAT statement.
For example:
READ (1, 15) ETA, PI

5-6

5.2.2

WRITE Statement
The WRITE statement is used to transmit information from the computer to a specified output

device. The WRITE statement assumes one of the fol lowing forms:
WRITE (d,f) list
WRITE (d ,f)
where dis a device designation (integer constant or integer variable) f is a format reference, and list is
a list of variables.
The first form of the WRITE statement causes the vaiues of the variables in the iist to be read
from memory and written on the device designated in ASCII form. The data is converted to external
form as specified by the designated FORMAT statement.
The second form of the WRITE statement causes information to be read directly from the
specified format and written on the device designated in ASCII form.

5.2.3

Device Designations
The I/O device designations are used in the READ and WRITE statements. The device codes

are:
Device Code

Designating
Teletype and !ow-speed reader and punch

High-speed reader and punch

For additional I/O information, see SABR Manual, DEC-08-ARXA-D.

5-7

CHAPTER 6
SPECIFICATION STATEMENTS

Specification statements allocate storage and furnish information about variabies and constants
to the compiler. The specification statements are DIMENSION, COMMON, and EQUIVALENCE, and
when used, must appear in the program prior to any executable statement.

6.1

COMMON STATEMENT
The COMMON statement causes specified variables or arrays to be stored in an area avail-

able to other programs. By means of COMMON statements, the data of a main program and/or the
data of its subprograms may share a common storage area. Variables in COMMON statements are
assigned to locations in ascending order in field l beginning at location 200 (see Appendix D). The
COMMON statement has the general form:

6.2

DIMENSION STATEMENT
The DIMENSION statement is used to declare identifiers to be array identifiers and to

specify the number and bounds of the array subscripts. The information supplied in a DIME NS ION
statement is required for the allocation of memory for, arrays. Any number of arrays may be declared
in a single DIMENSION statement. The DIMENSION statement has the form:
DIMENSION S ,s , ••• ,Sn
1 2
where S is an array specification.
Examples:
DIMENSION A (100)
DIMENSION Y(lO) ,PORT(25) ,A(lO, 10) ,J(32)

NOTE
When variables in COMMON storage are dimensioned,
the COMMON statement must appear before the DIME NSION statement.

6.3

EQUIVALENCE STATEMENT
The EQUIVALENCE statement causes more than one variable within a given program to

share the same storage location. The EQUIVALENCE statement has the form:
6-1

EQUIVA LEN E (V l

Iv2

I ••• ) I

Nk Iv k +l

I ••• ) I •••

where the V's are variable names.
The inclusion of two or more references in a parenthetical list indicates that the quantities
in the list are to share the same memory location.
For example,
EQUIVALENCE (RED I BLUE)
specifies that the variables RED and BLUE are stored in the same place. The subscripts of array
variables must be integer constants.
Example:
EQUIVALENCE (X,A(3) I Y2, 1)) I (BETA(2 ,2) ,ALPHA)
Identifiers may not appear in both EQUIVALENCE and COMMON statements.

6-2

CHAPTER 7
SUBPROGRAM STATEMENTS

7 .1

GENERAL

External subprograms are defined separately from {i.e., external to) the programs that cal I
them, and are complete programs which conform to all the rules of FORTRAN programs. They are
compiled as closed subroutines, that is, they appear only once in core memory regardless of the number or times they are used. External subprograms are defined by means of the statements FUNCTION
and SUBROUTINE.

7 .1 .1

Dummy Identifiers
Subprogram definition statements contain dummy identifiers, representing the arguments of

the subprogram. They are used as ordinary identifiers within the subprogram and indicate the sort of
arguments that may appear and how the arguments are used. The dummy identifiers are replaced by
the actual arguments when the subprogram is executed.

7 .2

FUNCTION SUBPROGRAMS
A function subpiOgiOm is a single=valued function that may be called by using its name as

a function name in an arithmetic expression, such as FUNC(N), where FUNC is the name of the subprogram that evaluates the corresponding function of the argument N. A function subprogram begins
with a FUNCTION statement and ends with an END statement. It returns control to the calling
program by means of one or more RETURN statements.

7 .2 .1

FUNCTION Statement
The FUNCTION statement has the form:
FUNCTION identifier (a ,a , ... ,an)
1 2

This statement declares the program which follows to be a function subprogram. The identifier is the
name of the function being defined. This identifier must appear as a scalar variable and be assigned a
value during execution of the subprogram which is the function value.
Arguments appearing in the list enclosed in parentheses are dummy arguments representing
the function argument. The arguments must agree in number, order, and type with the actual arguments
used in the calling program. Function subprogram may have expressions and array names as arguments.

7-1

Dummy arguments may appear in the subprogram as sea lar identifiers or array identifiers. A
function must have at least one dummy argument. Dummy arguments representing array names must
appear within the subprogram in a DIMENSION statement. Dimensions must be given as constants and
should be smaller than or equal to the dimensions of the corresponding arrays in the calling program.
A function should not modify any arguments which appear in the FORTRAN arithmetic
expression calling the function. The only FORTRAN statements not allowed in a function subprogram
are SUBROUTINE and another FUNCTION statement.

7.2.2

Function Type
The type of function is determined by the first letter of the identifier used to name the func-

tion, in the same way as variable names.

7 .3

SUBROUTINE SUBPROGRAMS
A subroutine subprogram may be multivalued and can be referred to only by a CALL state-

ment. A subroutine subprogram begins with a SUBROUTINE statement and returns control to the calling
program by means of one or more RETURN statements.

7 .3 .1

SUBROUTINE Statement
The SUBROUTINE statement has the form:
SUBROUTINE identifier {a ,a , .•• ,an)
1 2

This statement declares the program which follows to be a subroutine subprogram. The first identifier is
the subroutine name. The arguments in the list enclosed in parentheses are dummy arguments representing the arguments of the subprogram. The dummy arguments must agree in number, order, and type with
the actual arguments used by the calling program.
Subroutine subprograms may have expressions and array names as arguments. The dummy
arguments may appear as sea lar or array identifiers •
Dummy identifiers which represent array names must be dimensioned within the subprogram

by a DIMENSION statement. The dummy arguments must not appear in an EQUIVALENCE or COMMON
statement in the subroutine subprogram.
A subroutine subprogram may use one or more of its dummy identifiers to represent results.
The subprogram name is not used for the return of results. A subroutine subprogram need not have any
argument at all.

7-2

Examples:
SUBROUTINE FACTOR (COEFF ,N,ROOTS)
SUBROUTINE RESID U(NUM,N ,DEN ,M,RES)

SUBROUTiNE SERiES
The only FORTRAN statements not allowed in a function subprogram are FUNCTION and another
SUBROUTINE statement.

7.3.2

CALL Statement
The CALL statement assumes one of two forms:
CALL identifier
CALL identifier (argument, argument, ..• argument)

The CALL statement is used to transfer control to a subroutine subprogram. The identifier is the subroutine name.
The arguments may be expressions or array identifiers. Arguments may be of any type, but
must agree in number, order, type, and array size with the corresponding arguments in the SUBROUTINE
statement of the ca!!ed subrot.itine. Unlike a function, a subroutine may produce more than one value
and cannot be referred to as a basic element in an expression.
A subroutine may use one or more of its arguments to return results to the calling program.

If no arguments at all are required, the first form is used.
Examples:
CALL EXIT
CALL TEST (VALUE, 123,275)
The identifier used to name the subroutine is not assigned a type and has no relation to the types of the
arguments. Arguments which are constants or formed as expressions must not be modified by the sub-

....
rou.1ne.
7.3.3

RETURN Statementi~=,
The RETURN statement has the form
RETURN

This statement returns control from a subprogram to the calling program. Each subprogram must contain
at least one RETURN statement. Normally, the last statement executed in a subprogram is a RETURN
statement. Any number of RETURN statements may appear in a subprogram. The RETURN statement
may not be used in a main ·program.

7-3

CHAPTER 8
OPERATING INSTRUCTIONS

This chaptei desciibes how to compile., assemble, and execute a FORTRAN program using

the 8K FORTRAN Compiler, SABR Assembler, and Linking Loader. The PDP-8/I System User's Guide,
DEC-08-NGCB-D, is frequently referenced for loading instructions.
Except when loading. the Linking Loader (Section 8 .5), the DF setting is ignored because
all other system tapes have field settings coded on the tapes.

8. l

LOADING THE COMPILER

8 .l .l

Loading Into Core Memory
a.

Make sure the Binary Loader is in memory, say field l.

Place the FORTRAN Compiler binary tape in the reader.

c. Set the console switches as follows: (Data field is ignored) instruction field = 1,
Switch Register = 7777.
d.

Press LOAD AD Dress.

Depress Switch Register bit 0.

Press START

g. The FORTRAN Compiler has now been loaded into memory by the Binary Loader.
Parts of the compiler wi II load into field 0 and field l .

8 .1 .2

Loading on the Disk

Make sure the Disk Monitor ·is in memory. (Type CTRL/C or START at 7600 .)

When the Monitor responds with a dot, call the system loader by typing
. LOAD,/

(the J denotes typing the RETURN key)

Place the Compiler binary tape in the reader.

Answer the Loader command dialogue as fol lows:
* IN-R: J

*
*OPT-2
*ST= J
t <cTRL/P > t <CTRL/P > t <CTRL/P>

CTRL/C and CTRL/P are typed by holding down the CTRL key while typing the C or P key. They
do not echo (print) when typed, therefore, their presence are indicated by being enclosed in angle
brackets.

8-1

After typing the second CTRL/P remove the tape from the reader and place it back in
the reader for the second pass.
e • The FORTRAN Compi Ier has now been Ioaded into memory, parts into fie Id 0 and fie Id l .
It must now be saved on the system device as follows:
.SAVE FTCO ! 0 - 7577; 5363J
.SAVE FTCl ! 200, 1000 - 1577, 2600, 6000 - 16377; J

The Compiler has now been saved on the user's system device and may be called as

follows:
.FTCl J
.FTCO,;
The field l part must be called first.

8 .2

OPERATING THE COMPILER

It is assumed that the programmer has written his main program and possibly one or more
subprograms, and that these source programs have been punched on paper tape in ASCII format.
Remember that each source tape must have an END statement at the end of the tape.
After the Compiler has been loaded into memory, it is used to translate each FORTRAN
statement into one or more SABR assembler instructions. The Compiler output wi 11 be punched in two
parts separated by approximately three feet of blank tape. The first part, (executable code) wi II be
punched as the source tape is read. The second part, (variable storage and constants) wi II be punched
after the entire source tape has been read.

If the Compiler has been saved on the Disk Monitor System, it wi II halt after it is loaded
into memory. Be sure that the source tape has been placed in the reader and the punch has been
turned ON, then simply press CONTinue to begin step(d).

It may be desirable to suppress all compiler output the first time a particular program is
compiled, simply to check for errors. To do this it is necessary to load the Compiler and then deposit
3075 in location 0356 (field O), prior to executing step (c) below.
a. Set the console switches as follows: Data field= 0, Instruction field= l
Switch Register = 1000. (The Compiler may also be started at location 5364 in field 0 .)
b.

Place the FORTRAN program source tape in the reader / and press the punch ON.

Press LOAD ADDress and START.

d. As soon as the Compiler has typed out an identification number, it wi II begin compiling
the user's program. The Compiler output wi II generally be severe I times the length of the FORTRAN
source program.

8-2

e. If an error is discovered in the user's FORTRAN program, the Compi fer wi II type the
incorrect line, fol lowed by an error message. Although eompi ler output wi II be suppressed, the rest
of the user's program will be read, and additional error messages may be typed.

f. When the Compiler has finished punching both sections of tape it will halt. It may
be restarted to compi ie additionai programs by pressing CONTinue.
g. The FORTRAN Compiler may be restarted at any time by pressing STOP and going
back to step (a).

8 .3

LOADING THE SABR ASSEMBLER

8 .3 .1

Loading Into Core Memory
a.

Make sure the Binary Loader is in memory, say in field n.

b. Set the console switches as follows:
(Data field is ignored)

Instruction Field = n, Switch Register = 7777.

Press LOAD AD Dress.

Insert the SABR binary tape into the reader.

If using the high-speed reader depress Switch Register Bit 0.

Press START.

g. SABR wi II now be loaded into memory by the Binary Loader. Portions of SABR wi II
load into Field 0 and Field 1 •

8.3.2

Loading on the Disk
a.

Make sure the Disk Monitor is in memory. (Type CTRL/C or START at 07600 .)

When the Monitor responds with a dot, call the system Loader as follows:
.LOAD J

Insert the SABR binary tape in the reader.

Answer the loading command dialogue as follows:
*IN-R:,) for high-speed reader or

*IN-T: J for ASR reader

*
*OPT -2
*ST =,)
t <CTRL/P>

t <CTRL/P> t <CTRL/P> t <CTRL/P>

After typing the second CTRL/P remove the tape from the reader and place it back
in the reader for the second pass •
e. SABR is now loaded into memory, partly in Fiefd 0 and partly in Field 1. It may be
saved on the user's system device by responding to the monitor 1s dot as follows:
. SAVE SABR! 0-7177; 200 J
• SAVE SAB l! 12000 - 12427;_,

8-3

SABR is now saved on the user's system device and may be called as follows:
.SABl J
.SABR J
The Field 1 portion must be called first.

8 .4

OPERATING THE SABR ASSEMBLER
In addition to being a stand-alone assembler, SABR also serves as the second pass of BK

FORTRAN compi lotion. For this purpose the use of SABR is slightly different from that described in
the SABR manual. This difference in the operation of SABR is due only to the unusual format of the
FORTRAN Compiler output.
The Compiler, in one pass, converts the user's FORTRAN source into a symbolic machine
language program tape. SABR then converts the symbolic tape into relocatable binary. However,
the symbolic tape produced by the Compiler is not a standard format SABR language tape. It is
arranged as shown in the figure below.

L
E
A
D
E

B
L
A
N

0
R
T
R

Main part of program;
Executable code.

Symbol Definitions
Common, Arrays,
Data and
Program Entry
Point.

T
A

p
A

s
E

T
R
A

1
L
E
R

True Start

The tape is arranged this way because the data at the end of the tape cannot be inserted in
the midst of the executable code, and some of it which should be qt the beginning of the tape is not
known unti I later. Thus the true start of the symbolic program is near the end of the symbolic tape
preceded by a segment of leader/trailer code and fol lowed by a PAUSE statement.

To assemble such a tape with SABR one of three methods must be followed. Actually, the
general procedure is the same as that described in the SABR manual, but in particular detai Is it differs.

The differences are covered by the three methods explained belovv.

8 .4.1

Method 1
The simplest method is to cut the symbolic tape into two parts. The cut should be made at

the middle of the blank tape which separates the executable code from the symbol definitions. The
latter section of the tape should then be marked "Section 1 11 and the former section (the executable

8-4

code) should be marked 11 Section 2 • 11 Assembly then proceeds with the two part symbolic tape exactly
as described below.
After SABR has been loaded into memory, it is used to assemble the Compiler output. In
the first pass through SABR the reiocatabie binary version of the user's program is created and, at the
end of this pass, the symbol table may be typed and/or punched. Pass 2 is the listing pass. The assembly
is carried out as fo I lows.

If SABR has been saved on the system 1/0 device as in Section 8.3, it will start automatically
at step (c) below when called into memory. The source tape (first section) should be inserted in the
reader before operation begins.
It may be desirable to suppress all assembler output the first time a particular program is
assembled, simply to check for errors. To do this it is necessary to load SABR and then deposit 5370
in location 3165 (Field 0) before beginning step (a) below.
a. Set the console switches as follows: Data field = 0, Instruction field = 0,
Switch Register = 0200.
b.

Press LOAD ADDress and ST ART .

SABR now types a sequence of two or three questions;
11

HIGH SPEED READER? 11
HIGH SPEED PUNCH ? 11
11
LISTING ON HIGH SPEED PUNCH ? 11
11

These questions must be answered with 11 Y 11 if the answer is 11 yes. 11 Any other answer
is assumed to be 11 no. 11 The third question is typed only if the second is answered 11 Y 11 • If the third
is answered 11 Y, 11 both the symbol table and the listing will be punched on the high-speed paper tape
punch. Otherwise, they are typed on the teletypewriter. Incidentally, the user need not wait for
the ful I question to be typed before responding.
d. As soon as SABR has echoed the user 1s response to the last question, the punch device
and, if it is being used, the ASRreader should be turned on. If using the iow-speed reader, the error
message E indicates that the user has waited too long before turning the reader on. He will have to
start over.
e. At this point, Pass 1 begins. SABR reads the source tape and punches the binary tape.
After the binary tape has been completed SABR wi 11 type or punch the program symbol table.
f. If the source tape is in several sections (separate tapes with PAUSEs at the end of al I
except the last), SABR will halt at the end of each section. At this point the user should insert the
next section in the reader and then press CONTinue.
g.

At the end of Pass 1 SABR wi 11 halt.

h. If the user desires an assembly listing, he should now reposition the beginning of the
source tape in the reader and press CONTinue.

If the listing is going to be punched on the high speed punch, the user may want to
list the symbol table (at the end of the binary relocatable tape) before beginning Pass 2.

8-5

i. At the end of Pass 2 SABR will again halt. It may be restarted for assembling another
program by pressing CONTinue.

j. SABR may be restarted at any time by pressing STOP, setting the Switch Register = 0200,
pressing LOAD ADDress and START. However, Pass 1 must always be repeated.

8.4.2

Method 2
The user may avoid actually cutting the symbolic tape by cleverly manipulating the tape as

if it were two parts as explained above. The tape should initially be inserted in the reader with the
separator blank tape over the read-head. When SABR halts at the PAUSE statement at the physical
end of the tape, the user should reposition the tape, putting the physica I beginning of the tape in the
reader. Then press CONTinue. The assembly pass will end at the separator blank tape code. The
assembly listing can be produced in a similar manner, pressing CONTinue to start the Listing pass.

8.4.3

Method 3
The third method requires SABR to pass over the symbolic tape two times for each pass of the

assembly. However, it allows the tape to be inserted at its physical beginning. It is based on the
fact that a symbolic tape output by the FORTRAN Compiler has as its physical first line the special
pseudo-op, FORiR. ihis pseudo-op has no effect except when a symboiic tape output by the Compi ier
is assembled using this third method.
The method is this:
a.

Insert the symbolic tape in the reader at its physical beginning.

Start SABR as usua I .

c.
Sensing the FORTR statement as the first line, SABR ignores all further data until
after it passes over the END statement. SABR then begins the actual assembly by processing the
symbol definitions, etc., which are at the latter end of the tape.
d·. Then SABR ha Its at the PAUSE statement which is at the physical end of the tape.
At this time the user should reposition the symbolic tape in the reader at the physical beginning of the
tape, and then press CONTinue. SABR wi 11 now assemble the executable code portion of the tape in
the norma I way .
e. If the user desires an assembly listing, he should proceed as in Method 2 after SABR
finishes the assembly pass.
One further type of error may occur. This is an undefined symbol. Because SABR is
a one pass assembler, this can not be determined unti I the end of the assembly pass, so the error
diagnostic UNDF is given in the symbol table listing.

8-6

8 .5

THE LINKING LOADER
Relocatable binary program tapes produced by SABR are loaded into memory by the BK

System Linking Loader. The Linking Loader is capable of loading and linking a user 1s program and
subpmgiams in any fields of memory. It is even capable, in a speciai way, of loading programs over
itself. The Linking Loader also has options which give storage maps and core availability.
Generally speaking, the Linking Loader is capable of loading any number of user and
Library programs into any field of memory. These programs are loaded one after the other via the
high-speed reader (or the ASR reader). The choice of which field to load each program into is a
Switch Register option. Usually several programs may be loaded into each field.

Because of the space

reserved for the Linkage Routines, the available space in Field 0 is three pages smaller than in all
other fields.
Any COMMON storage reserved by the programs being loaded is allocated in Field 1 from
location 0200 upwards. The space reserved for COMMON is obviously subtracted from the available
loading area in Field 1. The program reserving the largest amount of COMMON storage must be
loaded first.
The Linking Loader uses the following special method to enable loading data over itself.
When the Linking Loader encounters data which must be loaded ovei itself, it punches this date onto
paper tape in RIM format. Then after the user has finished loading all his relocatable binary program
tapes, all that is necessary is to load the RIM format tape using the RIM loader.
The Run-Time Linkage Routine necessary to execute SABR programs are automatically loaded
into the required areas of every field by the Linking Loader as a part of its initialization. The user
needs to know nothing more about these routines than the particular areas of core they occupy. (See
Appendix D of the SABR manual.)
The 8K System Library subprograms, which may be used by any SABR program, are loaded
in the same way as any other refocatable binary programs. Only those Library programs which the
user's programs actually call need to be loaded. Refer to the SABR manual for additional information.
During the loading operation with the Linking Loader, two options are available to the
user to obtain information about what has been loaded so far.
The Switch Register is used to select these options. Either option may be selected after
any program has finished loading.

('Naming: if the ASR punch is turned on, it must be turned off

before selecting these options.) The Switch Register bits used are as follows.
BIT 0 = 1 selects the Core Availability option;
BIT 1 = 1 selects the Storage Map option.

8-7

The Core Availability option, when selected, causes the number of free pages of memory in
every field of memory to be typed in a list on the teletype. For example, if the user has a 16K configuration a list like the fol lowing might be typed.
0002
0010
0030
0036

(number of free pages in field 0)
(number of free pages in fie Id 1)
(number of free pages in fie Id 2)
(number of free pages in fie Id 3)

The number of pages initially available in field 0 is 0033 and in all other fields is 0036.
The Storage Map option, when selected, causes a list of all program Entry points to be typed
along with the actual address at which they have been loaded. Entry points of programs which have
been called but which have not been loaded are also listed along with a U flag for "undefined. 11
Such flagged programs must be loaded before execution of the user's programs are possible. The core
availability list is automatically appended to the storage map. A sample is shown below.
MAIN
READ
WRITE
I OH
SE TERR
ERROR

10200
01055
01066
03031
00000
00000

T TVf""\1 IT

COCCO v

H SOUT
TT YIN
H SIN
F DV
CLEAR
I FAD
FMP
ISTO
STO
FLOT
FAD
DIV
I REM
F SB
FLOAT
FIX
IFIX
CHS
0011
0033

00000
00000
00000
04722
05247
05131
04632
05074
04447
05210
04010
00000
00000
04000
05046
04513
04561
05231

\.JV I

8-8

u
u
u

8.6

LOADING THE LINKING LOADER
The Linking Loader must be loaded into the highest available field of memory.
a.

Make sure the Binary Loader is in memory, say in field l .

b. Set the console switches as fol iows: Data fieid = h, instruction fie id = i,
Switch Register = 7777. Where h represents the number of the highest field in the user's configuration.

8.7

Press LOAD AD Dress.

Place the binary paper tape of the Linking Loader in the reader.

If using a high-speed reader, depress Switch Register Bit 0.

Press START. The Linking Loader wi II now be loaded into memory.

OPERATING THE LINKING LOADER
The Linking Loader is used to load the user's relocatable programs and 8K system Library

subprograms as outlined below.
The program or subprogram which uses the largest amount of COMMON storage should be
loaded first.

(The Library subprograms do not use COMMON.)

a. After the Linking Loader has been loaded into the highest memory field, h, the user
should set the console switches as follows: Data Field = h, Instruction Field = h, Switch Register =0200.
b.

Press LOAD AD Dress.

c. Place the relocatable binary tape for the first program to be loaded in the reader. It
should be positioned with leader code in the reader.
d. The Switch Register should be set to 0000. Then, if loading via the ASR reader is
required, raise Switch Register Bit 6. If the user does not have high-speed punch, he should raise
Switch Register Bit 7. Finally the user should set Switch Register Bits 9-11 to the number of the field
into which he wishes to load the first program or subprogram.

Number of
Loading Field

Example:

If the user wishes to load his first program into Field 3, and if he has no high-speed

1/0 device, then he should set the Switch Register to 0063 before the next step.

* Al I other Switch Register bits are irrelevant.
8-9

Press START.

The user's relocatable binary program wi II now be loaded. When loading is completed
the Linking Loader wi II halt.

The user may now either load another program or select one of the options.

h. To load another program, insert the program relocatable binary tape in the reader,
set Switch Register Bits 9-11 to the number of the field the program is to be loaded into, and then
press CONTinue.
i.
CONTinue.

To select the Core Availability option, set Switch Register Bit 0 = l and press
To select the Storage lv\ap option, set Switch Register Bit 1 = 1 and press CONTinue.
WARNING

If the ASR punch is turned on for possible RIM format
data punching {as explained in Section 6 .2), be sure
to turn it off before selecting either of the options
and to turn it back on after the typing of the option
is completed.
k.

The user may continue loading more programs as in step (h) after using either of the

options.
Any time the Linking Loader halts the user may access memory directly via the
DE Posit and EXAMine console switches. After this is done the linking Loader may be restarted vie
the console switches at location 7200 (in the highest field, where the Linking Loader resides).
In general, all five parts of FORTRAN Library Tape I must be loaded before any
FORTRAN program can be executed. The five parts of Library II may be loaded selectively as
determined by the Storage Map option, and the following table:

8 .7. 1

Library Organization
Tape I

1)
2)

IOH 11
11
FLOAT 11

contains
contains

"INTEGER"
11
UTILITY 11

contains
contains

ERROR 11

contains

Tape II

1)
2)

SUBSC 11
"POWERS"

contains
contains

contains
contains
contains

SQRT 11
TRIG 11
11
ATAN 11

8-10

IOH I READ I WRITE
FAD I FSB I FMP I FDV I STO I FLOT I
FLOAT, FIX, IFIX, IFAD, ISTO, CHS,
CLEAR
IREM, ABS I IABS I DIV I MPV I IRDSW
TTYIN I TTYOUT I HSIN I HSOUT I
OPEN, CKIO
SETERR, CLRERR, ERROR
SUBSC {Subscripting routine)
IIPOW I IFPOW I FIPOW I FFPOW I EXP I
ALOG
SQRT
SIN I cos I TAN
ATAN

8 .8

EXECUTING THE FORTRAN PROGRAM
Determine the starting address of your Main program by using the Linking Loader Storage Map

option. The address wi II be typed in the form:

MAIN

dnnnn

Set Data field = d, Instruction field = d, Switch Register = nnnn.

Turn on paper tape punch and/or put data tape in reader as required.

Press LOAD AD Dress, and START.
Prngram execution will begin.

8-11

CHAPTER 9
DEMONSTRATION PROGRAM

The purpose of this program is to compute the factorials of the even integers from 1 through
34. The MAIN program calls the subprogram to perform the computation.
This demonstration program was run on a PDP-8/I with 8K words of core memory and a highspeed photoelectric reader and punch. The demonstration, from start to finish, required 15 minutes.
Actual Teletype printout is used below.

10
60

FORTRAN DEMONSTRATION PROGRAM
DIMENSION AC35)
DO 10 N=2,34,2
ACN)=FACTCN)
WRITE Cl,60)N,ACN)
STOP
FORMAT CI3,4H! = ,E12.7)
END

Both source programs were written using the
Symbolic Editor, listed on the Teletype for
inclusion here, and punched on the high-speed
punch.

C
1

3
4
5
6

FORTRAN FUNCTION TO COMPUTE FACTORIALS
FUNCTION FACTC N)
IF CN-34) 1,s,s
IF CN) 2,4,2
M=N-2
FACT=N
DO 3 K=l,M
C=N-K
FACT=FACT*C
RETURN
FACT=l •
RETURN
WRITE (1,6) N
FACT=0
RETURN
FORMAT CIS,30H! EXCEEDS CAPACITY OF MACHINE.)
END

PDP-8 FORTRAN DEC-08-A2Bl-3

This is the system program tape identification.
Loaded the FORTRAN Compiler and compiled both
source programs.

9-1

PDP-8 SABR DEC-08-A282-10
HIGH SPEED READER? Y
HIGH SPEED PUNCH? Y
LISTING ON HIGH SPEED PUNCH? N

CKIO

FACT
IOH
ISTO
MAIN
OPEN
SUBSC
WRITE
[0

\A
\N
\10
\60
tA
18

1C
tD
tE
tF

Loaded the SABR Assembler, responded to the initial
dialogue and assembled both compiled programs.

0000EXT
0000EXT
0000EXT
r?J000EXT
0352EXT
0000EXT
0000EXT
0000EXT
0 512
0200
0 351
0426
0 501
0 361
0473
0 411
0450
0463
0476
0 512

HIGH SPEED READER? Y
HIGH SPEED PUNCH? Y
LISTING ON HIGH SPEED PUNCH? N

FACT
FAD
FLOT
FMP
IOH
OPEN
STO
WRITE
[0

\C
\FACT
\K

\M
\N
\1
\2
\3
\4
\5
\6
]3
1A
18
tC
tD

0215EXT
0000EXT
0000EXT
0000EXT
0000EXT
0000EXT
0000EXT
0000EXT
0473
0205
0201
0204
0200
0 471
0251
0261
0 331
0354
0406
0445
0210
0305
0346
0 422
1471

9-2

DEC-08-A2B3-5
MAIN
OPEN
SUB SC

0 11 52
10325
11000

FACT

0i415

ISTC
WRITE

06062
02066
03744
10 321
06200
0 544LJ
0 5010
05623
02055
10400
10 503
10227
10255
10200
10245
0 5711
06237
0 611 7
06443
0 6616
05000
06034
05510
05556
06221
06636
06700
06400
06723
10344
10 431

IOH
CKIO
FLOT
STO
FP.D
FMP
READ
SET ERR
ERR CR
TTY OUT
HS OUT
TTYIN
HSIN
FDV
CLEA.R
IFAD
DIV
IREM
FSB

FLOAT
FIX
IFIX
CHS

ABS
IABS
MPY
IRDSW
EXIT
CLRERR
0003
0032

Loaded the Linking Loader and the Library programs
{al I of the first tape and the first section of the second
tape, and then set the switch register for the memory
map).

Loaded the relocatable binary tapes and started the

MAIN program at location 01152 (see memory map).
Received the expected results .

2! = • 2000000E+01
4! = .2400000£+02
6 ! = .7200000E+03
8! = • 40 32000 E+0 5
10 ! = .3628800E+07
12 ! = .4790016E+09
1 4 ! = .8717829E+ll
16 ! = .2092279E+14
.6402374E+16
18 !
20! = .2432902E+19
22! = ·1124001E+22
24! = • 620 4484E+24
26! = -4032915E+27
28! = .3048883E+30
.2652529E+33
30 !
32! = .2631308E+36
34! EXCEEDS CAPACITY OF MACHINE.
34! = ·0000000E+00

9-3

APPENDIX A
DECIMAL AND OCTAL REPRESENTATIONS OF THE CHARACTER SET

This version of FORTRAN uses six-bit characters. Other ASCII characters wi 11 be trimmed
to six bits and translated to even parity, resulting in one of the characters in the table.
A*

Carriage
Return
A
B

[

27
28
29

33
34
35

48
49
50

31
32
33

37
40
41
42
43
44
45
46
47
50
51
52
53

63
64

\
]

,.,,,.,,

01
02
03
04
05
06
07
10
11
12
13
14
15
16
17
20
21
22
23
24
25
,.,, ,

23
24
25
26

27
30
31
32

A*
B*
C*

character
decimal representation
octal equivalent

c
D
E
F
G
H
I
J
K
L
M
N
0
p
Q

l
2

3
4

5
6
7
8
9
10
ll
12

13
14
15
16
17
18
19
20
21
LL

Space
!
II

35
36
37
38
39
40
41
42
43
44
45
46
47

$
%
&
(
)

A-1

55
56
57

2
3
4

5
6

7
8
9
i

<
>
?

52
53
54
55
56
57
58
59
60
61
62
63

62
65
66
67
70
71
72

73
74
75
76
77

APPENDIX B
STATEMENT SPECIFICATIONS

r
FORM

STATEMENT

WHERE

(R or P indicates a required or
prohibited statement number,
N indicates nonexecutable
statement)

I COMMENT

C 11 in column l

columns 2 through 80 w! 11 be
ignored.

CONTINUE

contro I goes to next statement.

ARITHMETIC

v=e

variable name = expression.

GO TO

GO TO n

G 0 T0 ( n , ... , nm), i
1

n is a statement number

1 ~ i ~ m and control goes to
statement n .. i is a nonsubscripted
I

integer variable.
n- 'E'
1r
l J n ,n ,n
1 2 3

I
DO

{1\;j

contro I goes t

expression t

DO n i=m , m , m
1 2 3

"~} 0.
;.

repeated execution through statement n beginning with i=m ,
1
incrementing by m , while
3
; ;s loss +!.-.,.. .... """ ""''1'"'""4:1
o,.., ,,..I +............
1112•
I

........

1 llUll

'"'I

I""

m's and i may not be subscripted.

PAUSE

STOP

END

DO n i=m ,m
1 2

m assumed to be l .
3

PAUSE

temporary halt, resumed by
CONTinue key.

PAUSE n

octal equivalent of the integer n
displayed.

STOP

must be used to halt execution of
a main program.

STOP n

END

octal equivalent of the integer n
displayed.
an END statement at the end of a
subprogram tells the compiler
there is no more program.

B-1

Ir STATEMENT l!
r
~

WHERE

FORM

I READ

I WRITE

;

READ (d, f) L
WRITE (d, f) L

d is device number, f is a
FORMAT statement number and L
is list of variable names separated
by commas.

FORMAT (k , ... , kn)
1

k's are format specifications

I NP COMMON a, b, ... , n

a, ... , n are nonsubscripted
variable names

I
•

FORMAT
COMMON*

DIMENSION

DIMENSION a (k ), ... ,an (kn)
1 1

o's are array names and k's are
maximum subscripts.

FUNCTION

FUNCTION name (a , ... ,an)
1

a 's are dummy arguments and
function name must be defined as
a variable containing the value
of the function.

SUBROUTINE

SUBROUTINE name (a , ... ,an)
1

o's are dummy arguments and subroutine name may not appear
elsewhere in the subroutine.

CALL name (a , ... ,an)
1

o's are actual arguments of a
subroutine and may be expressions.

RETURN

for subroutines, control returned
to statement following CALL. For
functions, evaluation of expression
in calling program is resumed
using value of the function.

EQUIVALENCE (V , ... , Vn), · ·.,
1
(V I ... Iv )
m
p

V's are variables or subscripted
array names.

! CALL

RETURN

EQUIVALENCE

B-2

APPENDIX C
FORMAT SPECIFICATIONS

1
WHERE

FORM

KIND
Integer

rlw

r is the repetition count; w is total field width in
characters .

Floating Point
(Decimal}

rFw.d

r is the repetition count, w is field width including
sign and decimal point, and d is number of characters
to right of decimal.

Exponentia I

rEw.d

r is the repetition count, w is field width including
sign, decimal point, and d is the number of characters
in exponent.

Alphanumeric

rAw

r is the repetition count, w is field width.

H (Hollerith
or Literal)

nHcharacters
'characters'

n is total number of characters following H. Parentheses
in each format statement must balance. Characters enclosed within sing le quotes (SHIFT/7) are also printed.

Parentheses

n (specification)

format specification in parentheses is repeated n times.

Carriage
Control

indicates beginning of a new data record.

C-1

APPENDIX D
STORAGE ALLOCATION

D.1

REPRESENTATION OF CONSTANTS AND VARIABLES

D. 1. 1

Integers
Integers are each al located one machine word. They are represented in two's complement

binary.

sign

Two's complement magnitude

Positive numbers in two's complement binary are
represented as straight binary with the first bit zero.

lo I 111111111111
3777 =+2047 , the largest positive integer.
10
8
Negative numbers are represented by replacing each 0 bit with a 1 and each 1 bit with a 0,
then adding l to the binary result.
+l

o I 00000000001 I
-1

1111111111 o I

+1 = .._I1__.__1_11_11_1_11_1_11.......1

The largest negative number is -2048 which is represented by 4000
00000000000

D.1.2

=77778
or

Real Numbers
Real numbers are each allocated three machine words. They are represented as a binary

mantissa multiplied by 2 raised to a binary exponent:
Word 1

sign

exponent

D-1

mantissa

Word 2

Word 3

Imantissa

mantissa

T~e sign of the number is bit 0 of word l (O=+, l =-).

are obtained by subtracting l 0 000 000

(or 200 ) from bits l through 8 of word l.
8

Example l
110000001100

-0-0Sign:
Exponent:

1
2
l 0 000 0001

Mantissa:

. l 00

2
Exponent = 201 -200 =1
8
8 8
Mantissa = .4
8
i
No. = - . 4 X 2
8
8

= -1/2 x 2 = -1

Example 2
01000010110

-0-

-0Sign:

Exponent:

l 0 000 l Ol 2

Mantissa:

Mantissa

= .4 8

Exponent =
No.

The value and sign of the exponent

205 -200 =5
8
8 8

.4 X 25
8

1/2 x 32 = 16

D-2

D.2

STORAGE OF ARRAYS
Array variables are stored in core according to USA Standards, in columns and from top to

bottom. For example, the array IJ
DIMENSION IJ (5)
if started at location 0705 would be stored:

IJ (1) . - + - - - - - - - - I 0705
IJ (2)
0706
IJ (3)
0707
IJ (4)
0710
IJ (5)
0711
The rea I array, T
DIMENSION T (3)
starting in location 0612 would appear:
01

T ( 1)

T (2)

T (3)

1l
0612
0613
0614
0615
0616
0617
0620
0621
0622

Two-dimensional arrays are stored as shown below.
DIMENSION 1(4,2)

1(1,l)
I (2 I l)

0566
0567
0570
0571
0572
0573
0574
0575

I (3, 1)
I (4 I 1)
I ( l , 2)
I (2 I 2)
I (3,2)
I (4 I 2)

D-3

In the array
A(M(J, K) )
Mis a two-dimensional integer array stored as indicated above.

No element of M may be less than 1.

If the element
M(3, 4)
contains the integer 7, then A(M(3,4)) will be evaluated as A(7). The largest integer stored in M must
not exceed the dimensions of A.

D.2.1

Representation of N-dimensional Arrays
Although arrays of more than two dimensions are illegal, the values of the subscripts of

larger arrays may be calculated by using the fol lowing algorithm:

where the subscript values are i , i ... in in an array whose dimensions are D , D2 ... Dn.
1
1 2
Subprograms may be written to compute and insert subscript values in such ii legal arrays.
For example, in an array A(3,4,5), the following subprogram inserts the value of element A(Nl, N2, N3):

D.3

DIMENSION ARRAY (60)
READ (1,5) Nl I N2, N3, VALUE
I= N 1+3* (N2-1 )+-3*4 * (N3-1)
ARRAY(I)= VALUE
FORMAT (311IF5.3)
END

COMMON STORAGE ALLOCATION
Common storage begins in absolute location 200 in field l. Variables are assigned locations

in the common storage area in ascending order as they appear in COMMON statements.
For example:
COMMON A, JI K
DIMENSION A(2, 2), J(4)
would be stored as follows.

D-4

A(l I 1)

A(2, 1)

200
201

202
203
204
205
206
207
210

A(l, 2)
I

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

"> 1 1

LI I

212
213
214
215
216
217
220

I
NOTE

K does not appear in a DIMENSION statement.

If the COMMON statement of another subprogram defines
COMMON J
DIMENSION J(5)
J(l) through J(5) will be assigned to locations 1000 through 1004 respectively, thus overlapping the
variables A(l, 1) and A(2, 1 ). The Loader is not aware of this, therefore it is advisable to make COMMON
statements identical in all subprograms in which they appear.
However, the statements
COMMON DUMMY,J
DIMENSION DUMMY(2,2),J(4)
would not produce overlapping common and could be used in subprograms. In the example above,
DUMMY is an arbitrary variable which need not be used in the subprogram.

D-5

APPENDIX E
ERROR MESSAGES

E. 1

COMPILER
When an error is encountered during compilation of a statement, the incorrect statement

and an error message is printed. Further compilation of that statement is terminated, and output is
suppressed for the rest of the compi la ti on. The compiler, however, wi 11 scan the remaining statements
for errors, and wi II print an error message for any errors found.
An example of an error message fol lows:

A=B+M(6) +N(l)

MIXED MODE EXPRESSION
Note that an twas printed directly below the incorrect statement. This indicates that the error occurred
somewhere between that point and the beginning of the statement. In some cases the arrow may point
directly at the illegal character or word, but this cannot always be assumed.

If an error occurs in the middle of a series of continuation lines, all iemaining lines in that
statement will be printed with the error message ILLEGAL CONTINUATION.
Compiler error messages are self-explanatory:
ILLEGAL CONTINUATION

ILLEGAL VARIABLE

ILLEGAL ARITHMETIC EXPRESSION

ILLEGAL OR EXCESSIVE DO NESTING·

ILLEGAL STATEMENT

ARITHMETIC EXPRESSION TOO COMPLEX

ILLEGAL CONSTANT

MIXED t-AODE EXPRESSION

ILLEGAL STATEMENT NUMBER

EXCESSIVE SUBSCRIPTS

SYMBOL TABLE EXCEEDED

ILLEGAL EQUIVALENCING

SYNTAX ERROR (usually illegal
punctuation)

E .2

SABR
Because SABR is a 1-pass automatic paging assembler for binary relocatable pmgrams,

errors are somewhat difficult to handle. If there are errors in the source, the assembled binary code
will be virtually useless. Both errors E and Sare fatal. Assembly halts when they are encountered.
The other types of errors are not fatal, but they cause the line in which they occur to be treated as a
comment and thus essentially ignored. An address label on such a line wi II remain undefined and no
space is reserved in the binary output for the erroneous data.

E-1

During the assembly pass error diagnostics are typed on the teletype as they occur.
Example:

+0004

\10

This means that an error of type C has occurred at the 4th instruction after the location tag \10.
This would correspond to statement 10 in the source program.
During the listing pass the error letter is typed in the address field of the instruction line.
The following error diagnostics may occur.
A
C

means that too many or too few ARG 's follow a CALL statement.
means that an i !legal character appears on the line. This could possibly be an
8 11 or 11 9 11 in an octal digit string or an alphabetic character in a digit string.

means that a symbol is multiply defined. It is impossible to resolve multiple
definitions during Pass 2. Therefore, listings of programs which contain multiple
definitions wi II necessarily have unmarked errors. The M flag occurs only during
Pass l .
means that an illegal syntax has been used. Below are listed the types of illegal
syntax that may occur.
(1)
(2)
(3)
(4)
(5)

A pseudo-op with improper arguments.
A quote mark with no argument.
A non-terminated text string.
A memory reference instruction with improper address.
An illegal combination of microinstructions.

means there is no END statement.

means either one of two things:
(1)

The symbol table has overflowed. This can be corrected by
using fewer symbols, using more shorter symbols, or by
breaking the program into smaller parts.

(2)

Common storage has been exhausted.

One further type of error may occur. This is an undefined symbol. Because SABR is a onepass assembler, undefined symbols cannot be determined unti I the end of the assembly pass, so the error
diagnostic UNDF is given in the symbol table listing.

E .3

LINKING LOADER

If during the process of loading a program or subprogram the Linking Loader encounters an
error, the user is notified by an error message; the partially loaded program or subprogram is ignored,
removed from the field, and core is freed. The error messages are typed out in the form
ERROR
where n is the error code number.

E-2

OOOn

Error Code

Explanation
More than 6410 subprogram names have been seen by the
Loader (64 subprogram names is the capacity of the
10
Loader's symbol table).

The current field is ful I.

The current subprogram has too large a COMMON storage
assignment. (Subprogram with largest common storage
declaration must be loaded first.) This is a semi-fatal error.
Re-initialize the Linking Loader as explained below and
reload the programs in the proper order.

Checksum error on input tape. If the error persists 1 reassembly is necessary.

Illegal Relocation Code has been encountered. This can
occur only if the relocatable binary tape is bad or if the
user is using it improperly, e.g., not starting at the beginning of the tape 1 or reader error 1 or punch error. If the error
persists 1 re-assembly is necessary.

Recovery from Errors 2, 4, 5 is accomplished by repositioning the tape in the reader to the
leader code at the beginning of the subprogram and then pressing CONTinue. When attempting to
recover from one of these errors 1 no other program should be loaded before re loading the program which
caused the error. Obviously 1 on Error 2 a different field should be selected before pressing CONTinue.
The entire ioading process may be restarted via the consoie switches 1 at any ti me by reinitializing the Linking Loader. To do this 1 set the console switches as follows: Data Field = h (the
field where the Linking Loader resides), Instruction Field = h, Switch Register = 6200; then press START.

E.4

LIBRARY PROGRAM
During execution the Library programs check for certain errors and type out the appropriate

error messages in the form
11

XXXX 11 ERROR AT LOC NNNN

where XXXX specifies the type of error 1 and NNNN is location of the error. When an error is encountered 1 execution stops, and the error must be corrected •
When multiple error messages are typed, the location of the last error message is relevant
to the user program. The other error messages are to subprograms called by the statement at the
relevant location.

E-3

Explanation

Error Code
"ALOG"

Attempt to compute log of negative number

ATAN 11

Result exceeds capacity of computer

01vz

Attempt to divide by 0

EXP

FIPW 11

FMT1

Result exceeds capacity of computer
Error in raising a number to a power
Multiple decimal points

"FMT2 11

E or • in integer

FMT3 11

Illegal character in I, E, or F field

FMT4

FMT5 11

FLPW

FPNT 11

Multiple minus signs
lnvali d FORMAT statement
Negative number raised to floating power
Floating-point error may be caused by: Division by zero;
floating-point overflow; attempting to fix too large a number

SQRT

Attempt to square root a negative number

To pinpoint the location of a Library execution error:
a. From the storage map, determine the next lowest numbered location (external symbol)
which is the entry point of the program or subprogram containing the error.
b. Subtract in octal the entry point location of the program or subroutine containing the
error from the LOC of the error in the error message.
c. From the assembly symbol table, determine the relative address of the external symbol
found in step a and add that relative address to the result of step b.
d. The sum of step c is the relative address of the error, which can then be compared with
the relative addresses of the numbered statements in the program.

E-4

APPENDIX F
OPERATING PROCEDURES

This appendix is a condensation of Chapter 8. The figures referenced (in parentheses) are
found in the PDP-8/I System User's Guide, DEC-08-NGCB-D.

F. 1

LOADING THE FORTRAN COMPILER

Load RIM and BIN Loaders into Field 1 (Figure RIM-1, 2, 3, and BIN-1).

2. Load the FORTRAN Compiler using BIN (Figure BIN-2); IF=l SR=7777. When loaded,
parts of the Compiler wi II be in Field 0 and Field 1.
To load the Compiler on the disk, proceed in step sequence, otherwise, proceed at
step 5, below.
3.

With the Disk Monitor in memory, call the Disk System Loader by typing:
.LOAD

and load the FORTRAN Compiler onto the disk (see PDP-8/I Disk/DECtape Monitor System,
DEC-D8-SDAB-D.)

Save the compiler by typing:
.SAVE FTC0!0-7577: 5363
. SAVE FTC 1 ! 200 I 1000-1577 I 2600 I 6000-16377;

F .2

COMPILING (Pass l)
5.

Set DF = 0, IF= l , SR= 1000

6. Place FORTRAN source program tape in reader, press punch ON, LOAD ADD, and
then START; compi lotion commences.
7.

Error message? Either proceed or correct program and recompile.
Compiler will punch compiled tape in two sections, separated by a noticeable length

of blank tape.
8. More source program tapes to be compiled? Yes: insert source tape and press CONT.
No: proceed to next step .

NOTE
The FORTRAN Compiler may be restarted at any time
by pressing STOP and proceeding at step 5.

F-1

F .3

LOADING THE SABR ASSEMBLER

9. Load the SABR Assembler using BIN (Figure B-2); IF= l, SR =7777. When loaded,
parts of the Assembler will be in Field 0 and Field 1.
To load the Assembler on the disk, proceed in step sequence, otherwise, proceed at
step 12, below.
l 0.

Same as step 3, above .

11 .

Save the Assembler by typing:
.SAVE SABR!0-7177; 200
.SAVE SABl ! 12000 - 12427;

F .4

ASSEMBLING (Pass 2)
See Section 8 .4 for alternate methods of assembling.
12.

Insert Section 1 (the last section punched) of the compiled tape into the tape reader.

13.

Set DF=O, IF=O, SR=0200, press LOAD ADD, START, and answer SABR 1s initial

14.

Turn the appropriate punch and reader ON; the tape reads in and the binary tape is

dialogue.
punched.
15. Insert Section 2 (the first section punched) of the compiled tape into the tape reader
and press CONT; assembly is completed when SABR halts after producing the relocatable binary tape.
SABR may be restarted to assemble another program by starting over at step 12 above.
SABR may be restarted at any time by pressing STOP, setting the SR=0200, and pressing
LOAD ADD and then START.
To generate an assembly listing, proceed in step sequence, otherwise, proceed at step
18below.

F .5

16.

Insert Section 1 of the compiled tape into the reader and press CONT.

17.

Insert Section 2 of the compiled tape into the reader and press CONT.

LOADING THE LINKING LOADER
18.

Set DF =highest field in the configuration, IF= 1, SR =7777, and press LOAD ADD.

19. Insert Linking Loader tape into the appropriate reader: if ASR reader, turn reader ON;
if high-speed reader, set SR =3777.
20.

F .6

Press START; the Linking Loader will be read into core memory.

LOADING PROGRAMS AND SUBPROGRAMS
21.

Set DF and IF=to DF in step 18 above, SR=0200, and press LOAD ADD.

22. Insert relocatable binary tape (first, program or subprogram with largest amount of
COMMON storage) into the reader with leader code over reader head.

F-2

23.

Set SR as explained in Section 8.5.

24.

Press START; the relocatable binary program will be loaded into core memory.

Repeat from step 22 for subsequent program or subprogram tapes or select an option
(Core Availability or Storage Map) as explained in Section 8 .5.

F .7

EXECUTING THE FORTRAN PROGRAM

25. Set DF and IF =to fl eld of MAIN program, and SR =to starting address of MAIN program
{determined from the Storage Map).
26.

Turn punch 0 N and/or insert data tape in reader, as required.

27.

Press LOAD ADD and START.
Program execution wi II begin.

F-3

APPENDIX G
IMPLEMENTATION NOTES

G .1

INPUT/OUTPUT

G. l .1

Implied DO Loops
Because of core memory restrictions, 8K FORTRAN does not have implied DO loops in READ

and WRITE statements. However, a simple way to circumvent this restriction has been implemented.
Normally a carriage return/line feed (CR/LF) is produced at the end of each WRITE statement. The
CR/LF can easily be suppressed by terminating the WRITE statement with a comma. The CR/LF can
be generated explicitly in one of two ways:
a.

By using a WRITE (d,f) instruction.

By using a FiNi pseudo instruction.

The second method is more efficient since it generates only 4 words of code, whereas the
first method wi 11 generate somewhat more than that. For example, the fol lowing statements:

DO 10 J = 1,M
10 WRITE (l ,20) (A(J I K) I K = l, N)
20 FORMAT (10F7 .3)
which is not legal in SK FORTRAN, could be rewritten as follows:

10
15
20

DO 15 J=l,M
DO 10 K= l ,N
WRITE (l ,20) A(J I K) I
WRITE (l ,20)
FORMAT (F7.3)

10
15
20

DO 15 J = 1,M
DO 10 K =l ,N
WRITE (1 ,20) A(J I K) I
FINI
FORMAT (F7 .3)

The second method is preferred for more efficient utilization of core memory. Note that it is not
necessary to specify a repetition count in the FORMAT statement since the I/O handler initializes
itself to the beginning of the FORMAT statement each time the WRITE statement is executed.

G.1.2

FORMAT Handling
For more complicated FORMAT handling a somewhat different technique can be used. For

example,

G-1

WRITE (1 ,20) (A(K), K = 1, N)
FORMAT (F7 .2 ,2E15 .6)

which again is not legal in SK FORTRAN, could be written as follows

10
20

WRITE (1,20),
{comma suppresses CR/LF)
DO 10 K = l ,N
CALL IOH(A(K))
FINI
FORMAT (F7.2,2El5.6)

In the example above, the statement WRITE (1 ,20), generates the following assembly code
CALL 2, WRITE
ARG (1
ARG \20
The statement CALL IOH (A(K)} will generate code to call the subscripting routine SUBSC and will
then generate the following code
CALL 1, JOH
ARG [O
where [O is a temporary location generated by the compiler. Finally the FINI pseudo instruction wi II
generate the fol lowing
CALL 1, JOH
ARG 0
which wi II cause execution of the WRITE statement to be completed.
Although only WRITE statements have been shown in the previous examples, the same
techniques apply equally well to READ statements.

G.1.3

Numeric Input Conversion
In general, numeric input conversion is compatible with most other FORTRAN processors.

A few exceptions are listed below:
a. Blanks are ignored except to determine what field digits fall in. Thus numbers are
treated as if they were right justified within a field. In an F5 .2 format, the following
bbbl2
12bbb
. 12bb
00012
would all be read as the number 0.12.
b. A null line delimited by two CR/LFs will be treated as a line of blanks, and blanks wi II
be appended to the right of a Ii ne (if necessary) to fi II out a FORMAT statement. Th us
l 2(CR/LF)
12bbb
bbbl2

G-2

would all be identical under an F5.2 format. If an entire line is blank, numeric data from that line
wi I I be read as zeros .
c.

No distinction is made between E and F format on input. Thus

100.
100E2
1.E2

10000
would al I be read identically under either an F5 .2 or E5 .2 format.

G .1.4

Alphanumeric Data Within FORMAT Statements
Alphanumeric data may be transmitted directly from the FORMAT statement by two different

methods: H-conversion or the use of single quotes.
Hollerith (H) format is used in WRITE statements only. An attempt to use H format specifications with a READ statement wi 11 cause characters from the format field to be either typed or punched.
This may occasionally be a usefui feature since it provides a simple way of identifying data that is to
be read from the Teletype. For example, the following instructions

READ (1,30) A,B
FORMAT (4HAb =b, F7 .2/4HBb =b ,F7 .2)

would cause a =and B =to be typed out before the data was read.
The same effect is achieved by merely enclosing the alphanumeric data in single quotes.
The result is the same as in H-conversion; on input, the characters between the sing le quotes are replaced by input characters, and, on output, the characters between the single quotes {including blanks)
are written as part of the output data. For example, when referred to from a WRITE statement,
FORMAT ('PROGRAM COMPLETE')
would cause PROGRAM COMPLETE to be printed. This method eliminates the need to count characters.

G.1.5

E and F Format
When using the WRITE statement with either E format or F format with numbers less than

1 .0 a zero wi II not be typed to the left of the decimal point.

G.2

ARITHMETIC OPERATIONS

G .2.1

Floating-Point
In general, floating-point arithmetic calculations are accurate to seven digits with the

eighth digit being questionable. Subsequent digits are not significant even though severa I may be
typed to satisfy a field width requirement.

G-3

No definitive information is currently available on the accuracy of the functions except that
they are believed to be accurate to six decimal places for arguments which are neither extremely large
nor extremely small.
The floating-point arithmetic routines check for both overflow and underflow. Overflow
wi II cause the FPNT error message to be typed and program execution wi II be terminated. Underflow
is detected but wi II not cause an error message. The arithmetic operation involved wi II yield a zero
result. The arctangent function is accurate to six decimal places for arguments whose absolute value
is greater than .01. This is a temporary restriction.

G .2.2

Integer
Integer arithmetic operations do not check for overflow. For example, the sum 2047+2047

wi II yield a result of -2. For more information refer to Chapter l of Introduction to Programming
{Small Computer Handbook Series) or any text on binary arithmetic.

G.2.3

Exponentiation
Zero raised to a power of zero wi II yield a result of l. Zero raised to any other power

wi II yield a zero result. Numbers are raised to integer powers by repetitive multiplication. Numbers
are raised to floating-point powers by calling the EXP and ALOG functions. A negative number
raised to a floating-point power wi II not cause an error message but wi II use the absolute value. Thus,
the expression (-3.0)**3.0 will yield a result of +27.

G. 3

SUBSCRIPTING
Since excessive subscripting tends to use core memory inefficiently, it is suggested that

subscripted variables be used judiciously. For example, the statement
A= ( (B (I) + C2) * B(I) +Cl) *B(I)
could be rewritten with a considerable saving of core memory as follows:
T = B(I)

A = ( (T + C2) *T + C 1) *T

G.4

DO LOOPS
DO loops are treated slightly differently in 8K FORTRAN than in most compilers. The

index is tested before the range of the DO is executed. Therefore, in the following example

G-4

DO 20 N = 1, M

20 CONTINUE
the instruction between the DO statement and statement 20 wi II never be executed if M is less than one.

G .5

PAUSE STATEMENT
The PAUSE statement may be used for a variety of reasons to temporarily suspend program

execution. In some cases the PAUSE statement may be used to give the operator a chance to change
data tapes or to remove a tape from the punch. When this is done it is necessary to fol low the PAUSE
statement with a call to the OPEN subroutine. This subroutine initializes the I/O devices and sets
hardware flags that may have been cleared by pressing the tape feed buttons. Example:
PAUSE
CALL OPEN

G .6

EQUIVALENCE STATEMENT
Because of core memory restrictions within the compiler, variables may not appear in

EQUIVALENCE statements more than once. Thus,
EQUIVALENCE (A, B,C)
would be valid, but the statement
EQUIVALENCE (A,B), (B,C)
would not compile correctly.

G. 7

SPECIAL I/O DEVICES
I/O can be performed on devices other than Teletype and high-speed paper tape reader and

punch in severa I different ways:
1. If it is desired to use other devices in place of the high-speed paper tape reader and
punch, rewrite the Utility library subroutine defining the entry points for the desired input and output
devices as HSIN and HSOUT respectively. The source tape for the Utility subroutine is available from
the program library and is very short. Refer to the SABR manual for more information.
2. If it is desired to input or output on a special device but not in ASCII format, write a
subroutine to handle the particular device in the SABR assembly language. For more information refer
to the SABR manual.
3. If it is desired to add additional devices which can be used with READ and WRITE statements, then edit part I of the Library Subroutines IOH. New entries must be made in the device
transfer table at the beginning of IOH. Copies of this source tape and listings of the library subroutines
are available from the program library. The service routines for the additional I/O devices must be
written in SABR assembly language and can then be assembled along with the revised version of IOH.

G-5

Program written in SABR language can call PAL subroutines in various ways:
a)
A JMS 7000 instruction will call a PAL program which starts at location 7000
in the same memory field.
b)
A CONTINUE (or PAUSE) statement might be inserted in the user's FORTRAN
program. Then a JMS to the PAL subroutine may be inserted using the Switch Register.

It is possible to load any size PAL III program for linkage with an 8K FORTRAN program
by merely dimensioning an integer variable to the proper size for the PAL III program. This offers
two advantages, virtually unlimited size programs in PAL III can be linked to BK FORTRAN main
programs, and none of the library routines are disturbed by this linkage.

G .8

ERRORS
All compile time, assembly time, and execution time errors are fatal. For this reason it is

desirable to suppress punched output of the compiler and assembler unti I the source program is believed
to be correct. For specific instructions refer to Chapter 8.
Note especially that the compiler will not detect undefined statement numbers. Therefore
it is important to examine the assembly symbol table for undefined symbols before loading and executing
the program •
Do not attempt to load or run a program which has assembly errors. Do not attempt to
proceed after an execution time error by pressing CONTinue. Unpredictable results wi 11 be obtained
in either case.

G-6

HOW TO OBTAiN REVISIONS AND CORRECiIONS

Notification of new or revised DEC software and manuals available from the Program
Library is published in:
Digita I Software News for the PDP-8 Family
Digita I Software News for the PDP-9 Family

if you are not receiving the publication appropriate to your computer, please notify Software Information Service (see Reader's Comments card).
Revised software products and documents are shipped only after the Program Library receives
a specific request from a user (see title page for address).
Digital Equipment Computer User's Society (DECUS) maintains a library of user software
and publishes them in DECUSCOPE, a magazine available to both DECUS members and to non-members
who request it. Return the request card below to receive further information or to place your name on
the ma i Ii ng Ii st.

I
I
I
I

To: Dec us Office,
Digital Equipment Corporation,
Maynard, Massachusetts 01754

D Please send DECUS installation membership information.
D Please send DECUS individual membership information.
D Please add my name to the DECUSCOPE non-member mailing list.
Name
Company
~~~~--.....~~~~~~~---

Address

8K FORTRAN
PROGRAMMERS REFERENCE
DEC-08-KFXB-D
READER'S COMMENTS

Digital Equipment Corporation maintains a continuous effort to improve the quality and usefulness of its
publicafions. To do this effectively we need user feedback -- your critical evaluation of this manual.
Please comment on this manual's completeness, accuracy, organization, usability, and readability.

Did you find errors in this manual? Please explain, giving page n u m b e r s . - - - - - - - - - - - -

How can this manual be improved?---------------------------~

DEC also strives to keep its customers informed on current DEC software and publications. Thus, the
following periodically distributed publications are available upon request. Please check the publication(s) desired .

.
: 0 Digita I Software News for the
:

PDP-8 Family, contains current
information on software prob Iems,
programming notes, new and revised software and manuals.

0 PDP-8/I Software Manual
Update, contains addenda/
errata sheets for updating
·software manua Is.

0 PDP-8/I User's
Bookshelf, contains
a bibliography of
current and forthcoming software
manuals.

Please describe your position.-----------------------------~
Name - - - - - - - - - - - - - - - O r g a n i z a t i o n - - - - - - - - - - - - - - - - - - Street -----------------Department _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _~
City _ _ _ _ _ _ _ _ _ _ _ _ State _ _ _ _ _ _ _ _ _ _ _ __ Zip or Country------

························································································· Fold Herc .......................................................................................... :

............................................................................ Do Not Tear - Fold Here and Staple ............................................................... ..

FIRST CLASS
PERMIT NO. 33
MAYNARD, MASS.
BUSINESS REPLY MAIL
NO POSTAGE STAMP NECESSARY IF MAILED IN THE UNITED STATES

Postage will be paid by:

mamaama
Digital Equipment Corporation
Software Information Services
146 Main Street, Bldg. 3-5
Maynard, Massachusetts 01754

Digital Equipment Corporation
Maynard, Massachusetts

printed in U.S.A.

mnmnomn