Digital PDFs

DEC-08-OCOSA-E-D

March 1973

369 pages

Original

10MB

Document:	COS 300 System Reference Manual 197303
Order Number:	DEC-08-OCOSA-E-D
Revision:
Pages:	369
Original Filename:	http://bitsavers.org/pdf/dec/pdp8/cos-300/DEC-08-OCOSA-E-D_COS_300_System_Reference_Manual_197303.pdf

OCR Text

COMMERCIAL OPERATING SYSTEM

SYSTEM
REFERENCE
MANUAL

DEC-08-0COSA-E-D

c0 s

300

S YS T E M

RE F E RE NC E MANUA L

SOFTWARE SUPPORT CATEGORY
The software described in
this document is supported
by Digital Equipment Corporation under Category I
,
as defined on page iii of
this document.

For additional copies, order No. DEC-08-0COSA-E-D
from Software
Distribution Center, Digital Equipment Corporation, Maynard,
Massachusetts 01754

First Printing, Feb.
Second Printing, Mar.
Third Printing, Apr.
Fourth Printing, Jun.
Revised, Mar.
Your attention is invited to the last two pages of
this document. The "How to Obtain Software Information" page tells you how to keep up-to-date with
DEC's software. The "Reader's Comments" page, when
filled in and mailed, is beneficial to both you and
DEC; all comments received are acknowledged and are
considered when documenting subsequent manuals.
Copyright

@ 1972, 1973 by Digital Equipment Corporation

The material in this document is for information purposes and is subject to change without
notice.

The following are trademarks of Digital Equipment
Corporation, Maynard, Massachusetts:
CDP
Computer Lab
Comtex
DEC
DECtape
Dibol

Digital
DNC
Flip Chip
IDAC
Indac
KAlO

LAB-8/e
OMNIBUS
OS/8
PDP
PHA

PS/8
Quickpoint

RAD-8
RSTS
RSX
RTM
SABR
Typeset 8
Unibus

1972
1972
1972
1972
1973

SOFTWARE

SUPPORT

CATEGORIES

Digital Equipment Corporation (DEC) makes avai Iable four categories of software. These
categories refiect the types of support a customer may expect from DEC for a specified software
product. DEC reserves the right to change the category of a software product at any time.
The four categories are as fol iows:

Software

CATEGORY
Products Supported

Charge

This classification includes current versions of monitors, programming languages, and
support programs provided by DEC. DEC wi 11 provide ins ta Ilat ion (when applicable), advisory,
and remedial support at no charge. These services are limited to original purchasers of DEC
computer systems who have the requisite DEC equipment and software products.
At the option of DEC, a software product may be recategorized from Category I to
Category II for a particular customer if the software product has been modified by the customer
or a third party.

Software

Products

CATEGORY II
that Receive Support

for

Fee

This category includes prior versions of Category I programs and all other programs available from DEC for which support is given. Programming assistance (additional support), as
available, wi II be provided on these DEC programs and non-DEC programs when used in conjunction with these DEC programs and equipment supplied by DEC.

CATEGORY Ill
Pre-Release Software
DEC may elect to release certain software products to customers in order to facilitate
final testing and/or customer familiarization. In this event, DEC will limit the use of such
pre-release software to internal, non-competitive applications. Category Ill software is only
supported by DEC where this support is consistent with evaluation of the software product.
While DEC will be grateful for the reporting of any criticism and suggestions pertaining to a
pre-release, there exists no commitment to respond to these reports.

CATEGORY IV
Non-Supported Software
This category includes all programs for which no support is given

iii

PREFACE

This manual provides a
writing and operating
300 environment.

reference source for users interested in
DIBOL language and system programs in the COS

Chapter l
summarizes the use of the DIBOL language in program
preparation.
Chapters 2, 3, 4 and 5 cover the main system programs
which handle the loading and editing of programs, software system
configuration and logical unit assignment, and movement 0£ files and
programs between devices. Chapters 6, 7 and 8 contain the utility
programs which are useful in building, sorting and maintaining data
files. Chapters 9-14 contain additional utility programs.
Appendices A through
summaries.

provide

additional

reference

material

A glossary is provided at the end of this manual for those
are unfamiliar with the terms used.

users

and
who

The system analysts, planners and user programmers who read this
manual are expected to have a basic knowledge of DIBOL. Additional
background in DIBOL and the COS 300 system can be obtained from the
COS 300 Self-Teaching Manual (DEC-08-0CSTA-A-D).

CONTENTS
Page
PREFACE
INTRODUCTION

INTR0-1
PART I

DIBOL LANGUAGE

DIBOL LANGUAGE
CHAPTER 1
STATEMENT NUMBERS
1.1
1.2
STATEMENT LABELS
COMMENTS
1.3
STATEMENTS
1. 4
l. 4.1
ACCEPT
1. 4. 2
CALL
1. 4. 3
CHAIN
1. 4. 4
Data Manipulation Statements
1.4.4.l Clearing Fields
1.4.4.2 Moving Alphanumeric Data
1
A A ~
Moving DecL~al Data
1.4.4.4 Moving Records
1.4.4.5 Coftverting and Formatting Data
1. 4. 5
DISPLAY
1. 4. 6
END
1. 4. 7
FINI
1. 4. 8
FORMS
1. 4. 9
GO TO
1.4.9.l Unconditional GO TO
1.4.9.2 Computed GO TO
1. 4 .10
IF
1.4.11
INCR
1. 4 .12
INIT
1. 4 .13
ON ERROR
1. 4 .14
PROC
1. 4 .15
READ
1. 4 .16
RECORD
1.4.17
RETURN
1. 4 .18
START
1. 4 .19
STOP
1. 4. 20
TRACE/NO TRACE
1.4.21
TRAP
l.4.22
WlUTE
XMIT
1. 4. 23
.L.e-Z•~•....,

PART II
CHAPTER 2
2.1

2.2
2.2.1
2.2.2
2.2.3

2.2.4
2.2.5
2 .. 2. 6
2.2.7

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

1-4
1-5
1-7
1-9
1-11
1-14
1-15
1-15

1-16
1-16
1-19
1-23
1-25
1-27
1-29
1-29
1-29

1-31
1-33
1-35
1-39

1-41
1-43
1-45
1-59

1-61
1-63
1-65
1-67
1-.71
1-73

SYSTEM AND UTILITY PROGRAMS

MONITOR
OPERATING PROCEDURES
MONITOR COMMANDS
BATCH
DATE
DELETE
DIRECTORY
PLEASE

RUN
SAVE

Vii

2-1
2-1
2-3
2-5
2-7
2-9
2-11
2-13
2-15
2-19

2.3
2.3.1
2.3.2
2.3.3
2.3.4
2.3.5
2.3.6
2.3.7
2.4
2.5
2.6

EDITING COMMANDS
ERASE
FETCH
LIST
Line Number
Number Conunands
RESEQUENCE
WRITE
EDITING EXAMPLE
BATCH EDITING
ERROR MESSAGES

2-20
2-21
2-23
2-25
2-27
2-31
2-33
2-35
2-36
2-38
2-40

CHAPTER 3
3.1
3.2
3.3
3.3.1
3.3.2
3.4

SYSGEN
OPERATING PROCEDURES
SYSTEM SOFTWARE CONFIGURATION
LOGICAL UNIT ASSIGNMENT
DECtape Users
Disk Users
ERROR MESSAGES

3-1
3-1
3-1
3-4
3-6
3-6
3-8

CHAPTER 4
4.1
4.2
·4. 3

COMP
OPERATING PROCEDURES
CONDITIONAL COMPILATION PROCEDURE
ERROR MESSAGES

4-1
4-1
4-7
4-8

CHAPTER 5
5.1
5 .1.1
5 .1. 2
5.1. 3
5 .1. 4
5.1. 5
5 .1. 6
5 .1. 7
5.2

PIP
OPERATING PROCEDURES
Transfer Binary File
Copy Device
Transfer Data Files
Eliminate Spaces in Directory
Transfer Source Files
Transfer System Program
Return to Monitor
ERROR MESSAGES

5-1
5-1
5-3
5-5
5-7
5-9
5-11
5-13
5-15
5-16

CHAPTER 6
6.1
6 .1.1
6 .1. 2
6 .1. 3
6 .1. 4
6.2
6.2.1
6.2.2

BUILD
BUILD CONTROL PROGRAM
Field Descriptor Section
INPUT Section
OUTPUT Section
Storing the BUILD Control Program
OPERATING PROCEDURES
KBD - Data Input from Keyboard
SYS, CDR or RDR - Data Input from
System Device, Card, or Papertape
BUILD INPUT LINE
ERROR MESSAGES

6-1
6-1
6-3

6.3
6.4
CHAPTER 7
7.1
7 .1.1
7 .1. 2
7.2

SORT
SORT CONTROL PROGRAM
File Descriptor Section
INPUT/OUTPUT Section
SORT OPERATING PROCEDURES

viii

6-4
6-8
6-10
6-11
6-12
6-13
6-13

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

RUNNING SORT AS PART OF AN UPDATE PROCEDURE
MERGE OPERATING PROCEDURE
RULES FOR USING DEFAULT UNITS
ERROR MESSAGES

7-6
7-6
7-8
7-9

CHAPTER 8
8.1
8.1. l
8.1. 2
8.1. 3
8.1. 4
8.1. 5
8.1. 6
8.1. 7
8.2
8.3
8.4
8.5

UPDATE
UPDATE CONTROL PROGRAM
File Descriptor Section
UPDATE Statement
INPUT Statement
SORT Statement
KEY Statement
OUTPUT Statement
END Statement
UPDATE COM..1\1ANDS
UPDATE EXAMPLE
OPERATING PROCEDURES
ERROR MESSAGES

8-1
8-1
8-1
8-2
8-2
8-3
8-3
8-4
8-5
8-5
8-6
8-7
8-8

CHAPTER 9

9.1

BOOT
OPERATING PROCEDURES

9-1

CHAPTER 10
10.1
10.2

PATCH
OPERATING PROCEDURES
ERROR MESSAGES

10-1
10-1
10-4

CHAPTER 11
11.1
11. 2
11.3
11. 4

DAFT

OPERATING PROCEDURES
COMMANDS
DAFT ERROR MESSAGES
DAFT OUTPUT

11-1
11-1
11-1
11-4
11-6

CHAPTER 12
12.1
12.2

COS-OS/8 CONVERTER
OPERATING PROCEDURES
ERROR MESSAGES

12-1
12-1
12-3

7.3
7.4
7.5
7.6

9-1

CHAPTER 13
FORMAT PROGRAMS
13.1
RK8MRK
13.1.1
Operating Procedt:~re
13 .1. 2
Error Messages
RKEMRK
13.2
13.2.l
Operating Procedure
13.2.2
Error Messages
TDMARK
13.3
13.3.l
Operating Procedures
13.3.2
Error Messages
13.4
DTMARK
13.4.1
Operating Procedures
13.4.2
Error Messages

13-1
13-3
13'-3
13-3
13-7
13-7
13-7
13-9
13-9
13-10
13-13
13-13
13-14

CHAPTER 14
14.l
14. 2

14-1
14-1
14-8

CREF
OPERATING PROCEDURES
ERROR MESSAGES

APPENDIX A

COS CODES

A-1

APPENDIX B
1.0
2.0
2.1
3.0
3.1
4.0
5.0

LOADING COS
AUTOMATIC LOAD
TC08 DECTAPE AS THE SYSTEM DEVICE
SYSTEM RESTART ON DECTAPE
DISK AS THE SYSTEM DEVICE
SYSTEM RESTART ON DISK
TD8E BOOTSTRAP
PDP-12 USERS

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

APPENDIX C

SIZE OF CODE PRODUCED BY DIBOL COMPILER

C-1

APPENDIX D

DESIGNING A RECORD

D-1

APPENDIX E

BUILD CHECKDIGIT FORMULA

E-1

APPENDIX F
1. 0
1.1
2.0

HARDWARE DESCRIPTIONS
VT05 TERMINAL
VT05 Start Up Procedures
HIGH-SPEED PAPER TAPE READER
AND PUNCH UNIT
DECTAPE TRANSPORT UNIT
CARD READER
LINE PRINTERS
LP08 Line Printer
LS8-E Line Printer
RK08 DISK
Mounting and Dismounting the
RK08 Disk Cartridge
Mounting
Dismounting
Mounting and Dismounting the
RK8E Disk Cartridge
Mounting
Dismounting

F-1
F-3
F-7

APPENDIX G
1. 0
2.0
3.0
4.0
5.0

COS FILES
COS SOURCE FILES
COS DATA FILES
COS DIBOL COMPILER BINARY
COS SYSTEM PROGRAMS
SYSTEM AND DATA TAPE FORM.ATS

G-l
G-1
G-1
G-2
G-2
G-2

APPENDIX H
1. 0
2.0

DIBOL DEBUGGING TECHNIQUE (DDT)
ENTERING DDT MODE
DDT COMMANDS

H-1
H-1
H-1

APPENDIX I

DIBOL STATEMENT SUMMARY

I-1

3.0
4.0
5.0
5.1
5.2
6.0
6.1
6 .1.1
6 .1. 2
6.2
6.2.1
6.2.2

B-4
B-4
B-6

F-8
F-10
F-13
F-15
F-15
F-17
F-20
F-22
F-22
F-23
F-23
F-23
F-24

APPENDIX J

ERROR MESSAGE SUMMARY

J-1

APPENDIX K
1.0

COMMAND SUMMARY
COMMANDS FOR MONITOR FUNCTIONS
COMMANDS FOR EDITING FUNCTIONS

K-1
K-1

ADVANCED PROGRAMMING TECHNIQUES
ACCEPT AND DISPLAY
Background
Interaction of ACCEPT and DISPLAY
Statements
Simple Example Using ACCEPT and DISPLAY
Generalized ACCEPT Subroutine
DIRECT ACCESS TECHNIQUES
Background
The Reason for Direct Access
How It Works In DIBOL
Unsorted File
Sorted File
Rough Table, No Index File
Rough Table Plus Index File
Summary
DIRECT ACCESS.NOTES
XMIT Statements Used With Direct Access
Truncating a File
Appending to a File
Rewriting a File
NUMERIC FIELD VERIFICATION
CHAIN STATEMENT NOTES
Interaction of CHAIN and INIT (n,SYS)
Conununication Between CHAINs
File Status
Clearing CHAINed Records
Transferring Variable Values
Multiple CHAIN Entry Points
Miscellaneous-CHAIN Facts
DIBOL PROGRAMMING OF SYS FILES
Operating Procedures
Data Section
PROCedure Section

L-1
L-1
L-1

2.0
APPENDIX L
1. 0
l.l
1.2

1.3
1. 4
2.0
2.1
2.2
2.3
2.4
2.5
2.6

2.1
2.8
3.0
3.1
3 .1.1
3.1.2
3 .1. 3
4.0
5.0
5.1
5.2
5.2.1
5.2.2
5.2.3
5.3
5.4
6.0
6.1
6.2
6.3

K-2

L-1
L-1
L-2
L-8
L-8
L-8
L-9
L-9
L-10
L-10
L-12
L-12
L-13
L-13
L-13
L-14
L-14
L-14
L-15
L-15
L-16
L-16
L-16
L-16
L-17
L-17
L-18
L-18
L-18
L-18

GLOSSARY
TABLES

1-1
1-2
1-3
2-1
11-1

Source Program Limitations
Terminator Codes
Special Characters
Monitor Key Commands
DAFT Command Summary

1-4
1-6
1-18
2-2
11-2

FIGURES

INTR0-1
INTR0-2
INTR0-3
1-1
3-1
4-1
4-2
6-1
11-1
14-1
14-2
14-3

Business Application System
Typical Disk Operation Direct Processing
Programmer's System
Sample Program
SYSGEN Flow Chart
Compiler Listing
Storage Map Listing
Sample BUILD Control Program
DAFT Sample
Square Root Subroutine
CREF of Square Root Subroutine
Output of Square Root Subroutine

xii

INTR0-3
INTR0-4
INTR0-5
1-2
3-2
4-3
4-5
6-2
11-7
14-2
14-5
14-6

INTRODUCTION

THE LANGUAGE
DIBOL (Digital Equipment Corporation Business Oriented Language) is a
COBOL-like language used to write business application programs. The
DIBOL language consists of data definition and procedures specified in
easy to learn statements. Figures INTR0-1 and INTR0-2 illustrate a
possible application work flow using DIBOL and system programs.

THE COS SYSTEM
Basic hardware requirements are:
PDP-8, 8/E, 8/I, 8/L or PDP-12 computer with BK of core
4 DECtapes, or LINCtapes
User terminal
Line printer
For a detailed description of the basic hardware and
supported equipment, refer to Appendix F of this manual.

additional

Also provided as part of COS are the following system programs:
MONITOR/EDITOR

controls the calling and executing of all
other programs in the COS system; provides
the I/O control for the peripheral devices
and an editing capability for correcting user
programs.
This program is
referred
to
throughout the manual as Monitor.

SYS GEN

specifies the
system
assigns logical device
startup.

COMP

translates DIBOL language source statements
into a binary object program which can be run
on the COS 300 computer.

PIP

transfers data, source
between two devices.

BUILD

creates data files based on the
data fields typed by the user.

SORT

sequences records according to key characters
or fields specified by the user. Records may
be sorted into ascending {0-9 or A-Z)
or
descending (Z-A or 9-0) sequence.

UPDATE

performs maintenance on data files.
UPDATE
makes the specified changes, creates a new
file and prints a report showing all changes,
deletions and insertions.

INTR0-1

configuration
and
numbers for system

binary

files

contents

Figure INTR0-3 suggests possible usage of
DIBOL program development.

these

system

programs

COS FILE STRUCTURE
There are four types of files used in the COS
compiler binary, and system program.
Data files are completely devoted to
processed by DIBOL or system programs.

the

system:
storage

source,
of

data

data,
to

Source files contain control programs (for BUILD, SORT, etc.) or
DIBOL programs.

user

Binary files are the output of the compiler and contain user
programs translated into a code which can be executed by cos.

DIBOL

System program files include those programs
(MONITOR, SYSGEN, PIP,
BUILD, COMP, SORT and UPDATE) supplied as part of the COS package.
DIBOL source programs, control programs and compiled binary programs
can also be saved on system devices.
(Data files cannot be saved as
system files.)

INTR0-2

INVOICE
MAILING

1--....;....;..;..;.,....;........;....---1 LIST

------i--- BUILD
UPDATE,DIBOL
.___ _ _ ____, PROGRAM

This figure illustrates
the possible flow
of
work from a
source
document (such as an
invoice
or
mailing
list)
through
the
creation of a data file
to the eventual output
of a report or listing.

Figure INTR0-1.

Business Application System

INTR0-3

TRANSACTION
ENTERED VIA
CARDS

PROCESS
INPUT
PROGRAM

i------1

D~~~WR~~~~~T
SUPDATE

DIBOL

SAVE TRANSACTIONS FOR
RETENTION
CYCLE ON DISK

FILE.

DIBOL

PROCESS MASTER
FILE TO PRODUCE
REPORTS
DIBOL

I
I

e
This figure illustrates
the input
of
data
(transaction file) from
terminal,
card reader
of DECtape;
the
processing against the
master file by a user
created DIBOL
program
and
the
output
(depending on the DIBOL
program being run)
of
reports and files.

Figure INTR0-2.

Typical Disk Operation Direct Processing

INTR0-4

Transfers files,
Enlarges binary
scratch area

DIBOL

program
creation (also
SYSGEN tables
and control
programs) and
editing

Compiles
source
programs
into object
(binary) program

Dump a:nd

Cross

Fix da-c.a
files

reference
of DIBOL
program

This figure illustra-c.es the COS
system programs of particular
interest to the programmer.

Figure INTR0-3.

Programmer's System

INTR0-5

MANUAL AND SYSTEM CONVENTIONS
The symbols and terms used throughout this manual are described below.
SYMBOLS USED

EXAMPLE

Lower-case
characters

PROC n

Represent information that must
be supplied by the user, such as
values,
names
and
other
parameters.

Upper-case
characters

NO TRACE

Words or characters that must be
used exactly as shown •

• • • (_Ellipsis}_

RU prog(, ••• ,progn)

Indicates optional
of arguments.

continuation

Underscored
characters

COS MONITOR 2.1108
Indicates
system.

from

{}

START00

) (carriage
return key)

the

Braces indicate a choice of
of the items enclosed.

START[0~]

]

output

Indicates a space or blank.

RUL-IPIP

L-1

[

EXPLANATION

Brackets
feature.

indicate

one

optional

Line and command terminator •

• R COMP, PROG)

TERMS
Expressions
Are variables, constants or arithmetic expressions (made
of variables, constants and the operators#,+,-,*,/).

Filnam,pronam,label and keyword
Are used to identify names assigned to files, programs
statements, and input lines, these names may be of any
length but only the first 6 characters are recognized.
Pronam can contain any characters
comma, plus, period and minus.

INTR0-6

except

space,

slash,

PART I

DIBOL LANGUAGE

CHAPTER 1

DIBOL LANGUAGE

This chapter provides a reference source on the DIBOL language.
If
more basic information is desired, refer to the COS 300 Self-Teaching
Manual (DEC-08-0CSTA-A-D).
A DIBOL program consists of a series of
sections, Data and PFDCedure.

statements

arranged

two

The Data section contains the RECORD statements which specify the type
and size of the information to be used in the program and w;1ere it is
to be stored.
The PROCedure section consists of:
ACCEPT
CALL
CHAIN
DISPLAY
END
FINI
FORMS
GO TO
IF
INCR
INIT
ON ERROR
PROC
READ
RETURN
STOP
TRACE/NO TRACE
TRAP

WRITE
XMIT
and data manipulation statements. These statements are arranged in a
logical order for program execution.
Refer to the sample program
shown in Figure 1-1. Use the editing conunands (refer to Chapter 2) to
enter newly created DIBOL programs into the COS system.
Spaces and
tabs may be inserted in the program for clarity and are ignored by the
Compiler except when enclosed in single quotation marks (').
Statement length is limited to 120 characters and spaces.
When the
!21st character is typed, an error message LINE TOO LONG is displayed
and that line is lost.
CTRL/U deletes the entire line.
If the
terminal line is filled before the 120 characters are entered, the
Editor executes a carriage return/line feed to the terminal and the
line can be continued.

1-1

START
RECORD

RECORD

RECORD
PROC 2
LOOP,

EOF,

INBUF
STOCKN,
DESC,
UCOST,
QORDER,

'
OUTBUF,X

'
ECOST,
LABEL,

04
A25
DS
04
09
04
A25
DS
04
09
A6,P

INIT(l,IN,LABEL)
INIT(2,0UT,'OUTPUT')
XMIT(l,INBUF,EOF)
IF(STOCKN.LT.lOOO)GO TO LOOP
ECOST=UCOST*QORDER
XMIT(2,0UTBUF)
GO TO LOOP
FINI (2)
FINI(l)
STOP

END

Figure 1-1.

1.1

Sample Program

STATEMENT NUMBERS

Statement numbers are assigned manually by the user when typing in the
program or automatically by the Monitor when the program is entered
for editing (Refer to Chapter 2). These numbers control the order of
the statements in the file. The DIBOL language makes no use of these
numbers, but error messages and the DIBOL debugging refer to them.

1.2

STATEMENT LABELS

Labels are the symbolic names assigned to identify the location
statement in the procedure section of a program.
TEST, A=B

1-2

Labels consist of a sequence of letters or digits the first of which
must be a letter and only the first six characters are significant.
Labels are separated from the statement by a comma.
Statements with
labels are the transfer point for other statements such as GO TO, IF
and CALL.
Examples:
GO TO BEGIN

(

label )

BEGIN, A=B+4

GO TO (PRT,SEC,TO) ,K

PRT, XMIT(lS,RPT,EOF)

( label

READ ( 10, CUST 1 8_8)

TOT, SUB=COST+TAX

c
l. 3

label

COMMENTS

Comments and notes to explain the source program may be added
following a semicolon (;) on any statement line. A program line that
begins with a semicolon contains only comments.
Examples:
RECORD CUST ; THIS IS THE CUSTOMER RECORD
; THIS PROGRAM PRINTS THE ACCOUNTS
;PAST DUE REPORT
Table 1-1 summarizes DIBOL source program limitations.

1-3

TABLE 1-1.

1.4

SOURCE PROGRAM LIMITATIONS

Maximum characters

about 8,000 per file

maximum number of
source files per
program

Maximum number of
symbols in 8K
systems

365

in 12K and
larger:

511

Line numbers
available

0-4095

Maximum characters
per line

120

STATEMENTS

There are six types of statements in DIBOL:
1.

Compiler statements
(START, END and PROC) which label
the
beginning, end and procedure section of the program. These
statements are non-executable and START and END are optional.

Data specification statements (RECORD) which describe the type and
size of data to be stored and the location in which it is to be
stored.

Data manipulation statements and !NCR, which control the
of data within memory.

Control statements, (GO TO, IF, CALL, ON ERROR, RETURN, TRAP,
CHAIN, STOP) which govern the sequence of execution of statements
within a program. Without these statements, the program would be

movement

executed in the order written.
S.

Input/Output statements (ACCEPT, DISPLAY, !NIT, XMIT, READ, WRITE
and FINI) which control data movement within memory or between
memory and peripheral devices, and to open and close the files
used by the program.
Line spacing to the line printer is
controlled by the FORMS statement.

Debugging statements, (TRACE, NO TRACE) which
program execution.

1-4

permit

tracing

1. 4. 1

The ACCEPT statement takes input from the keyboard, stores it in the
specified alphanumeric field and causes the decimal equivalent of the
last key typed to be stored in the specified decimal field. ACCEPT is
primarily used with the DISPLAY statement (Section 1.4.5) and has the

form:
ACCEPT (dfield, afield)
where
df ield

is a decimal
equivalent of
stored.

afield

is an alpha variable indicating
keyboard input is to be stored.

variable where the decimal
the last key typed is to be
where

the

The code stored in the decimal field depends on the last key typed
(special terminator character). The keys and equivalent codes for the
allowable terminator characters are shown in Table 1-2. For example,
when all characters are entered, filling up afield, the code 00 is
stored in the specified decimal field, which should be defined as two
or more characters.
This code can be tested later in the DIBOL program.
When an ACCEPT statement is encountered during program execution, the
system waits for keyboard input.
Execution continues when the
characters typed fill the afield or when a special character such as
carriage return, line feed, altmode, rubout, CTRL/U, etc. is typed.
Any non-COS character terminates the ACCEPT and sets df ield to the
decimal equivalent of the last character typed. When the ACCEPT is
terminated before afield is full, the remaining character positions in
the afield are unchanged.
Examples:
Legal

Illegal

ACCEPT (A, B)
ACCEPT(A(3),B{4,5))

ACCEPT ( 5 , CHAR)
ACCEPT(TCHAR,CHAR+3)

See Appendix L for more examples.

1-5

TABLE 1-2.
Decimal
Equivalent
00
01
02
04

05
06
07

09
10
11

12
13
16
17
18

19
20
21
22
23
24

25
26
27
28
29
30
31
32
61
62
63
94

TERMINATOR CODES

Teletype Console
Character

VT05 Console
Character

NULL OR END OF FIELD
CTRL/A
CTRL/B
CTRL/D
CTRL/E
CTRL/F
CTRL/G
CTRL/H
CTRL/I OR TAB
CTRL/J OR LINE FEED
CTRL/K
CTRL/L
CTRL/M OR CARRIAGE RETURN
CTRL/P
CTRL/Q
CTRL/R
CTRL/S
CTRL/T
CTRL/U
CTRL/V
CTRL/W
CTRL/X
CTRL/Y
CTRL/Z
ALT MODE
CTRL/\
CTRL/]
CTRL/t
CTRL/+
RUBOUT
ALT MODE
ALT MODE

or CTRL/H

tor CTRL/K
CTRL/M, CARRIAGE
RETURN, OR CTRL/=

CTRL/X or

ALTMODE or CTRL/[
HOME or CTRL/]
CTRL//\
not used
not used
/\
not used
not used

ESCAPE
PREFIX
NOTE
Unless otherwise indicated,
codes same as for Teletype.

1-6

VT05

characters

and

ICl\LL
1. 4.2

CALL

The CALL statement causes control to branch to a
the form:
CALL

subroutine

and

has

label

where label is the label of the first
PROCedure section of the program~

state~ent

a subroutine in the

The CALL statement saves the return location in a
(pushdown)
list.
Additional subroutine CALL and RETURN statements may be nested within
a subroutine to a depth of SO.
Examples:
Legal

Illegal

CALL SET

CALL
CALL (SUBR+3)

SET, PROFIT=PRICE-COST

RETURN

1-7

fcHAIN'

1.4. 3

CHAIN

The CHAIN statement allows a DIBOL program which will not
available memory to be split up into two or more programs.

fit

The form of the CHAIN statement is:
CHAIN decimal expression
where decimal expression refers to the position of the
in the Monitor RUN command and is in the range 0-7.

CHAIN

program

When the CHAIN .statement is encountered in a DIBOL program, execution
of the current program is halted and the CHAIN program referred to by
the decimal expression in the CHAIN statement is loaded. Execution of
the CHAINed program begins with the statement immediately following
the PROC statement. All CHAIN programs must be properly declared in a
.RUN command. The program initially run is chain O, the next program
chain 1, etc. For exanlple, a CHAIN 2 statement refers to program TRY2
in the .RUN command •
• RUN START+HELPl+TRY2+STOP3
Refer to section 2.2.6 for a description of the RUN command, section
1.4.16 for the RECORD ,c statement and to Appendix L for more
information about the CHAIN statement.

Idata manipulation statements
1.4.4

Data Manipulation Statements

Data manipulation statements clear data fields, move data between
fields, calculate decimal expressions, convert data from decimal to
alphanumeric and vice versa and format data.
The form of this
statement is
destination field = source field
where source field is a field name, variable, constant or expression,
and destination field is the field where the data is to be stored.
The contents of the source field are moved to the destination field.
The destination field must have been defined in the data definition
section as either an alphanumeric or decimal field.
If
the
destination or source field is decimal, the data from the source field
is right-justified; if both fields are alphanumeric, left-justified.
The source data is always converted to the type defined for the
destination field.
Record names can be used in data manipulation
statements.
Arithmetic Expressions:
Arithmetic expressions are allowed on the right side of the equals
sign.
The expression can contain decimal elements, subscripted data
elements, constants, variables and arithmetic operators.
There are
five binary arithmetic operators:

# (rounding)

Order of priority:
1. rounding
2. multiplication and division
3. addition and subtraction

I (division)

* (multiplication)

+ (addition)
-

(subtraction)

These operators require decimal operands.
Operators with the same priority are executed left to right.
The result of a division operation
example, 5/3 is 1 and -14/5 is -2.

truncated

toward

zero.

For

An error occurs when a divide by zero is attempted or when the
of an
small.

addition,

subtraction

result
multiplication is too large or too

There are three unary operators (-,+,#). The (+) has no effect; the
unary - negates the decimal value to which it is specified. More than
one unary minus can be specified for a number, e.g., A= - -3 is
equivalent to A= +3 and A= --+-4 is equivalent to A= -(-(+(-4))) or
A = -4.
The number sign (#) operator is also used to convert an alphanumeric
character to its equivalent internal code and make that decimal number
(0-63) available to the program.

1-11

The # operator takes that value from the alphanumeric or decimal field
on the right side of the expression, checks the first character
(left-most) , determines the equivalent internal code and returns the
code as a decimal number. The result is a decimal value.
In this
application, the number sign (#) appears before the character.
Examples:

A = #B

3 + #B

A is a decimal field and B is an
alphanumeric field.
If B contains the
characters XYZ, X is converted to its
internal
code, 71, and the decimal
equivalent, 57, is stored in A.
A is a decimal field and B is an
alphanumeric field.
After conversion,
the decimal equivalent of the first
character of B, 57, is added to 3 and
stored in A.

= 10

Fl = lOO*A/2+3-1
The order of priority can be altered by the use of parentheses, since
expressions in parentheses are executed first.
The above expression
equals 502.
Fl=lOO*(A/2+3-1)
The result of this example is 700.
Illegal
A+(B*C
C*3)-F
INTRST*((4*AMT+2)-PAYMNT/2
LENGTH+WIDTH+
6/3005434+'ABCDE 1
Rounding:
Variables may be truncated and rounded.
The resulting variable is
rounded upward if the digit which followed the remaining variable
before truncation is 5 or more. The variable is rounded in terms of
magnitude; sign is unchanged.
The format for rounding is:
A#B
where A is truncated by B places and rounded in magnitude.
B cannot
be greater than 7 and should be a positive integer.
If B is negative,
it is treated as positive. When used for rounding, the number sign
(#) appears after a character. For example:

1-12

MONEY,

TEMP=MONEYf 2
If MONEY equals 123456, TEMP equals 1235.
If MONEY equals 123446, TEMP equals 1234.
If MONEY equals -1473, TEMP equals -15.
Typically, this feature would be used for rounding to the dollar.
Variables:
Variables may take on one of the following three formats:
a.

narne

b.
c.

name {subscript)
name (position 1, position 2)

name - consists of a sequence of letters and digits beginning with
a letter, only the first 6 characters are significant. Used in
record name, field name, etc.

name{subscript) ~name should be a subscripted array and the value
of the subscript must be between 1 and the dimension specified in
the array.
The subscript must be decimal and can be
an
expression.
If either name is not an array or the value of the
subscript exceeds the dimension other locations in core outside
the limits of name are referenced.
No error message is generated
unless the user tries to reference data so far away from his data
area that he is referencing his program or the system.

name (position!, position2) -position! must be less than or equal
to position2; Position2 should be less than or equal to the
maximum size of the array associated with name.
Positions 1 and 2
must be both positive and decimal. This form of subscripting
references those characters from position! to position2 inclusive.
If the variable name is .sub.scripted, yo.u may c.onsider the
successive array elements strung out left to right. For example,
in referencing A(3,9) in the array A declared as 4D4, the 7 digits
are referenced as follows: the last 2 digits of A(l), the entire
A(2), and the first digit of A(3).

Decimal Literals:
These consist of a sequence of from 1 to 15 decimal digits.
Alpha literals:
These consist of a sequence of any legal COS characters (except single
quotes) enclosed in single quotes.

1-13

Example:
'ABCD'
RECORD Literals:
A RECORD literal is a constant that can be used anywhere in the
PROCedure section of a DIBOL program where a record is required.
It
is similar to an alpha literal except it begins with a double quote
(") and ends with a single quote ('). For example:
REC="THERE I
!NIT ( 8, TTY)
XMIT ( 8, "HELLO I)
XMIT ( 8 , REC)

1.4.4.1

Clearing Fields

Data manipulation statements clear fields when used in the form:
field name =
The destination can be a single field or an array.
If alphanumeric,
it is cleared to all spaces; if decimal to all zeros. Fields, parts
of fields or fields in an array, can also be cleared by the use of
subscripts
(expressions in parentheses). Any part of an array can be
accessed in a program statement by listing the position of the field
of character(s) in paren~heses. Specifying an array name without any
subscripts, clears the first array element only.
Examples:
Field Name

Meaning

Fl(S,7)

clears character positions 5,
field Fl

Fl(S)

clears the fifth element in an array

Fl (A)

=
=

clears the Ath element in array Fl

Fl(l,l)

and

clears the first character.

To clear a RECORD area to all spaces use the form:
RECNAM=
Record names can be subscripted to allow reference to RECORD areas as
though they were in an array. All records to be so referenced must
follow one another and be of the same length. For example:

1-.·14

RECORD Z
,A3
RECORD Z2
,A3

Z(2)=
Clears the next record defined after record Z in the data
the program.

1. 4. 4. 2

section

Moving Alphanumeric Data

To move the contents of one alphanumeric field to another alphanumeric
field, use the following form:
Alpha destination = Alpha source

the

source

is shorter than the destination, the
data
is
with the right=most characters of the destination field
undisturbed. If the source is longer than the destination, the
right-most characters of the source field are not moved into the
destination field. The source field retains the original data.
left~justified

For example:
RECORD
A,A5, I ABCDE I

B,A3,'FGH'
PROC
A=B
A now contains

FGHDE

RECORD A
NAME ,A4 I I FRED I
NAME1,A7,'JOHNSON'
PROC
NAME=NAMEl
The name in RECORD A is now JOHNJOHNSON.

1.4.4.3

Moving Decimal Data

Use a statement of the following form to
decimal field to another decimal field.

move

the

contents

one

decimal destination = decimal source
If the value in the source field is shorter than the length of the
destination field, zeros are inserted on the left. If longer, the
most significant digits are not moved.

1-15

1.4.4.4

Moving Records

An assignment statement may be of the form
RECl=REC2
where RECl and REC2 are record names or subscripted record names or
REC2 may be a record literal.
Its effect is to move the contents of
REC2 into the space reserved by RECl. If REC2 is smaller than RECl,
the right-most characters of RECl are undisturbed.
If REC2 is larger,
the right-most characters of REC2 are not moved. Extreme caution must
be used if the destination record is subscripted.

1.4.4.5

Converting and Formatting Data

Decimal values can be converted to alphanumeric and vice versa for the
purposes of I/O and calculations. Any data field may be reformatted
to contain spaces and punctuation marks which are not stored with the
records on DECtape or disk and cannot be present during calculations.
Alphanumeric to Decimal Conversion - The form used for this conversion
is:
Decimal variable = Alpha expression
Alphanumeric values to be converted to decimal must be 16 or less
characters in length. The source is an alphanumeric expression which
contains an optionally signed representation of a number.
The
destination is a simple or subscripted decimal variable.
Example:
DEC= ALPHA
The numeric alphanumeric field (ALPHA) is converted to decimal and
stored in the decimal field (DEC).
If the decimal field is larger,
the data is right-justified and zero-filled.
If the decimal field is
smaller, the excess characters to the left are not moved. If the
alpha field contains characters other than digits, spaces and signs (+
or-), the message BAD DIGIT results at run time. The statement should
be preceded by an ON ERROR statement if · the contents of ALPHA may
contain BAD DIGIT's
(refer to the ON ERROR statement for more
information). Spaces and signs are not counted as characters that are
moved.
Spaces in the alpha field are ignored. Signs may be irnbedded
anywhere in the alpha field. Two minus signs are equivalent to a
plus; three minus signs are equivalent to a minus; etc.
Decimal to Alphanumeric Conversion - The form of this conversion is
Alpha variable = decimal expression
The source is a decimal expression and the
alphanumeric variable. Example:
ALPHA = DEC

1-16

destination

simple

The contents of the decimal field (DEC) are placed in an alphanumeric
field
(ALPHA).
If the alphanumeric field is larger, the data is
right-justified and space-filled on the left.
If the alphanurt$ric
field is smaller, the excess characters on the left are not moved.
Leading zeroes in the decimal field are converted to spaces. However,
if the decimal field equals zero, the alpha field is set to spaces
except for the right-most character, which is a zero.
Data Formatting - Decimal fields can be
alphanumeric fields using the form:

reformatted

and

stored

Alpha variable = decimal expression, format

The source is a decimal
alphanumeric variable.
inserted with numbers.

expression and the destination a simple
Format specifies special characters to be

Example:
A= D, '-XXX,XXX.ZZ'
The eight-digit decimal number at D is converted to alphanumeric code,
reformatted with specified punctuation and stored in alphanumeric
field A. The format string must be an alpha expression.
The formatted
value
is
placed
in
the
destination
field,
right-justified and padded with spaces if necessary.
If the formatted
value is larger than the destination field, the left-mo~t characters
are not moved.
Most printing characters on the line printer or terminal can be used
in a format string but the following characters must be used with
care: X,Z,*,-,.,', and comma. Table 1-3 shows the special characters
to be used in format strings.
Examples:
TOTL=TEMP,'XXX,XXX.XX-'
If TEMP contains 12345678, TOTL will contain 123,456.78
AMT=PAY,'*XXX,XXX.XX'
CDATE=DATE,'XX/XX/XX'
GTOT=TOTl+TOT2,'-XXX,XXX.ZZ'
W'AGES=RATE'9t'HOURS,~XXXX.XX'

RATIO=(TOT1/TOT2)#2,'XXX.XX'
TOTl is divided by TOT2; the result is rounded two places; that result
is formatted and stored in RATIO.
Al=NUM, I xxxo I

If NUM contains 987, Al will contain 9870.
A2=DATE,'.XX/XX/XX 1
A=Q, 'Q=XX'

1-17

Also:
TOTFMT, All, •-xxx,xxx.zz•
GTOT=TOTl+TOT2, TOTFMT
TOTFMT is specified in data section and is used to specify
the PROC section.

format

When using a comma, period, slash, minus sign or any other notation,
each must be counted as a character position.
In the first example
above, for instance, TOTL must be defined in a RECORD statement as an
eleven-character alphanumeric field.
In the example A=B,C! B contains a decimal value being converted to
alpha and stored in A (an alpha field). The C represents an alpha
field which contains the format strings to be used in the command.

TABLE 1-3.
Character

SPECIAL CHARACTERS
Explanation

Used in a format to arrange a decimal field for
printout.
Each X represents a digit and leading zeros
are automatically suppressed.

Special character
formatting output.

Used in a format string to replace leading zeros and
eliminate trailing spaces on printout.
If the * is not
the first character in the string. digits may also be
replaced.

used

suppress

digit

when

Inserts an arithmetic sign in a number to be printed.
The sign may be placed before or after the number.
If
the number is positive, a space is inserted where the
is placed.
If the hyphen is placed in a position
following the first significant digit, but previous to
the last position of a format string, it is printed
lik-e any other inserti-on· character.
Inserts a decimal point in a format string and
zeros
to the right of the decimal point
significant.

forces
to be

Used to insert a comma in a format string if there
significant digits to the left.
All other characters
characters.

1-18

are

treated

are

insertion

I DISPLAY
1.4.5

DISPLAY

The DISPLAY statement is used primarily with the VTOS terminal to
display a message on the scope and to move the scope cursor to the
specified line and character position. DISPLAY is used in conjunction
with ACCEPT to display questions on the scope and store the replies.
The form of the DISPLAY statement is:
DISPLAY (y ,x,

'literal;')
afield ) )
{ df ield

where
y

is a decimal expression representing the scope
line number (1-20). If the line number specified
is not within 1-20, the results are unpredictable
when used wi t..11

NOTE
If y is zero, no positioning is done and
x is ignored.
The statement with Y
equal to 0 is used to output a message
without a carriage return/line feed i.e.
DISPLAY{O,O,'NAME').
x

is a decimal expression representing the character
position (1-72).
If the character is not within
1-72 the results are unpredictable.

'literal' is an alphanumeric or decimal character string.
An alphanumeric string must be enclosed in single
quotes (') and is displayed at the character
position
specified.
There
is
no carriage
return/line feed after the message and the cursor
remains in the character position at the end of
the message.

A decimal literal must be
codes:

one

the

following

O=position cursor.
!=clear to end of scope.
2=clear to end of line.
25=emit a bell or beep sound.
These are the only codes recognized by COS and use
of any other codes causes unpredictable results.
afield

is a field defined as alphanumeric which
a message to be displayed.

1-19

contains

df ield

is a field defined as decimal which contains
code that causes a particular operation.

O=position cursor.
l=clear to end of scope.
2=clear to end of line.
25=emit a bell or beep sound.
These are the only codes recognized by COS and use
of any·other codes causes unpredictable results.
Examples:
DISPLAY{0,0,25)

;ring terminal bell.

DISPLAY(l,1,1)

;clear entire scope.

DISPLAY(l0,1,1)

;clear from character position 1
;line 10 to end of scope.

DISPLAY(2,20,DAY)

;display the contents of
;alpha field DAY on line
;2 starting at character
;position 20.

DISPLAY(l,10,0)

;position cursor at first
;line, 10th character position.

DISPLAY(ll,1,2)

;clear line 11 starting at
;character position 1.

DISPLAY(ll,37,2)

;clear second half of line
; 11.

DISPLAY(ll,12,25)

;position cursor at line 11,
;character position 12 and
;emit beep.

DISPLAY(Y,X,BEEP)

;position cursor at coordinates
;defined in X and Y and
;execute code defined in
;BEEt>. If the programs
;data division contains
RECORD

Y,02,20
X,D2,36
BEEP,D2,25
;the cursor is set at line 20,
;character positon 36 and an
;audible beep is sounded.

1-20

DISPLAY(20,36,FIRST)

;if the data division contains
RECORD
FIRST, A4 , 'NAP'..E 1
;NAME is displayed at line
; 20 character
;position 36. The cursor is
;left in the 40th position of
;line 20.

DISPLAY(l,l,'HELLO')

;display HELLO in the
;upper left hand corner
;of the screen ..

DISPLAY(l0,10,BEEP)

;If the data division contains
RECORD DISP
POST,D2,00
CSCOPE,D2,0l
CLINE,D2,02
BEEP,D2,25
;the cursor is placed at
;character position 10 on line
;10 and an audible beep
;is emitted.

See Appendix L for more examples.

1-21

1.4.6

END

This is an optional statement and is used to indicate the
of a program. The form of the statement is

termination

END [/x]
where
/x

is an optional switch to the compiler
/N

no compiler listing of source program
map

List source
printer

List source program storage map on terminal.

program

storage

map

storage
line

If no options are specified, a compiler listing is
produced as usual. The END option switch or lack
of it can be overridden with an option switch in
the compiler RUN command. If no symbol table is
printed, the symbol count and number of free
locations are not printed.

1-23

I
1. 4. 7

FINI

The FINI statement writes an end-of-file mark, closes the file,
rewinds the tape (if DECtape) and writes the file length in the file
label (write only).
(In multi-reel files the file length is written
only on the last reel.) The form is:
FINI (n)
where n is a decimal expression whose value is 1-15 specifying a
channel number which was previously specified in an INIT statement.
FINI disassociates the channel number from the device specified in the
INIT statement.
(See Section 1.4.12.) Assignments of channel numbers
and devices can then be changed with another INIT statement. FINI is
necessary for mass storage output files only but it is considered good
programming practice to FINI each data file INITed.
If specified for
an input file (other than from SYS), reading of that file stops. The
file may again be read {from the beginning), if another INIT statement
is executed~
Examples:
Legal
FINI(!)
FINI(OOS)
FINI (A+B)

Illegal
FINI
FINI(
FINI ( 4
FINI,3
FINI(O)
FINI(l6)

If an output file is not INITed, the end-of-file mark is not written
and records may be lost.
If an update file is not FINied, the last
few records written may be lost.

1-25

I FORMS
1.4.8

FORMS

The FORMS statement is used to format line printer (only)
has the form:
FORMS

output

and

(channel, skip-code)

where
channel

is a decimal expression
(value 1-15)
associated
with the line printer in a previous !NIT statement
(refer to Section 1.4.12).
If
the
channel
specified is not associated with a line printer,
the statement is ignored.

skip-code

is a decimal expression with a value of 0 to 4095
which causes the line printer to go to the top of
the page or skip the number of lines specified.
The codes are:
n
v

1-4095
-1

-2

-85

go to top of next page
skip this number of lines
skip to channel 2
(LSBE printer only)
and
move paper forward n lines as
specified on vertical forms
control
tape.
print enlarged characters (LSBE printer
only)
for next XMIT statement. Allows
only half as many characters per line as
normal size printing.
sound alarm bell (LSBE printer only).

Other
negative
numbers
cause
unpredictable
results.
The skip-code is compiled modulo 4096.
For example, FORMS(6,4096) causes a top of form to
be executed.
Example:
!NIT {l ,L)

FORMS (1,3)
1 is the channel number specified in a previous !NIT statement

is the nwilber of lines to be left blank.

1-27

and

INIT(S,L)

FORMS(S,O)

5 is the channel number and 0 specifies top of line printer page.
INIT(S,L}

FORMS ( 5 ,-1)

5 is the channel number and -1 causes the page to move up the number
of lines specified on the vertical forms control tape of the LS8E
printer.

1-28

GO TO

1.4.9

GO TO

The GO TO statement has two forms:
(1) unconditional and
(2)
computed.
It transfers control to the statement whose label is
specified in the GO TO statement.

Unconditional GO TO

1.4.9.l

The form of the unconditional GO TO is:
GO TO label
where
label

is a statement label assigned to the statement in
the PROCedure section where control is to be
transferred.

Example:
Legal

Illegal
GO TO
GO TO,TARGET

GO TO SET

SET, TOT=TOT+SUB

1.4.9.2

Computed GO TO

The computed GO TO has the form:
GO TO (labell,label2, ••• ,labeln) ,variable
where
labell,label2, ••• ,labeln are statement labels.
number of labels.)
variable

(There can be any

is a decimal variable
representing a value.

expression

Control
is
transferred
to
the
statement
labeled
label!,
label2, ••• labeln if variable has the value 1,2, ••• n respectively.
If
variable is negative, zero or greater than the number of labels,
control passes to the next statement in sequence.

1-29

Example:
GO TO (LOOP,LIST,TOT) ,KEY
KEY acts as a switch transferring control to LOOP if
KEY=2, or TOT if KEY=3.

KEY=l,

Other examples:
Legal
GO TO{LAB1,LAB2,LAB3) ,VAR
GO TO{TRY,LOOP2,TRY,TRY,TOP) ,Q+l
GO TO(TEST) ,IN
GO TO(Ll,L2,L3,Ll,L2,L3,0VER,L2,L3,LA83,W,Ll) ,NOW
GO TO(EI,EI,C) ,(TEST+2)*1-FLAG
Illegal
GO TO(
GO TO(LAB1,LAB2
GO TO{Ll,L2)
GO TO{Dl,D3),
GO TO{Dl,D4)KEY
GO TO{) ,COUNT
GO TO(SO,LO,) ,PEAR
GO TO(Ll,,L3) ,KIT

1-30

LIST

1. 4.10

An IF statement conditionally executes certain statements on the basis
of the result of a comparison of expressions.
The form of the
statement is
IF {expressionl.rel.expression2)stmnt
where

expressionl,expression2

are
a
combination
of
constants,
variables or arithmetic expressions of
the same type.

.rel.

is one of the following
relational
operators
.EQ. Equal
.NE. Not equal
.LT. Less than
.LE. Less than or equal
.GT. Greater than
.GE. Greater than or equal

stmnt

is a control statement which is executed
if the relation is TRUE.
"stmnt" must
be one of the following
GO TO label
STOP
CALL label
TRACE
RETURN
NO TRACE
ON ERROR label

The two variables must be of the same type: both decimal or both
alphanumeric.
Numeric data may be in decimal or alphanumeric form.
In a decimal comparison the shorter field is internally filled to the
length of the longer field, then the comparison is made between the
longer field and the zero-filled field.
In an alphanumeric comparison
the compare is made on the number of characters in the shorter field,
i.e. in comparing a 3 character field and a 5 character field, only
the 3 left-most characters of each field are used.
If the result of comparison is false, the next statement
is executed.
Examples:
Legal
IF (A .EQ. B) GO TO LABEL3
IF (SLOT.NE. 2) CALL BAD
IF{SALES.LT.PROFIT+TAX-RENT) NO TRACE
IF(A+B+C .GT. D+E*F) RETURN
IF(NUMBER.LT.D+3) CALL FIXUP

1-31

sequence

Illegal
IF
IF
IF
IF
IF

(C.LE.76)
(D2.EQ.'ABC') THEN GO TO RAL
{3.NE.5) D='FALSE'
{Q.LE.' l, GO TO TRADE
{DIST.GT.RATE*TIME) GO TO (Ll,L2) ,KEY

IF
IF
IF
IF
IF
IF

)E.GT.8) GO TO LAND
(ABC ,GT. DEF) GO TO CHI
(TIME.GT:lO) GO TO BED
{P-1 .LT. O, GO TO UPDATE
(D.LT.l .AND. D.GT.-3) GO TO NOW

IF (COST .GR.
IF (TAX .LTE.

EXPNSE) GO TO WR
100) GO TO BUY

1~32

IINCR I
I

1.4 .11

INCR

The INCR (increment) statement adds 1 to the
has the form:
INCR

specified

variable

and

variable

where
variable

is a decimal variable to be incremented by 1.

INCR can be used with positive numbers only and cannot be the object
of an IF statement. !NCR is typically used to add one to a counter
and is faster than a data manipulation statement. For example:
A2=A2+1
or
INCR A2
Legal
INCR A2 (3)
INCR B(S)
INCR C (H,6)

Illegal
IF (A.LT.16)INCR B
!NCR A+B
INCR 3

1-33

!NIT

1.4.12

!NIT

The !NIT statement associates a channel number with a mass storage or
character oriented device for input or output and initializes the
device. The form of the !NIT statement is:
!NIT {channel, dev[,data file name][, unit])
where
channel

is a decimal expression which evaluates to a
number 1-15 specifying a channel to be associated
with a logical or physical device.
(Refer to
section 3.3 for logical unit assignments.) This
number is then used in other statements such as
XMIT to refer to the same device.
If the number specified exceeds 15,
modulo 16.

taken

If not specified, the following channels are
initially
associated with devices at program
startup.
(The user may change these assignments
with an !NIT statement.)

5 = Paper Tape Reader
6 = Line Printer
7
Keyboard

8 = Terminal scope or printer

A channel that is associated with a logical
must be FINied prior to another !NIT.

unit

dev

is the designation for the COS device to be
associated
with
the
channel number.
These
designations can be abbreviated, since only the
first character is checked. The designations are:

Designation

Abbreviation

Mass storage
input.

device

used

for

OUT

Mass storage
output

device

used

for

UPDATE

Mass storage device to be used for both
input and output
{direct access) {See
READ and WRITE)

KBD

Input from terminal keyboard

TTY

Output to terminal printer or display

Meaning

1-35

Designation

Abbrevia ti on

LPT

Line printer

CDR

Card reader

PTP

Paper tape punch

RDR

Paper tape reader

SYS

Input from a file located on the system
device.
This file name must have been
specified in a RUN statement.
(Refer to
Appendix L, Section 6.0.)

Meaning

data file name is an alpha constant or variable that identifies
the data file on the logical COS unit. A data
file name is necessary with the I, o, and U
devices and illegal with other devices. If the
name is more than six characters only the first
six are used.
If less than six, the name is
left-justified and padded with spaces on the
right. It can also be the name of a field defined
in the data section with the P option which allows
flexibility in the files specified to the program.
unit

is an optional decimal expression used with I,O
and U device codes which specifies the logical
unit (1-15) where the data file name is stored or
to be stored.
{If the number specified exceeds
15, it is taken modulo 16) On input this unit is
checked for the data file name if it is not
present, a MOUNT message is displayed; on output
this unit is checked for the data file name or a
temporary (scratch file) name, if another file is
present, a REPLACE? message is displayed.

The Monitor uses the name specified in the !NIT statement in the run
time MOUNT message
(refer to Chapter 2). The reply to the MOUNT
message is the logical COS unit where the file is stored (INPUT) or to
be stored (OUTPUT). The logical unit is then checked by the MONITOR
to be sure the named file is indeed stored there. In the case where a
logical unit is specified in the !NIT statement, the Monitor checks
the logical unit for the named file and if found bypasses the MOUNT
message.
If not found on that logical unit the Monitor displays the
usual MOUNT message.
Examples:
!NIT (15,I,'RENEW')
initializes channel number 15 for input,
the input file.

1-36

(I).

"RENEW " is the name of

RECORD D
TAPEID,A6,P

PROC

INIT (3,IN,TAPEID}
The previous code causes the message ENTER TAPEID to be displayed.
The user responds with the name of the file; a MOUNT message is then
displayed and must be answered with the proper logical unit number.
This allows different file names to be specified each tiMe the program
is run.
INIT(l+l,IN,FILEl)
initializes channel 2 for input (I).
Legal

Other examples:
Illegal

INIT(l,IN,'FILEOl')
INIT(2,INPUT,'FILE02')
INIT(3,I,'FILEOS')
INIT(4,IXYDREQW,'LABEL 1 )
INIT(S,OUT,'TAPEl')
INIT(6,0UTCH,'TAPE2')
INIT (7 ,O,RCRD)
INIT(l,UPDATE,'LABEL')
INIT(3,UPD, 'L')
INIT {6 , u, ' L3 I )
INIT(6,UP,'T453')
INIT ( 5, KBD)
INIT ( 5 , TTY)
INIT(4,K)
INIT(l,TELETYPE,NAME)
INIT(9,LPT)
INIT ( 10 ,CDR)
INIT ( 11,RDR)
INIT ( 12 , PTR)
INIT(N,SYS)
INTT (15, IN MASTER I ,A+B)

!NIT
INIT(
INIT(4
INIT<4>
INIT,3

1-37

ioN ERROR
1.4.13

ON ERROR

This statement can be inserted in the source program to cause a branch
to the statement with the specified label when a non-fatal executed
error occurs in the next program statement.
The form of this
statement is:
label

label is a statement label assigned to the statement in
PROC section where control is to be transferred.

where

the

The ON ERROR statement can be the object of an IF statement.
Examples:

ON ERROR ERRTRP
TEMP=lOO*ORIG/NEW

ERRTRP,

ON ERROR TRAP
DEC = ALPHA

TRAP,
The ON ERROR statement eliminates a return
following run time errors.

the

Monitor

for

the

zero divisor
bad digit in an alpha to decimal conversion
more than 15 digits in decimal field used in calculation
an end of file label was not specified in an XMIT statement
input line overflowed block it was read into
no file specified in
statement

RUN

command

satisfy

INIT

(SYS)

in a random-access operation (see READ and WRITE)
the file
contains variable length records, record number is too large
or O, or length specified in the record header word does not
match the length of the XMIT block.

1-39

IPROC

1. 4.14

PROC

The PROC statement separates the two sections of a DIBOL program, the
data definition section and the procedure section. It is of the form:
PROC [n] [/x] [;comment]
where
n

is a single digit, 0-7,
(not an expression)
indicating the maximum number of mass storage
logical units which the program will have open
simultaneously.
If no number is specified, the
compiler assumes 7. The available core is divided
into buffers to handle the number of logical units
specified.
The more buffers
necessary,
the
smaller they must be and the slower the I/O
process.

is an optional listing switch to the compiler
/N

temporarily halts compiler listing
program

source

resume listing source program
line printer

and

errors

resume listing source program
terminal

and

errors

The PROC option switch is active until disabled by
a START or END statement with an option switch.
There can be only one PROC statement.
If no
option switch is specified, a compiler listing is
produced as usual. The PROC option switch can be
overridden with an option switch at compile time.
(Refer to Chapter 4.)
comment

is an optional string of text preceded by a
semicolon which is stored for output as a heading
for the procedure section 0-f the compiler listing.
When the compiler encounters the PROC statement
the line printer moves to the top of the next
sheet of paper and outputs the comment, if any, as
a header line.

If /N is specified in the RUN COMP command, it overrides any /L's or
/T's in the source program.
However, /L and /T determine on what
device errors are listed. If the program simultaneously opens more
mass storage logical units than were specified in the PROC statement,
a run time error occurs.

1-41

Examples:
Legal

Illegal

PROC
PROC 3;BILLING V3
PROC 3/N
PROC;
PROC 4/L; TEST PROG.

PROC 32
PROC, 4
PROC 5+2
PROC BILLING V3
PROC A

1-.,42

1. 4.15

READ

The direct access READ statement allows a specified data record to be
moved from a named file to a specified area in core and has the form:
READ (channel, record,rec#)
where
channel

is a decimal expression with a value of 1-15
specifying a channel number which links the READ
statement to the related INIT statement.
(The
INIT statement must specify INPUT or UPDATE as the
COS device. )

record

is the record into which data is to be read.

rec#

is a decimal arithmetic expression specifying
sequence ntL~ber of the record to be read.

the

If the program READS past the end-of-file mark, the results
undefined. See Section 1.4.22 for restrictions on usage.

are

Examples:
READ (5,REX,88)
reads the 88th record of the device linked to the channel which was
INITed with the INIT (S, ••• ) statement and places it in the core area
labeled REX.
READ (6 ,BLT ,EXPR)
reads the record specified by the expression EXPR and stores it in the
core area labeled BLT.
Refer to Appendix L for more
access techniques.

examples

1-43

and

discussion

direct

RECORD

1. 4 .16

RECORD

The RECORD statement reserves areas of core where records are stored
during processing. Block can be used interchangeably with RECORD.
The RECORD statement is of the form:
RECORD [name]
and is followed by statements of the form
fldnam,xn
where name labels the record area; fldnam names a field within the
record and xn specifies whether the field contains alphanumeric or
decimal data, and the number of characters in the field.
Storage is
allocated in the order that the data appears. The RECORD name can be
any number of alphanumeric characters in length (only the first 6
characters
are
siqnificant)
and the first character must be
alphabetic. The name-is optional unless the program is to transfer
(XMIT,READ or WRITE) the data.
If no name is specified, the data is
assigned to contiguous locations but is not available for input or
output.
Therefore, RECORD statements with no names are used for
temporary storage.
Examples:
Legal

Illegal

RECORD
RECORD ALPHA
RECORD Tl376
RECORD R

RECORD $
RECORD, ABC
RECORD 564
RECORD,

Each RECORD statement should be followed by one or more data field
definitions.
The data field name (fldnam) is optional and may be any
number of alphanumeric characters in length (only the first 6
characters are significant). The first character must be alphabetic.
A corrana may be used without a field name if the program does not
reference the individual field.
This specification is used when
formatting an output line for the printer, for example, so that
intercolumn spaces do not require a data name, only a comma followed
by the type and size.
Data field type
(x) is alphanumeric (A) or decimal (D).
All
input/output data to the line printer, terminal or high-speed paper
tape reader should be in alphanumeric form (A).
Decimal (D) is
specified for data to be used internally by the program. However, the
alpha representation of a positive decimal number is the same as its
decimal representation.
This is not true of negative numbers (refer
to Appendix A).
Data stored in
alpha
fields
is
generally
left-justified and if necessary padded with blanks on the right. Data
stored in decimal fields is always right-justified and padded with
zeros on the left if necessary. The data field size (n) is a decimal

1-45

number 1-510 specifying the number of characters in the field.
Decimal fields to be used in arithmetic operations have a maximum
length of 15 digits. If data size is omitted, 1 is assumed.
The
total size of all data fields in a record must not exceed 510, if the
record has a name (i.e. can be input or output}, otherwise the total
size must not exceed 4094.
Examples:
Legal
RECORD MASTER
A, D3

Ml, D4
M2, AG
AG
RECORD TRANS
COST, Al6
ADD, A30
KEY,

D03
Illegal
RECORD NUMBER
SETNO Dll
ORDERS, D632
RECORD NUMBER
B,
A-6

1-46

Inserting Initial Values
Initial values can be assigned to data fields as part of the field
statements following a RECORD statement. The initial value may be set
as part of the data field definition or marked for insertion via the
terminal when the program is ready for execution. The maximum size
(limited by the Monitor) of an initial value is approximately 110
characters.
If no initial value is assigned to a field it is
initially set to all blanks (if alphanumeric) or zeroes (if decimal).
If entered as part of the program, the alphanumeric or decimal
constant is placed after the type and size specification. A comma
must separate the constant from the type and size specification.
The
initial value to be entered must agree with the size and type
specified for the field. The alphanumeric constant must be enclosed
in single quotation marks
(' ')
and may contain any printable
characters except\,' ,or+. Tabs inside the single quotation marks
are treated as spaces, 1 space per tab. A decimal constant is a
string of decimal digits optionally preceded by a sign (+ or -)
and
optionally followed by a minus sign. The signs (+ or -) and quotes do
not count as part of the size of the initial value. The compiler does
not insert leading zeros or trailing spaces in initial values.
Examples:
Illegal

,A4, 'TEXT'
M,06,000400
PA,A2,'ll'
AA,03,123
AC,A3,'123'
A,03,-146
B,03,146C,03,+123

,A4:
,A4,

'NOTE'
'3Q'

,A3,345
,04,
,D4,

I WEED'
12

,04,12345
A3 '2'
A3,
'HELLO'
AlT
AD,A3,

If the initial value is to be entered when the program is executed,
put a comma and the character P after the type and size specification.

1-47

Example:
fldnam,A8,P
The character P is used in the data field specification to allow
user to enter data at run time. Before execution, the message

the

ENTER fldnam
is displayed on the terminal.

The specification

QZB135,A6,P
causes the message
ENTER QZB135
to be printed at the terminal at run time.
characters

The

may

use

six

999999
The terminal entry is stored in the field named.
If the value entered
at run time in response to the ENTER message is shorter than the
field, the value is padded with blanks, if longer, an error will
occur.
The P should only be used with named fields and is executed
once each time the program is run.
NOTE
When using the P option all information entered
from the terminals is treated as alphanumeric
e.g.,
LABEL,A6,P
in the data section causes the message
ENTER LABEL
If the user replies
ABCD)
LABEL will equal 'ABCD

The sequence

LABEL,07,P
ENTER LABEL
AB123)
causes LABEL to equal AB12300.
A variable to be initialized with the P option can also
initial default value.

1-48

given

Example:
NAME ,A3, 'DOE' ,P
The default value must agree in type and size with that specified for
the field.
The program uses the default value if CTRL/Z is typed in
answer to the run time ENTER message.
Examples:
Legal
ENT,A6,P

Illegal
,D2P

DUP,D8,P
zIP, AS , I 0 0 0 0 0 I , p

1-49

The character D is used in the data field specification to insert the
current date (as specified to the Monitor at start up) in a field to
be used in a report heading. The date is stored in the form mmddyy,
e.g., June 25, 1972 would be stored as 062572. The field must be
exactly six characters long or the date stored in it will be
incorrect.
This eliminates the necessity of specifying the P option
and entering the current date each time the program is run. The P and
D options cannot be used on the same field.
Example:
RECORD A
DATE,D6,D
Arrays
A number of values can be represented with a single field name by
placing a repetition count before the type and size specification.
All elements of the array must be the sa..~e size and type.
For
example:
TAG, 402
defines TAG to be 4 decimal fields, each 2 digits long. Data can be
initially entered in an array as a continuous string separated by
commas. For example,
TAG, 4D2,88,44,22,55
It is not necessary to initialize all fields of an array.
However,
those fields to be initialized must be the first field and any
continuous number of fields after the first field, for example:
TAG, 402, 88 ,44
initializes the first two fields.
The
initialize the first array element only.

option

Examples:
Legal

Illegal

C,4D3
D,5A2

,2D3, 34
,203,4545

,2A4, I BLUE I

,1D7
,16Al
GOAL,302,22,23,33
UP,4A2,'AB','CD','EF','GH'
INPUT ,3D5 ,P

, 2A2, I AMERICA'

LABEL,103,Al
,3E4
,OD6

,-3A4
,6D-3

GOAL, 302,222333
UP,4A2,'ABCDEFGH'

1-51

can

used

The character S is used in the data field to assign the value of
variable (A2) equal to the value of the options used at run time.
Example:
VAR,A2,S
.i- ......
will set 't7'1\'0
V.M."'1..V

lvvl
A.I.

•

'OTT
A.'\.\.I

prona..Ttt/XY 1S

1--53

run.

RECORD Overlay Option
The X option in a RECORD statement allows multiple definition
fields within a RECORD. Specify the overlay option as follows:

RECORD,X
er

RECORD name ,x

The X option must be preceded by a comma. There can be one or more
overlays defining the same record but there must be a preceding RECORD
statement without an x.
The overlay (X) record must be equal to or less than the size
record being overlayed.

the

Examples:
Illegal

RECORD ONE
RECORD X,X
RECORD, X

RECORD TIME,X,
RECORD X ( X IS A LEGAL RECORD
NAME BUT IN THIS FORM DOES NOT
INDICATE AN OVERLAY)

The X overlay option is useful in reformatting
record area. For example,

previously

defined

RECORD COLOR
Fl,A4,'BLUE'
,Al
F2 , A3 , ' RED'
has three alphanumeric fields with a total of
Fields Fl and F2 can be referenced individually.

eight

characters.

RECORD RAINBO,X
ALL,A8
overlays the previously described record and permits treatment as one
field.
The variable ALL contains BLUE RED and setting ALL to blanks
in the PROC section sets Fl and F2 to blanks.
The X option also
allows access to a particular portion of the record by specifying

.o,x
,A4
Q6,A4

RECORD

which skips the first four characters and uses the last four.
contain the quantity ' RED'.

Q6 will

The most common use for record overlaying is with print records. Each
unique format for a report is kept in a separate record overlay. For
example:

1-55

RECORD HEAD!
Hl, AlO, 'HEADING il'
RECORD HEAD2
H2, A23, 'THE SECOND HEADING LINE'
RECORD PRINT
PR132, Al32
RECORD ,X
CUSTNO, AlO
;CUSTOMER NUMBER
,
A20
CUSTNM, A30
;CUSTOMER NAME
RECORD, X
,
A35
CUSADD, A20
;CUSTOMER ADDRESS
,
AS
CUSZIP, DS

PR132=Hl
XMIT(6,PRINT)
PR132=H2
XMIT(6,PRINT)
PRINT=
CUSTNO=NO
CUSTNM=NAME
XMIT(6,PRINT}
PRINT=
CUSADD=ADDRl
XMIT(6,PRINT)
CUSADD=ADDR2
CUSZIP=ZIP
XMIT(6,PRINT)

1-56

Records in programs which are loaded directly by the .RUN command are
automatically cleared.
That is, all decimal fields are set to zero
and all alpha fields are filled with blanks, unless an initial value
is specified.
However, if the record is loaded as the result of a CHAIN statement,
the record will retain whatever contents it may have had on the
previous CHAIN unless the clear option (,C) is specified in the

record.
See Appendix L and Section 1.4.3 for more detail on CHAINing records.

1-57

I
1.4.17

RETURN

The RETURN statement is placed at the logical end of a subroutine
has the form:

and

RETURN
The RETURN statement causes control to return to the next
after the last CALL (or call implied by a TRAP statement).

statement

It is illegal to attempt a RETURN if no CALL has been
previously and the message RETURN WITHOUT CALL resultse

executed

Example:

(~L
1 .

LIST

LIST,

1-59

ISTART
1. 4 .18

START

This is an optional statement and can
program.
It is of the form:

inserted

anywhere

the

START[/x] [;comment]
where

is an optional listing switch to the compiler
/N

temporarily suspends listing of source program

resumes listing of the source
errors on the line printer

program

and

resumes listing of the
errors on the terminal

program

and

source

The START option switch is active until disabled
by another option switch. The START/x statement
may be inserted in a source program as of ten as
necessary.
If no option switch is specified a
compiler listing is produced on the line printer.
The switches can be overridden with an option
switch at compile time.
(Refer to Chapter 4.)
comment

is an optional method to store a line of text to
be output as a heading on the compiler listing.

Each time a START statement is encountered during compilation, a
eject occurs and a new page heading is printed.

page

Examples:

START
START;KR3766-CUST UPDATE
START/L
START/T;TEST PROG. LISTING
START;
Illegal
ST
STARTKl 76-REPORT
START/
If /N is specified in the RUN COMP command, it overrides any /L's or
/T's in the source program.
However, /L and /Twill then determine on
what device errors are listed.

1-61

1.4.19

STOP

STOP terminates program execution and returns control to the
The form of the statement is:

Monitor.

STOP
and it may be inserted anywhere in the PROCedure section.
There may
moL"~
than um::: STOP statement in a program. STOP does not close
files; a FINI statement must be inserted before the STOP to close
files previously INITed.

The END statement has the same effect as STOP but an END can
used as the last statement in a program.

1-63

only

ITRACE/NO TRACE
1.4.20

TRACE/NO TRACE

These statements are helpful
anywhere in the PROCedure
statement is:

debugging tools and can be inserted
section of a program. The form of the

TRACE

NO TRACE
Execution of the TRACE statement enables program tracing.
Execution of the NO TRACE statement discontinues
and is optional. The initial mode is NO TRACE.

the

TRACE

function

The appearance of TRACE statements in a program does not cause any
TRACE output to be generated unless the /T option is specified when
running the program.
When program tracing is enabled and the /T switch is
statement executed causes the line

on,

each

DIBOL

AT LINE xxxx
where xxxx is the source line number of the statement to be output on
the line printer.
This tracing continues until execution of a NO
TRACE. If the statement is a data manipulation statement, the value
which is produced and stored by the statement is printed on the
following line.
For example:
AT LINE 0200
000006
NOTE
Indiscriminate placement of TRACE statements may
cause excessive output to be printed on the line
printer - this is often as bad as no output at
all. To use the TRACE statement to best advantage
an attempt should be made to "pin down" the
problem to a certain part of the program flow and
then, using conditional
(IF)
statements, trace
only that part of the program.

1--65

Example:
If a certain DIBOL program works correctly on records 1-53 of a file,
but produces incorrect results on the 54th record and an inspection of
the program fails to reveal the reason, the following statements can
be added:
In the data definition section:

RECNO,D6

at the beginning of the procedure
section:

RECNO=O

after reading an input record:

!NCR RECNO
NO TRACE
IF (RECNO.EQ.54)TRACE

This will provide a complete trace of the operations
processing record 54, and only that record.

1-66

performed

when

l
1. 4 .21

TRAP

TFAP

The TRAP statement allows a DIBOL program to spool output to the line
printer while executing the program. The format of the statement is:
TRAP label
where
label

is the label of a line printer routine.

Whenever the line printer becomes free, DIBOL statement execution
temporarily terminates and a call is made to the label specified in
the last TRAP statement executed.
{If no TRAP statement was executed,
no call is made.) When a RETURN is made from this call, normal program
execution resumes. The TRAP statement normally precedes a FORMS or
XMIT statement.

The following example prints numbers

1~500

some other task is being performed:
RECORD A
N,

PROC
TRAP SUB
FORMS(6,0)

;START LPT
;PERFORM TASK

LOOP,
SUB,

IF (N.LT.500) GOTO LOOP
STOP
N=N+l
IF (N.GT.500) RETURN
XMIT (6 ,A)
RETURN

1-67

the line

printer

while

Typically, the TRAP statement would be used in a program as follows:
RECORD

LP REC

ABO

RECORD
LINCT, D2
LPFLAG, Dl
RECORD HDG
A60, I • • • • • I

PROC n

INIT (l,IN,'INFILE')
INIT (6 ,LPT)
CALL PGTOP
;PERFORM TASK
DONE,

IF(LPFLAG.EQ.O) GO TO DONE

;LPFLAG=O IF TASK IS
;COMPLETED BEFORE
;PRINTING

STOP
LPTRTN, IF(LPFLAG.NE.O) RETURN
LINCT=LINCT-1
IF(LINCT.EQ.O) GO TO PGTOP
XMIT (l,LPREC,LPEOF)
TRAP
LPTRTN
XMIT (6 ,LPREC)
RETURN

PGTOP,

LINCT=SO
; INITIALIZE FOR NEXTPG
TRAP
PGTOPl
;WHERE TO GO NEXT
FORMS(6,0)
;SKIP TO NEW PAGE
RETURN ; GO BACK
PGTOPl, TRAP
PGTOP2
;WHERE TO GO NEXT
XMIT(6,HOG)
RETURN
PGTOP2, TRAP
LPTRTN
FORMS(6,3)
RETURN
LPEOF, LPFLAG=l
FORMS(6,0)
RETURN

1-68

NOTES
1.

If the line printer buffer becomes empty during execution of
an INIT, XMIT, READ, WRITE, DISPLAY or FINI statement while
I/O is in progress the trap will be delayed until execution
of the statement is complete.

If the line printer buffer becomes empty during execution of
an ACCEPT statement the ACCEPT statement will be interrupted
and resumed. The user typing in response to the ACCEPT
statement
will
not
be aware that the statement was
interrupted and the ACCEPTed data will be correct provided
the TP~..P subroutine takes no longer than approximately 150
milliseconds to execute. In the example above the execution
time
of
the
TRAP routine is much shorter than 150
milliseconds if INFILE is on disk.
If INFILE were on DECtape
the TRAP routine might take up to one-half second to execute
and the ACCEPT statement might lose some input.

Always construct a TRAP subroutine so that output to the line
printer is immediately followed by a RETURN statement (as in
the example).

Due to a lack of line printer buffer space in the Monitor,
users with TD8/E DECtape and Centronix line printer are
limited to output lines of 76 characters in length when using
the TRAP statement. Outputting longer lines will result in
the program spending all of its time servicing line printer
TRAPs, thus cancelling the advantage of the TRAP statement.

The TRAP statement will not work on Analex line printers.

In the above example, the main DIBOL program will be slowed
down by approximately 5 to 10% by the TRAP processor
providing that INFILE is on disk.

1-69

jwRITE

1. 4. 22

WRITE

A direct access WRITE statement allows a specified data record to
moved from an area in core to a specified file and has the form:

WRITE {channel,record,rec#)
where
channel

is a decimal expression with the value of 1-15
specifying
channel which links the WRITE
a
(The
statement to the related INIT statement ..
INIT statement must specify UPDATE as the cos
device.)

record

is the record from which data is output.
is a decimal expression specifying
of the record to be written.

rec#

the

sequence

nu~~er

Example:
WRITE ( 5 , REX, 8 8)
returns the 88th record from the core area REX to the device
associated with the channel specified with the INIT (S, •• ~) statement.
NOTE
Because of the harm which the READ and WRITE
statements could cause if improperly used, several
restrictions have been placed upon their use:
1.

2.
3.

4.
5.
6.

The channel involved must refer to a mass
storage device.
The file to be accessed with READ or WRITE
operations must contain records of a uniform
size.
Only one reel of a multi-reel file
(the one
currently
mounted) may be accessed by a
READ/WRITE statement.
The record which is specified must be exactly
the same size as the records of the file being
accessed.
Attempting to READ
or
WRITE
over
the
end-of-file mark results in an error message
and termination.
In general, READing or WRITing a record which
comes after the end-of-file results in an
error
message;
however,
a
fortuitous
configuration of words on the unit being
accessed may cause the operation to succeed
and results in garbled data (on a READ) or the
loss of the output record (on a WRITE) but
does not crash the COS system.

1-71

Unless the user provides a FINI statement
before terminating his DIBOL programs which
have UPDATE files, the data from the last few
WRITE statements may not be output properly.

Ref er to Appendix L for more
access techniques.

examples

1-72

and

discussion

direct

XMIT

1. 4 .23

XMIT

The XMIT statement transfers a data record and is of the form:
XMIT {channel, record[, eof label])
where
channel

a decimal expression whose value is 1-15
specifying a channel number which associates the
XMIT statement with the related INIT statement.

record

is a name previously used in a RECORD statement
which identifies the area in core to which or from
which data is to be transmitted.
It may be a
simple
or
subscripted variable or a record
literal.

One must be very careful
record names.

when

using

subscripted

A single subscript, such as REC(3) should only

be
used if there are consecutive records, beginning
with REC and all of the same length as REC, i.e.,
RECORD REC
,D6

;REC(l)

,AlO
RECORD
,D6

; REC (2)

,AlO
RECORD
,06

; REC ( 3)

,AlO
If a double subscript form is used, i.e., REC
{n,m), then n must be less than m-1, n must be odd
and m must be even (or the last character in the
record).
This double subscript form refers to
characters n-2 through m-2 inclusive in record
REC.
{If n=l, it refers t.o characters 1 through
m-2.) Whenever an XMIT occurs referring to the
record, two characters before character n in the
record are destroyed. This is the COS 300 word
count
(see Appendix G) •
The user need not be
concerned about this if n=l.
eof label

is the label of a statement to which the program
branches if an end-of-file is read.
Used with
input files only.
(Refer to Statement Labels at
the beginning of this chapter.) Label is optional
but if not specified, an error message is output
when an end of file occurs.

1-73

Examples:
XMIT ( 3 , INV, EOF)
transfers a record from the input file associated with the statement
INIT
(3,IN, ••• ), to the RECORD area in core labeled INV.
If
end-of-file is reached, control branches to
PROCedure
section
statement labeled EOF.
If the length of the record being read is
greater than the defined size, an error message is output at run time.
If
the length is less than the defined size, the record is
left-justified and padded with spaces on the right.
XMIT (l,CUST,NEXT)
transfers a record from the input file
associated
with
the
INIT(l,I, ••• ) statement to the RECORD area CUST. At end of file, it
branches to the statement labeled NEXT.
The statement
XMIT (2,BUFF)
takes a record from RECORD area
associated with the INIT(2, ••• )
initialized for output, LPT etc'.) •

BUFF and
statement

puts it in the file
(assuming channel 2 is

operator's

XMIT (8, "HI THERE')
would output the message HI THERE
channel 8 was INITed to the TTY.

the

terminal

XMIT (8,CUST(l,7) ,EOF)
accesses the first five characters of the record area CUST.
Other examples:
Legal
XMIT{N,B2)
XMIT{FUNCT(I)+2,FOO)
XMIT(7,ROD,LAB3)
XMIT(9,BLKNUM,L)
XMIT (3 ,Bl)
XMIT (M,B23A, I)

XMIT(l,HHAA,FFPP)
XMIT ( 8, "MOUNT INVOICE FORMS')
XMIT(lS,REC(l,33))
XMIT(CHAN,REQ(l,INVOC+6))
The XMIT statement may be used with direct access READ and WRITE
See Appendix L for details.
statements for operations on files.

1-74

PART II

SYSTEM AND UTILITY
PROGRAMS

CHAPTER 2
MONITOR
The Monitor is the master control program for the COS system.
It
contains all the system I/O handlers (TTY, DECtape, RF08, RKOB, RKSE,
Line Printer, Card Reader, High-speed Reader and High-speed Punch) and
enables the user to edit, compile, save and execute user programs.
The Monitor maintains a directory of all programs stored on the system
device, labels files and opens and closes files as needed.
During program execution the Monitor produces the messages which
instruct the user to mount files.
It also provides the means for
BATCHing several programs for sequential execution.
Table 2-1 lists the special key commands that are available in on-line
operations.
The editing feature of the Monitor can be used to create SYSGEN tables
and source files on the system device for later use.

2.1

OPERATING PROCEDURES

The Monitor must be loaded via a bootstrap routine each time the
system is restarted.
(Refer to Appendix B.) Monitor signals that is
is loaded into memory by displaying the message
COS MONITOR 2.1108 (or current version number)

type the DAte command, a space, the current date, and the CR key in
the form .DA rnm/dd/yy. For example, .DA 1/25/72. The date must be
entered before proceeding. This date is used during program execution
to date reports, files, and new programs created.
The Monitor
responds to the date with a dot to indicate it is ready to accept any
of the commands shown in Section 2.2.

2-1

TABLE 2-1.
Key

MONITOR KEY COMMANDS
Function

CTRL/C

Returns control to the Monitor. Monitor displays a dot and
awaits a command. If already in the Monitor, CTRL/C has the
same effect as a CTRL/U.

CTRL/N

Turns on imbedded numeric keypad at terminal and can be used
in a BUILD or DIBOL program for numeric input. Type CTRL/N
a second time or return to Monitor and the feature is turned
off.
When the feature is turned on certain keys are
interpreted as numbers as indicated below:
Key

Value

K
L

3
4

NOTE
Keys 7, 8, and 9 have values 7, 8, and
These ten keys form a numeric keypad.

CTRL/O

Suppresses terminal echo of typed output.
If echo is
already suppressed, CTRL/O restores the terminal echo.
CTRL/O is also used to halt and resume output from an LI
command or the compiler. The echo always resumes the next
time the dot is printed.

CTRL/U

Deletes the current input line.

CTRL/V

Suppresses terminal echo of Monitor message, COS MONITOR
2.1108.
The Monitor displays a dot (.) when CTRL/V is
typed.
If message is suppressed, type CTRL/V to enable
message. CTRL/V has the same effect as CTRL/U.

CTRL/Z

Signals the end of input and returns control to the Monitor.
Halts output of line numbers from LN command.

RUBOUT

Erases the last character typed and echoes the deleted
character.
RUBOUT cannot be used in all instances: if the
character to be deleted does not echo when RUBOUT is typed,
type CTRL/U and retype the line.

2-2

2.2

MONITOR COMMANDS:

The commands
BATCH
DATE
DELETE
DIRECTORY
PLEASE
RUN
SAVE
apply to Monitor functions.

Only the

first

two

characters

the

command need be typed.
{R is sufficient for the RUN command.) Any
additional characters up to the first blank are ignored. All commands
must be terminated by typing the CR key. With the exception of RUN,
the Monitor commands have no effect on system programs
(COMP, SORT,
etc.). Refer to Section 2.3 for editing commands.

2-3

2.2.1

BATCH

With the BATCH command a string of Monitor commands can be executed
sequentially.
As soon as the execution of one program ends, another
is automatically started. The form of the BATCH command is:
.BATCH pronam
where
pron am

is the name assigned to a list of Monitor commands
previously created via editing commands and saved
with a WRITE command.

For example:
.0090

RUN COMP 1 JOBl

.0100

SAVE JOB!

.0110

RUN SYSGEN, SY STAB

.0120

RUN JOB!

.0130

RUN JOB2

.0140

RUN SORT, SRCL

.0150

RUN SYSGEN,OLDTAB

.0160

DE JOBl/B

All programs batched must be on the system device. The Batch
can contain another BATCH command only as its last line.

program

When a Monitor command is read from the BATCH file, it is displayed on
the terminal and executed. Type CTRL/C to terminate a BATCH program.
The BATCH must then be restarted at the beginning.
All of the necessary programs and data files must be available; if an
error occurs, BATCH Monitor mode terminates, control returns to
interactive mode, and a dot is displayed on the terminal.
Correct the error in the program being executed and restart the BATCH
program at the beginning or individually type each corranand left to be
executed.
When the BATCH program is finished, control returns to the Monitor and
a dot is displayed on the terminal.

2-5

2.2.2

DATE

The DATE command stores a date to be assigned to any program created
or reports printed until a new DATE command is issued or the system is
restarted. The DATE command is of the form:
.DA

mm/dd/yy

mm
dd

is a number 1 - 12
is a number 1 - 31
is a number 72 - 79

where

If the date entered does not conform to
error message is displayed.

the

above

restrictions,

This command must be entered whenever the Monitor is restarted.
It
may also be used whenever it is necessary to change the system date.
Examples
.DA 1/25/72
.DATE 12/5/73

2-1

2.2.3

DELETE

The DELETE command removes the named source, binary, or system program
from the specified device directory.
The DELETE command is of the form:
.DE pronam[,dev]/X
where
pron am

is the name of the file to
directory.

removed

dev

is the 3-character designation for
device where the file is stored

the

from

the

physical

DECtape = DTO-DT7
disk
= DKO-DK3
If no device is specified, the
assumed.

system

device

Is a 1- or 2-character code indicating whether the
file to be deleted is a source (S), binary (B) or
system (SV) file.
This code is necessary to
differentiate between three files with the same
name but of different types. The code SV is used
rather than V to make it more difficult to
inadvertently delete a system file.

Examples:

•
Legal

Illegal

.DE

JOBl, DTl/B

.DE JOBl

.DE

PROGA,DK3/S

.DE CONVEX/V

.DE

INV/S

.DE CONVEX/SV
See Chapter 3-for a description of device designations.

2--9

2.2.4

DIRECTORY

The DIRECTORY command prints a list of programs stored on
device or the label of the file stored on a logical unit.
the command is:
• DI

a physical
The form of

( [physical device] [/T] ~

l /logical unit#

where
physical device

is one of DTO-DT7 or DKO-DK3 and must be
preceded
by a comma or space.
If not
specified the system device is assumed.

logical unit#

is the number
(0-15) of the logical unit
assigned with SYSGEN. A logical unit # must
be preceded by a /.
Specifying a logical
u..~it

causes a data file label to be printed=

If the unit is a physical device which has
more than one logical unit, the DI command
must be repeated for each logical unit to be
labeled.

is an optional switch which causes
directory to be output to the terminal.

the

A directory contains the current date, names of programs, types of
programs, length in 256-word blocks (LN) and the date the program was
stored. Be sure the line printer is on-line before issuing the DI
command.
A file label contains the file name, sequence number (if a multi-reel
file), the date the file was created, file length in segments, and
logical unit number where the file was stored when the label was
requested. Segments are 16 256-word blocks long.
See Chapter 3 for a description of
unit numbers.

device

2-11

designations

and

logical

Examples:
.DI DKO

causes a directory similar to the following to be printed on the
printer:
DIRECTORY
NAME

TYPE LN

15-FEB-72
DATE

COMP
v 14 19-JAN-72
s 10 15-FEB-72
MORE
<0006 FREE BLOCK8>
s 07 12-FEB-72
TST2
<0007 FREE BLOCKS>
TST4
s 07 15-FEB-72
GLOP
s 10 15-FEB-72
<0579 FREE BLOCKS>
The command
..!.DI/3

outputs a file label similar to one below.

****************************

SEQ.
DATE
*
* NAME
*
*
DEP
#01 18-NOV-75 *
*
*
*
LENGTH: 0046 UNIT: 3 *
*
*
*
****************************

2-12

line

2.2.5

PLEASE

The PLEASE command outputs a message to the terminal
program execution.

during

BATCH

The form of the PLEASE is:
PLEASE text string
The message is displayed exactly as entered and the terminal alarm is
sounded.
After taking the action requested in the PLEASE message,
type any one key to continue the BATCH program. To make a two-line
PLEASE conunand, the first line can be terminated with an AND and the
second line begun with another PLEASE. This lets the operator know
more message is to follow.
For example:
0020
0030
0040
0050
0060

RUN JOBl
PLEASE PUT INVOICES IN LINE PRINTER AND
PLEASE TYPE 3 TO THE NEXT MOUNT MESSAGE.
RUN JOB3
PLEASE PUT REGULAR PAPER IN LINE PRINTER

When this BATCH program is executed, JOBl will be run, the first
PLEASE message will be displayed and the terminal alarm sounded. The
system waits for a key to be typed in reply to the PLEASE message then
displays the next PLEASE message. When a key is typed in reply to the
message, JOB3 is executed and the last PLEASE message displayed.
Control returns to the Monitor when a key is typed in reply to the
last PLEASE message.
If a PLEASE command is given when in a non-BATCH mode, following
CR, the bell rings and the system waits for a key to be typed.

2-13

the

2.2.6

RUN

The RUN command loads and executes the named system or binary program
using the named file. This is the command which provides access to
all other system programs such as:
SYS GEN

To change or check system configuration

BUILD

To build data files

SORT

To sort data files

UPDATE

To update data files

PIP

To move information between physical devices.

COMP

To compile a user source.program into
program.

the

binary

The RUN cornmand has the form:
.RU [pronam] [+chain!+ ••• +chainn] [,filnaml, ••• ,filnamn] [/xx]
where
pron am

is the name of the system program to be run or
pronam+chainl...
are binary files which are part
of a program and constitute one large program
broken up into several chains. For example:
.RUN P~OG+CHAINl+CHAIN2+CHAIN3
would execute program PROG.
PROG would then
determine
whether programs CHAINl, CHAIN2 or
CHAIN3 would next be run. See section 1.4.4 for a
description
of the CHAIN statement.
If the
program name is omitted, Monitor
loads
and
executes the last compiled DIBOL program from the

compiler• s scratch -ar~rn.
filnaml, ••• ,filnarnn
are source files which must be on the system
device.
If one of the system programs is executed
via the RUN command and no source files are
specified as input, the file in the editing
scratch area is used as input (system programs
only).
The maximum number of binary and source files per
program is eight. Multiple files are concatenated
and passed to system programs as one large file.

2-15

is one or a combination of the option switches
including /T which enables the TRACE feature and
/D which enables the DDT mode.

/xx

Other options are described with the programs to which they apply.
Refer to the appropriate chapter for more detailed information on
these programs.
If the program specified is not found, an error message is produced.
The RUN command is the only Monitor command which can
to one character, R.

abbreviated

Examples:
Command

Explanation

.RU

executes previously compiled DIBOL program from
the Compiler's scratch area •

• RU JOBl

runs program called JOBl

.RUN COMP, CHECK

this command compiles the source program CHECK,
prints a compiler listing and temporarily saves
the binary on the system device •

• R BUILD, FD, IO,FILE1,FILE2
this command builds a data file in the format
specified in the Field Descriptor (FD) and I/O
section {IO) of the BUILD control program with the
data in FILEl and FILE2.
MOUNT Messages
The Monitor outputs MOUNT messages to the terminal whenever an input
or output unit must be specified. These messages have the form:
MOUNT xxxxxx #nn FOR INPUT:
MOUNT xxxxxx #nn FOR OUTPUT:
where
xxxxxx

is the name of the data file
program currently executing.

#nn

is a sequence number from 1 to 63.

desired

the

Answer the MOUNT message with the appropriate logical COS unit (1-15).
If an error is made in the reply, type CTRL/U and the correct reply.
The MOUNT message is displayed (and the sequence number incremented)
whenever
the program reaches the end of the device and more
information remains to be read or written.
When default units specified in control programs are not available,
question mark precedes the MOUNT message, i.e.,

2-16

?MOUNT xxxxxx #01 FOR INPUT:
The message
REPLACE xxxxxx #nn ?
is displayed when a file is already stored on the logical unit
specified to a MOUNT message. Answer REPLACE with a Y and the CR key
to replace the old file or any other key to keep the old file.
File
labels beginning with any character other than A-Z, r. It, or J are
considered to be temporary files, and no REPLACE message is displayed
when such files are about to be written over. The following error

messages can occur:
IN USE
ILLEGAL UNIT
NOTE
Occasionally, if output is to a scratch unit, a
garbled filnarn or sequence number in the "REPLACE"
message may be typed. This is because the area on
the unit where the label should be is simply
random characters.
Just answer "YES" to the
REPLACE message as usual.

2-17

2.2.7

SAVE

This command stores the binary program from the compiler scratch
on the named device. The form of the SAVE command is:

area

.SAVE pronam[,dev] [/Y]
where
pron am

is the name to
binary output.

assigned

the

compiler's

dev

is any mass storage device DTO-DT7 or DKO-DK3
which has a directory. If no device is given, the
system device is assumed.

is used to bypass the REPLACE? message and YES
name
would be
response
when
a
duplicate
encountered. Normally used in a BATCH mode to
bypass operator response.

The SAVEd program must be moved to the system device before it can
executed.

2-19

2.3

EDITING COMMANDS

The COS Monitor contains a line editor; that is, all insertions,
changes, and deletions are done with line numbers. All text input to
the Monitor is assigned a sequence of line numbers. For example:
0100
0110
0120
0130
0140
0150
0160
0170
0180
0190
0200

START
RECORD A
A64
Al,
PROC 2
INIT {2 , IN , I MINT I )
XMIT (2 ,A,EOF)
LOOP,
XMIT (8 ,A}
GO TO LOOP
FINI (2)
EOF,
STOP
END

The conunands:
ERASE
FETCH
LIST
Line Number
Number Commands
RESEQUENCE
WRITE
apply to the editing functions. These commands can be entered at any
time in response to the dot displayed by the Monitor. Only the first
two characters of the command need be typed and all commands must be
terminated by typing the CR key.

2-20

2.3.1

ERASE

This conunand erases text from the edit buffer (that portion of core
where the Monitor stores text being edited). The form of the command
is:
.ER [nl] [ ,n2]
where
nl

is the line nUIPber of the line to be erased or the
first line number of two which delimit the lines
to be erased.
If omitted, erasing starts at the
beginning of the text buffer.

is the second of two line numbers indicating where
erasing ends.
If n2 omitted, but the comma is

included, erasing continues to the end of the edit
buffer.
If no line numbers are entered, the ERASE command clears the entire
edit buffer.
Use this command to erase the edit buffer before
entering material to be edited.
Examples:
.ER

clears the entire edit buffer •

• ER 5

clears line 5 •

• ER ,5

clears from the start of the buffer to line 5 •

• ER 5,10

clears from line 5 to line 10, inclusive.

2-21

2.3.2

FETCH

The FETCH command loads the named source file into
specified device. The form of the FETCH command is:

core

from

the

• FE pronam [ ,dev]
where
pron am

is a previously created program
brought into core.

dev

is the 3-character designation of the physical
device, DTO-DT7 or DKO-DK3 where the program is
stored. dev may be separated from pronam by a
space or comma.
If device
assumed.

omitted,

Monitor searches for the named program.
Monitor displays the message

the

which

system

device

If the program is not

found,

FILE NOT FOUND
Retype the command with the correct program name er device.
If another program is run while this program is in the active work
area, Monitor temporarily stores this file in the editing scratch area
on systems device, then returns it to core when program execution is
complete.
Examples:
.FE RICH

Moves program RICH from the system device to the
edit buffer •

• FE PAYROL,DT2 Moves program PAYROL from DECtape unit 2 to the
edit buffer.
.RUN XYZ

Runs a program called XYZ. On completion of this
program, the PAYROL source file is in the edit
scratch area.

The FETCH command erases the edit buffer before inserting the
program.

2-23

FETCHed

2.3.3

LIST

The LIST command outputs the specified lines or the entire edit buffer
to the high-speed punch, line printer, or terminal.
The LIST command has the form:
.LI [nl] [,n2] [/x]
where
nl

is the line number of the line to be listed or the
first line number of two which delimit the lines
to be listed.
If omitted, listing starts at the
beginning of the text buffer.

is the second of two line numbers indicating where
listing ends.
If omitted, but the comma is
included, listing continues to the end of the edit
buffer.

is the one letter code which indicates the
device
/L

line printer
high speed punch
omitted)

(line

output

numbers

If no device is indicated, the list is
the terminal.

output

are
to

If no line numbers are specified, the entire edit buffer is output to
the specified device. Type CTRL/O to stop output from an LI command.
Examples
.LI /L

List entire edit buffer on line printer.

• LI 5

List line 5 on the terminal •

.LI , 5

List from start of buffer to line 5 on the
terminal •

• LI 5,/P

List from line 5 to the end of the buffer on
the high-speed punch.

• LI 5,10

List lines 5-10, inclusive, on the terminal •

2-25

2.3.4

Line Number

The Line Number command automatically outputs line numbers
at
specified increments so new lines can be entered without typing each
line number.
The Line Number command has the form:
• LN

[nj [, incj [ixj

where
n

is the starting line number.
If no starting
number is specified, 100 is assumed.

line

If the corrana after the starting number is omitted,
the starting number and increment number are the
sa~e; i.e.; TN 100 causes a starting nu..~ber and a~
increment of 100.
If the command is LN 100, the start line number is
100 and the increment remains unchanged unless the
Monitor is read back into core; at which time the
increment returns to 10.

inc

is the increment between line nu.~hers.
increment is specified, 10 is assumed.

is the input device.
/R

/C
/T

high-speed reader
(All rubouts are ignored so
that paper tapes with TABS followed by RUBOUTs
will be read correctly.)
card reader
low-speed paper tape reader.

If no device is specified, the terminal
is
assumed.
Line numbers are specified by the LN
command and cannot be part of the text read from
the. inpu.t device.
If Line Number mode is terminated and some editing has been done, type
the line number command, LN, with no arguments to display the next
number in sequence. When inputting from t-hP card reader, all 80
columns are stored (40 columns with mark-sense cards).
This command does not clear the edit buffer. Line Numbers 0 to 4095
are available. Under the default conditions (start at 100, increment
by 10) , the program can be approximately 400 lines long.
The maximum number of characters on a line, including the line number
and space, is 120.
In automatic line number mode the line number and
space are counted as two characters. When the terminal line is full,

2-27

Monitor executes carriage return/line feed and since
continued line does not display the next line number.

this

is a

If the 120 characters limit is exceeded, the Monitor gives
message LINE TOO LONG and the line is lost.

the

error

If a CR is the first character typed, no line is entered and
number increments. To obtain a blank line, type space and CR.

line

Tabs can be used to increase the readability of a program.
The tab
key on most terminals is set to produce up to 8 spaces. The first tab
goes to column 13 because the line number and space take the first
five positions.
Type CTRL/Z to indicate the end of input and halt the automatic line
numbers. When input is coming into the Editor from the card reader, a
beep sounds to indicate no more cards.
Load more cards into the
reader and type any character to continue the input or type CTRL/Z to
indicate the end of input.
Examples:
Requests line numbers starting at
100
with
increments of 10. The terminal is assumed to be
an input device.

.LN

The terminal display would be
.0100
Type the first line of the program and the CR key.
Editor displays the next line number each time the
CR key is typed •
• 0100 START
.0110
.!.LN 10,5/R

Requests line numbers starting
at
10
with
increments of 5 to be assigned to input from the
high-speed reader.

.!.LN ,100

Request line number starting at 100 (default) with
increment of 100. Input is from the terminal.

..!..LN 10/C

Requests line number starting at 10 with increment
of 10 (default). Input is from card reader.

If an error is made when using automatic line numbers, type RUBOUT or
CTRL/U, RUBOUT erases the last character typed and echoes the erased
character. CTRL/U erases the entire line.
Monitor redisplays the
line number. CTRL/C acts the same as CTRL/U.
If the edit buffer is full, the error message TEXT AREA FULL appears,
and the last line entered is lost. The edit buffer must be separated
into two or more source files. This is done by:
1.

WRiting the edit buffer as file B.

2-28

ERasing the last half of the edit buffer.

WRiting the edit buff er as file A.

FEtching file B.

ERasing the first half of the edit buffer.

WRiting the edit buffer as file B.

2-29

!number commands
2.3.5

Number Commands

Any line beginning with a number inserts the line number and
the edit buffer.

text' in

The form is:
nnnn
or
nnnn text
where
nnnn

is a line number.

text

is the data to be inserted in the buffer.
The
data must be separated from the line number by a
space or tab.
(If tab is used, it becomes the
first character of text.)

If text is already at that line number, the new text replaces it.
Files are stored in increasing line number order and a new line number
is inserted in numerical order. If a CR is typed right after the line
number, data at that line number is cleared from the buffer.
The input line is limited to 120 characters including the line
which counts up to four characters.
Examples:
erases text and number at line 40
inserts RECORD at line 45

.40
:;:_45 RECORD

Text before editing:
0035
0040

PROC
IN IT (I , V, IN)

0047

XMIT(6,B)

0060

END

Editing commands:
.35

~40

PROC 1
INIT(l,IN,'LABEL' ,2)

..!.47
Text after editing:
0035
0040

PROC 1
INIT(l,IN,'LABEL' ,2)

0060

END

2-31

number

2.3.6

RESEQUENCE

The RESEQUENCE command renumbers
additions and deletions.

the

program

lines

adjust

for

starting

line

The form of the command is:
• RE [n 1 [ , inc]
where
n

is the starting line number.
If no
number is specified, 100 is assumed.

If the comma after the starting number is omitted, the
starting number and increment number are the same,
i.e., RE 100
causes a starting number and increment
of 100.
If the command is RE 100, the starting line
number is 100 and the increment remains unchanged
unless the Monitor is read back into core; at which
time the increment returns to 10.
inc

is the increment between line numbers.
is specified, 10 is assumed.

If no increment

If line number exceeds 4095. on a resequence, the error message, LINE #
TOO LARGE, results.
Enter another RESEQUENCE command with smaller
increments. If this is not done, the text will be only partially
RESEQUENCEd and duplicate line numbers may result.
Examples:
.RE

Resequences line numbers of program in the edit
buffer using 100 as starting line number and 10 as
increment •

• RE 10,5

Resequences line numbers of program in edit buffer
using 10 as starting line number and 5 as the
increment •

• RE ,100

Resequences line numbers of program in edit buffer
using 100 or the last specified line number as the
starting line and 100 as the increment.

50,

Resequences line numbers of program in edit buffer
using 50 as the starting number and the increment
remains unchanged.

• RE

2-33

2.3.7

WRITE

The WRITE command stores a new or modified source program on the
specified device so it can later be compiled or FEtched for editing.
The form of the command is:
.WR pronarn[,dev] [/Y]
where
pron am

is the six character name assigned to the
to be stored.

program

dev

is the three character designation of the physical
device (DTO - DT7, DKO - DK3) where the program is
to be stored. If no device is specified, the
system device is assumed.

is used to bypass the REPLACE? message and YES
response
when
a
duplicate
name
would be
encountered. Normally used in a BATCH mode to
bypass operator response.

If the name specified is a duplicate name, Monitor asks
REPLACE?
Type Y or YES and the CR key to replace the old file with the new
file.
Type N or any other character and the CR key to leave the old
file and return to the Monitor.

2-35

2.4

EDITING EXAMPLE

.FE

ACC08S

.LI
0010
0020
0030
0040
0050
0060
0070
0080
0090
0100
0110
0120
0130
0140
0150
0160
0170
0200
0210
0220
0230
0240
0250
0260
0270
0280
0290

0500

;LIST PROGRAM TO BE EDITED
START
;ACC08S ;THIS PROGRAM UPDATES A MASTER FILE OF STOCK RECORDS, WITH A
;TRANSACTION FILE OF RECEIVALS AND WITHDRAWALS.
;A NEW MASTER FILE AND A REPORT ARE PRODUCED.

RECORD MAST I
;MASTER INPUT RECORD
D9
DlO
;UNIT COST
DlO
;DATE OF LAST TRANSACTION (MMDDYY)
D6
BLOCK,X
MIWHOL, A35

MIPART,
MICOST,
MIQTY,
MIDATE,

RECORD
D9
DlO
DlO
D6

MAS TO

MOPART,
MOCOST,
MOQTY,
MODATE,

RECORD
Al
D9
D7
D6

TRAN

TYPE,
TPART,
TQTY,
TDATE,

;MASTER OUTPUT RECORD

;TRANSACTION FILE
;TYPE ( R=RECEIVAL, W=WITHDRAWAL)
;PART NUMBER
;QUANTITY
;DATE (MMDDYY)

END
;Editing

!.150
~230

.290
:0090
.0110
:0228
:0229

;PART NUMBER
MIPART, D9
;QUANTITY ON HAND
MIQTY, DlO
BLOCK,X ; REDEFINITION
MOWHOL, A35

2-36

;LIST EDITED PROGRAM

~LI

0010
0020
0030
0040
0050
0060
0070
0080
0090
0100
I'\, ,

I'\

START
;ACC08S
;THIS PROGRAM UPDATES A MASTER FILE OF STOCK RECORDS, WITH A
;TRANSACTION FILE OF RECEIVALS AND WITHDRAWALS.
;A NEW MASTER FILE AND A REPORT ARE PRODUCED.

RECORD
MIPART, D9
MICOST, DlO
'9.•Tl""\m,7

n i f\

MAST I
;MASTER INPUT RECORD
;PART NUMBER
;UNIT COST
• l"'ITT7\ll.TFT1TfT1V

1"111.T

t.17\11.Tn

VJ..J.V

J.Yl.J.\,!J. l. I

0120
0130
0140
0160
0170
0200
0210
0220
0228
0229
0240
02.SO
0260
0270
0280

;DATE OF LAST TRANSACTION {MMDDYY)
MIDATE, D6
BLOCK,X
MIWHOL, A35
;MASTER OUTPUT RECORD
RECORD MAS TO
MOPART, D9
MOCOST, DlO
MOQTY, DlO
MODATE, D6
BLOCK,X
; REDEFINITION
MOWHOL~ A35
;TRANSACTION FILE
RECORD TRAN
TYPE,
Al
;TYPE ( R=RECEIVAL, W=WITHDRAWAL)
TPART, D9
;PART NUMBER
;QUANTITY
TQTY,
D7
;DATE (MMDDYY)
TDATE, D6

0500

END

UJ.V

j~U.CU"l4.J..J.....L..L

VJ..'

.l..1..rl,l.'41L.J

2-37

2.5

BATCH EDITING

As the following example illustrates, editing can be done via a
program •

BATCH

• ERASE
.LN 5,5/R
• LIST

;AUTOMATIC LINE NUMBERS START AT
;5 AND INCREMENT BY 5. INPUT FROM
; PAPER TAPE READER •
;LIST THE FILE READ IN.

0005 FETCH TEMP
0010 LIST
0015 ERASE 4
0020 ERASE 30,34
0025 35 PROC 1
0030 40 INIT(l,IN,'LABEL',2)
0035 47 FORMS (6,0)
0040 RE 5,5
0045 LIST
0050 WRITE TEMPA
0055 DELETE FILE/S

;BATCH Program

• WRITE FILE

;WRITE BATCH PROGRAM CALLED "FILE" •

.BATCH FILE

;RUN BATCH PROGRAM "FILE"

FETCH TEMP

;FIRST COMMAND OF BATCH PROGRAM LOADS
;TEMP FOR EDITING
; SECOND COMMAND LISTS TEMP

LIST
0000 RECORD A
0004
,Dl
0005
,AS
0010
,A6
0015
,D6
0020
,DlO
0025
,A25
0030
,D2
0031 RECORD B
0032
,02
,Al
00-33
0034
,D2,
70
0035 PROC
!NIT ( l , V, IN)
0040
XMIT (l ,A,EOF)
0045
0046
XMIT ( 6 ,A)
0047
XMIT (6 ,B)
0050 EOF,
FINI (1)
0055
STOP
0060
END

; TEMP PROGRAM

2-38

ERASE 4
EP..ASE 30,34
35 PROC 1
40 INIT(l,IN,'LABEL',2)
47 FORMS (6,0)
RE 5,5

LIST
0005 RECORD A
0010
,AS
0015
,A6
0020
,D6
0025
,DlO
0030
,A25
0035 PROC 1
0040 INIT(l,IN, 'LABEL' ,2)
0045
XMIT (l,A,EOF)
0050
XMIT (6,A)
0055 FORMS (6,0)
0060 EOF,
FINI (1)
0065
STOP
0070
END
WRITE TEMPA
DELETE FILE/S

;THIRD COMM&~D ERASES LINE 4.
;FOURTH COMMAND ERASES LINES 30-34.
;FIFTH COMMAND INSERTS NEW TEXT
;AT LINE 35.
;SIXTH COMMAND INSERTS NEW TEXT
;AT LINE 40.
;SEVENTH COMMAND INSERTS NEW TEXT
;AT LINE 4 7.
;EIGHTH COMMAND RESEQUENCES LINE
;NUMBERS STARTING AT 5 AND
;INCREMENTING BY 5.
;NINTH COMMAND LISTS EDITED TEMP.

;STORE TEMPA ON SYSTEM DEVICE
;DELETE SOURCE FILE OF BATCH
;PROGRAM

2-·39

2. 6

ERROR MESSAGES

Monitor Errors
Message

Explanation

BAD COMPILATION

User tried to SAVE
that had errors.

BAD DIRECTORY

compiled

binary

Directory does not
example, requesting
data tape).

look right
a directory

of a

BAD LABEL

Tape has no
incorrect.

form

DT n?

The expected tape is not available on
DECtape drive.

label,

its

(for

For example, if following
a
MOUNT
message the operator has specified that
a file should be written on unit 3, and
unit 3 is not selected, or is WRITE
LOCKed, this message is output.
CTRL commands are not operative at this
point,
i.e., CTRL/C does not work. To
recover, set switch to ENABLE the unit
or set the dial selector to the proper
number.
Then type any key to continue.
On LINCtapes there is no message.
ENTER "DATE MM/DD/YY"

Typed an unrecognizable date.

ERROR IN COMMAND

Miscellaneous.

FILE NOT FOUND

After, for example, FETCH FILEX.

LINE TOO LONG

Greater than 120 characters.

LINE # TOO LARGE

Greater than 4095.

REPLACE?

Duplicate
replace.

TEXT AREA FULL

Greater than 8,150 characters.

?ILLEGAL UNIT

The unit specified is
DTO-DT7; for example,

file

names.

Type

not

DK0-DK3

WRITE FILE,XY3
WRITE FILE, RK9
!*!MEMORY FAIL!*!

A hardware error has occurred.
the program.

2-40

Restart

Explanation

Message
?NO FILE TO SAVE

Nothing in the edit work area when WR~TE
command is issued.

*NO INIT DONE

Program attempted to read or write on a
device that was not opened by the system
program.

Run-Time Error Messages
BAD DIGIT

A character other than +, -, space, or
the digits 0-9 was encountered in an
alpha to decimal conversion;

*BAD PROGRAM

Attempting to run a binary program which
contains a compilation error.
Check
compilation listing for error flags.
Correct flagged errors, and recompile.

*DIBOL FILE NUMBER IN USE

In INIT, the channel

n~~ber

already

associated with a device.
*DIBOL FILE NUMBER NOT INITED
An attempt was made to XMIT,
READ, or
WRITE with a channel numbeL that was not
associated with a device.

END OF FILE

The last record of

input

file

has

been read and the end of file mark
encountered, but no EOF
label
was
specified in the XMIT statement.
* I/O ERROR ON xx, RETRY?

System failed in three attempts to read
from or write to a device. xx = DT or
DK. Type Y if retry is desired; any
other character if no retry and return
to Monitor is desired.
In either case,
type only one character.

*ILLEGAL DEVICE

Attempt to WRITE on a file that was not
INITialized for UPDATE or attempt to
READ from
a
fil.e
that
was
no.t
INITialized for INPUT or UPDATE.

*ILLEGAL SUBSTRING

A DIBOL PROCedure section statement

attempted to access a
(m ,n) , but m=O or m>n.
ILLEGAL RECORD #

has
data field, Fl

Record number is O, past the end of the
physical unit or the length specified in
the record header word does not match
the length of the XMIT block (records in
data file are all not the same length).

2-41

Message

Meaning

LINE TOO LONG

An input line

*NO BUFFERS LEFT

Not enough core available
for
I/O
buffers. An I/O buffer of some multiple
of 512 characters is set up for each
active
mass
storage file.
Another
possibility:
too
few
files
were
specified in the PROC statement.

NO FILE

No file specified in RUN statement
satisfy !NIT (SYS) command.

NUMBER TOO LONG

A decimal field longer than
was used in a calculation.

*PROGRAM TOO BIG

Binary program does not fit in available
core. Reduce program size.

*PUSHDOWN OVERFLOW

Either (1) a statement is too complex or
(2)
subroutines are nested to a depth
greater than 50, or a combination of the
two.

*RETURN WITHOUT CALL

The program tried to execute a RETURN,
but there was no place to go: there was
no corresponding CALL statement.

*SUBSCRIPT TOO BIG

Program
attempted
to
destroy
the
run-time system or itself by using a
subscript larger than that defined in
the Data Definition section. Note that
the run-time system does not detect all
illegal subscripts; only those which
would cause the user's program or the
system to be destroyed.

ZERO DIVISOR

The program attempted to divide by zero.

(record)
block it was read into.

overflowed

NOTES
1.

The messages marked with an asterisk (*) cannot be
checked with an ON ERROR statement.

AT LINE nnnn is displayed under all run-time error
messages.
nnnn is a DIBOL source program line
number in which the error occurred.
If COMP/O
were
specified
for
this
program, nnnn is
meaningless.

2-42

the

digits

CHAPTER 3
SYS GEN

SYSGEN is a conversational program used to reconfigure the software
system to handle the hardware available, change the logical unit
assignments and print a table of those assignments for reference.

3.1

OPERATING PROCEDURES

To load SYSGEN type:
.RU SYSGEN[, filnam]
or
.RU SYSGEN/x
where
filnam

is optional and is a file previously created and
stored with the Editor which contains the table of
logical unit assignments to be made.

is one of three options used with SYSGEN.
/C

change the hardware handlers in the system.

take new
terminal.

print
table
assignments.

logical
of

unit
current

assignments

from

logical

unit

If no file or options are specified SYSGEN takes the information
stored in the edit buffer as the new logical unit assignments.
Refer to the appropriate section for
further
reconfiguration and logical unit assignments.

3. 2

explanation

SYSTEM SOFTWARE CONFIGURP.TION

As shipped, the COS software system is set up to handle the
hardware configuration on the PDP-8:
TC08 with 8 DECtapes (or TDSE for the 8/E)
RK8E, RKOB or RFOB disks
Terminal
LPOB BO-column Line Printer
Card Reader
High-speed Paper Tape Reader and Punch

3-1

following

or
LINCtape (PDP-12)
RKBE, RK08 or RF08 disks
LP08 80-column Line Printer
Terminal
Card Reader
High-speed Paper Tape Reader and Punch
on the PDP-12.
NOTE
No change is necessary if your COS hardware
configuration matches the one described here.
If, however, the system has either a 132 column line printer or an
Analex line printer or the system device is to be changed, type the
command
.RU SYSGEN/C
and the CR key.
(Refer to Figure 3-1 for a flow chart of the question
and answer sequence.)

_:_ RU SYSGEN/C

WHAT IS NEW
SYSTEM DEVICE

ASSUME 80
COLUMN PRINTER

SYSTEM IS
RECONFIGURED
AND CONTROL
RETURNS TO
MONITOR

Figure 3-1.

SYSGEN Flow Chart

3-2

SYSGEN displays the following questions on the terminal:
DO YOU WANT TO CHANGE THE SYSTEM DEVICE?
Type NO and the CR key to keep the current
skips to the questions on the line printer.

system

device.

SYS GEN

Type YES and the CR key to change the system device from DECtape to
Disk or Disk to DECtape. This is the only way to move the Monitor and
DIBOL run time system (RSYS) • If the system device is changed, the
system must be restarted with the bootstrap appropriate to that device
(see Appendix B or Chapter 9, BOOT). SYSGEN then asks:
WHAT IS THE NEW SYSTEM DEVICE?
Type the three character designation of the new system device, DK0-DK3
or DTO-DT7.
NOTE
Since only an RK08, or RK8E, or RF08 disk may be
resident
on
the system, disks are commonly
addressed as DKO-DK3. An RF08 disk must always be
addressed as DKO.
DO YOU WAi.~T TO TRANSFER YOUR FILES?

Answer YES and the CR key to move the rest of the system programs
(PIP, COMP, etc.) to the new system device. Transfer of the system
programs destroys anything stored on the new system device. Answering
NO and the CR key will zero the directory of the new device.
DO YOU HAVE AN ANELEX PRINTER?

NOTE
The Anelex printer must be configured into the
system with SYSGEN before it is turned on or it
will not function properly.
Answer YES and the CR key and the next question is
IS EVERYTHING CORRECT?
Answer NO and the CR key and SYSGEN assumes an LP08
and the question

LS8E

printer

132 COLUMN PRINTER?
is displayed. Answer YES and &he CR key if there is a 132 column line
printer.
Answer NO and the CR key and SYSGEN assumes an 80 column
line printer. The next question is:
IS hVERYTHING CORRECT?
Check all answers made to the above questions. Type YES and the CR
key if the answers are correct. Type NO and the CR key to correct any
errors. SYSGEN repeats the questions starting at the beginning.

3-3

SYSGEN/C does not reset the logical unit assignments to reflect the
new area occupied by the system on a disk. Run SYSGEN/T to reassign
logical units if the system is moved to a disk or the system may be
destroyed by a data file.

3.3

LOGICAL UNIT ASSIGNMENT

The assignment of logical units to physical mass storage devices
provides greater utilization of the storage area and a certain device
independence. The device independence is available at run time; any
mass storage device can be specified for I/O and the program executes
using the specified devices.
The COS system handles storage using the following hierarchy
2 characters = 1 word
256 words
= 1 block
16 blocks
= 1 segment
Refer to sections 3.3.l and 3.3.2 for background on logical unit
assignment before proceeding.
Devi~es
not currently part of the
system configuration can not be assigned a logical unit.
To make logical unit assignments from the terminal, type
• RU SYS GEN /T
and the CR key. Type the 3 character device code, a comma and length
in segments in the order they are to be assigned to the logical units.
The comma and length in segments can be omitted and SYSGEN assigns all
the space left on the specified device to the next logical unit
(DECtape = 46 segments; RK08 Disk, 202 segments; RK8E, 404 segments;
RF08 disk, 64 segments times the number of platters). The first
device entered is automatically assigned to logical unit 1, the second
to unit 2, etc. For example,
.RU SYSGEN/T
ful,46
DT2,46
OT3~46

DT4,46
DT5,46
DTG,46
DT7,46
DK0,46
DK0,46
DKO

;logical unit 10 is 110 segments (assuming DK is
;the RK08 disk and DECtape is the systems device)

assigns DTl to logical unit 1 with a length of 46 segments, DT2 to
logical unit 2, etc.
Logical units 11-15 are left unassigned. Be
sure to follow the format shown in the example.
If an error is made
when typing the table, type CTRL/C then retype the RUN SYSGEN/T
command and start the table over again.

3-4

Logical assignment can also be accomplished by typing the table into
the Monitor.
This method all~ws error correction with the Monitor
commands. If the following logical unit assignments are typed to the
Monitor,

0010
0015
0020
0025
0030
0040
0050
0060
0070

DTl
DT2
DT3
DT4
DT5
DT6
DT7
DK0,46
DK0,46

and the command
• RUN SYSGEN
and the CR key is typed, SYSGEN makes the logical unit assignments
according to the table in the edit buffer. Logical units 1-7 are
DECtape units 1-7 with a length of 46 segments and logical units 8 and
9 are portions of RK08 disk (DKO} ~ each 46 segments long. A listing
of the SYSGEN assignments is always printed on the line printer upon
successful completion of the SYSGEN program.
The table created with the Monitor can also be stored with
command for future use. Use the command

the

WRITE

file.

For

,.RU SYSGEN, file
to assign the logical units
example

with

previously

created

.RU SYSGEN,TABLl
where TABLl contains

0010
0020
0030
0040
0050
0060
00.70
0080
0090
0100

DTl
DT2
DT3
DT4
DTS
DT6
.DT7
DKl,202
DK2,101
DK2,101

Occasionally, it is desirable to check the logical
table. To obtain a printout of the table, type

device

assignment

.RUN SYSGEN/L
and the SYSGEN table of the current logical
printed on the line printer. For example:

3-5

device

assignments

LOGICAL
UNIT #
01
02
03
04
05
06
07

08
09
10
11
12
13
14
15

DEVICE
NAME

LENGTH
(SEGMENTS)

DTl
DT2

0046
0046
0046
0046
0046
0046

DT3
DT4
DTS
DT6

-UNDEFINED-UNDEFINED-UNDEFINED-UNDEFINED-UNDEFINED-UNDEFINED-UNDEFINED-UNDEF INED-UNDEFINED-

This table should be made available for reference by all persons using
the COS system.
Physical mass storage devices are limited by the hardware; but can be
subdivided
into logical units.
Logical units are composed of
segments'; the number of segments per unit is up to the user.
COS
allows one file per logical unit.
The COS system is shipped with logical units 1-7 assigned to DECtape
units 1-7 and each unit is originally assigned 46 segments. Logical
units 8-15 are left unassigned.
The user may assign and reassign any
of the 1-15 logical units as desired.
Logical unit 0 is always the
system device and cannot be reassigned.
It is advisable to create logical units of varied sizes since file
lengths vary and a short file in a long logical unit wastes storage.

3.3.1

DECtape Users

Although more than one logical unit can be assigned to a physical
DECtape unit, it is ac,lvisable to assign one logical unit per DECtape.
The use of two or more units per DECtape may cause excessive tape
spinning; for example, if reading from one unit and writing to another
located on the same physical device.
In the same example, if the
output unit filled and had to be replaced, it might be necessary to
dismount the tape, thereby also removing the needed input unit.
A DECtape is 46 segments long so the logical unit assigned
physical DECtape unit should be assigned 46 segments.

3.3.2

the

Disk Users

The easiest method of assigning logical units to a disk is to think in
terms of a sequence of DECtapes.

3-6

A DECtape is 46 segments long.

1
DECtape

A RF08 platter is 64 segments; and a second platter
segments.
64 65

1
T'\T".'11"\t'\

J:U.'VO

platter

another

128
RF08 platter

adds

I
I

The second platter is treated as a continuation of the first.
The RK08 disk cartridges are each 202 segments and there can be up
four drives {0-3).

202

DKO
202

DKl
202

DK2
202

1
DK3

The RKBE disk cartridges are each 404 segments and there can be up to
four drives
{0-3).
If the disk is the system device, approximately
200 blocks (or 12 segments) must be left unassigned to hold the
operating system and system programs.
In addition, space should be
left on the system device to store source programs, control programs
and binary programs.
The remainder of drive 0 may be used for this
purpose.
However, fifty segments are
suggested
which
leaves
approximately
140
segments
(342
for RK8E)
for logical unit
assignments. Logical units on the system device are automatically
assigned at the end of that device (the opposite is true of all other
devices). For example:

3-7

DKO (assuming RK08)
12

SYSTEM
PROGRAMS

202

Space for user
source and
!binary files

Logical Unit 0
which leaves approximately 140 segments for logical unit assignments.
A sample assignment might be
DKO (assuming RK08)
202

User source
& Binary
Files

SYSTEM

46
SEGMENTS

"""'
Logical
unit
0

46
SEGMENTS

Logical
unit 8

Logical
unit 9

46
SEGMENTS

Logical unit
10

The area between logical unit 8 and the system can be left unassigned
for system program overflow, if desired. On devices other than the
system device SYSGEN starts assigning logical units at the beginning
of the device.

3.4

ERROR MESSAGES
Message

Explanation

# TOO LARGE

Number entered was greater than 4095.

ALL UNITS ASSIGNED

Attempt to assign more than 15 units.

BAD SWITCH

Not /T, /C, or /L.

ILLEGAL DEVlCE

Device other than DTO-DT7 or DKO-DK3.

INSUFFICIENT SPACE ON DEVICE
Attempt to allocate more segments
there are available on a device.

than

NOT ENOUGH ROOM FOR SYSTEM AND FILES
Device
specified
too
small
to
accommodate system programs and files.
SYNTAX ERROR

Missing corn.ma., extra characters, etc.

3-8

CHAPTER 4
COMP

The Compiler takes a DIBOL language program and converts it to a
binary language program which can be executed by the run time system
(RSYS). In the process of creating the executable program, the
Compiler outputs a listing of the source program and a storage map of
the records and fields which are used by the program.
Comniler sets aside storaae soace for the constants, variables and
stat~~~~t~-- ~s~d by the program.
Refer to Appendix c for an
explanation of the amount of code generated by the Compiler.
Th2

Be sure to document any new programs completely and accurately and to
prepare a set of instructions for execution including any necessary
changes to logical device assignments, names of files, and where each
is to be mounted, the format of any entries to be made to the program,
restart procedures and any other essential information.

4.1

OPERATING PROCEDURES

The command to run the Compiler has the form:
• RU COMP [, filel, ••• , file?] [/xx]
and the CR key.

(Be sure the line printer is turned on.)

Where
filel, ••• , file?

are files on the system device to be compiled into
one binary.
These files contain segments of one
program, not several programs. If no files are
specified, the program in the edit scratch area is
compiled.

/xx

is one or a combination of the following options:
the

suppresses the output of
listing and storage maps.

compiles the source program and if no errors
are detected, executes the binary program
with no user intervention. /N is implied.
The
message
LOADING is displayed when
compiling is successfully completed.
The
program must have an END statement to be
compiled and executed with the /G option, if
!NIT SYS is used in the program.

/1-7

specifies a physical unit (DECtape only) for
temporary storage during compilation. The
unit specified should contain a scratch
tape. File is recopied on the system device

4-1

Compiler

when compilation is complete.
Use
of
temporary storage units minimize tape motion
during compilation.
/T

enables the TRACE function and
/G option. For example,

implies

the

RUN COMP/T2
/D

performs an implied /G option and then
chains
to
the
DDT
(DIBOL
Debugging
Technique) mode which is
explained
in
Appendix H.

compiles the source program in less than the
amount of memory space required by the other
options, at the expense of eliminating the
TRACE feature
(if stated in the source
program)
and accurate
error
reporting.
Execution speed of the compiled program is
increased by as much as 20%.
This option
can be combined with either the /N, /G, /D,
/T, or /1-7 option.

Use the SAVE command to store
Compiler.

the

binary

program

prepared

the

option,

To calculate how much memory space is saved by using
the following rules apply:
1.

Saves 1 location for each executable statement.

Saves 1 additional location for each label.

Costs 1 location for each ON ERROR statement.
ERROR statements are executable.)

(However,

the

Since the TRACE statement in a program compiled with the /0 option has
no useful effect, and since any run-time error reported in such a
program will be reported to have occurred in the PROC statement, it is
suggested that the /0 option only be used on those programs which have
been thoroughly debugged.
Unless the /N or /G option is specified in the RUN command, the
Compiler produces a listing of the source program (Figure 4-lj and a
storage map (Figure 4-2} on the line printer or the device specified
in the START, PROC or END statements. Most errors detected by the
Compiler are indicated by underscoring the number of the line where
the error exists and inserting a caret ( ) pointing in most cases to
the error. Other errors are listed in the symbol map.
These errors
must be corrected before an attempt is made to execute the program.
The number of errors detected by the
terminal as
nn ERRORS

compiler

displayed

the

CQ~

01801..

0l•MARw7~ THU~

DATA Df VISlQN

ST4RT

00i0
'5i13il
HUI

0IU

00u

""'"

HU
H90
0~~2'

0u'
"UZ

COMPILATION LISTING
VlZJ28 PAGE 01
R 0 0 T S UB R 0 U! t N£

s Q U~ R£

s Q uA~ t

R0 0 T s ua R 0 uT I Ne

Rf;eORD PRIN!
A15
AU

RECORO

SQRTX,

Ol.5
0:1.5

TE;MPX1
Na

015
03 1 Hi

FIGURE 4-1.

COMPILER LISTING

(sheet 1 of 2)

4-3

QQS OIBOL
~!•MAR•73 TMUR
PROCEOURE OIV!SIQN
9J~H
e~H

rop0

INtTC11~PTt

L,OOP1

s ~ SQRTx,1xxx;xxx!

e;90

INCR N

0?00
0~U

XHlTCl1PRlNP

If cN;L~~99) GOTO ~OOP
STOP

022il
0~;se

"2,2'
0260
0~10
0~BrD
~290

0 ~""

X:!!N•1HHH
CAL:L. SQRT

AIN

0~U

02421

VUt28

PROC 1

"~421

0u0

COMPILATION ~!S,tNG

SQRT,

SQ~OOP 1

ENO

TEMPX!K/2

SQRTX~CTEMPX+(X/T~MPX))/Z

If CSQRTX•TEMPX 1 EQ 1 Z) RETURN
TEMPXeSQRTX
GO'l'O SQl.,OOP

FIGURE 4-1.

JTRlAL VA~UE

JNEWTON'S METMOO

COMPILER LISTING

(sheet 2 of 2)

4-4

PAfil l!2

cos DI 801,,.

THUR

NAME

TYPE

STORAGE MAP L.ISTING
DIM SilE
OR lG IN

0003.

REcORo

00'!!~

Al,P~A

01
01

~1,,."1AR""73

0003
0004

s
x

0005

SQRTX

0006
0007

TEMPX

00UJ

Al..PMA
DECHAL.
DECMA~

OECMAI..
DE CM Al,.

N
t ·,1

OE CM AL,

~OOP

01
01

15
15

211
01

~0117

L..ABEL,

OECMAL.

021!3
0014

SQRT

L.ABEL.
AL.P~A

OECMAl,.

"~
0'1

, , 99

£11
00
01

0016

• ·, 2

OE CM A~

0017
02120

SQL.OOP

L.ABEL

.• ·, 0

OECMM'!

001;

H2,6

•• xxx

2000~

2"2121
20042
2H61
20u0

20122
1H71
20123

~ t 1000

'"' 1l.
00~~

PAG£; 03

2H0~

V1Zl28

Ut.3~

20132
2"'14:1

10141

20144

~U43

SVM90~S

NO ERRORS DETEC re:o~,

08 K eoRE REQUIRED C3940 F'REE:

FIGURE 4-2.

STORAGE MAP LISTING

4-5

~ocs

• 14 BVH'E~SJ

The storage map lists the names of all the records, fields, and labels
as they were processed by the Compiler. The information is arranged
in six columns as follows:
this column contains the internal number of the symbol named
in column 2 and may be ignored.
NAME

Column 2 gives the name of any symbol
(field name, record
name, program label)
or constant used in the compiled
program.
Constants are
decimal
constants
or
alpha
characters which appear in the PROC section of the source
program. Only the first six characters of a symbol name are
used;
any characters after the sixth are ignored.
Only the
first four characters of a constant are used and each
constant is preceded by two periods ( •• ) to distinguish it
from a symbol.
Constants with four characters or less
appear only once on the symbol map even though they may
occur more than once in the program.
Constants with more
than four characters are listed each time they occur in the
program. Record literals begin with a" •• ".

TYPE

Column 3 states the symbol type.

The types are:

ALPHA

Symbol is an alpha field name or an alpha constant
(character .string).

DECMAL

Symbol is
constant.

RECORD

Symbol is a record name or record literal.

LABEL

Symbol is a program label.

REDEF

Symbol is multiply defined
(redefined).
All
attempts at definition after the first are flagged
as errors in the source listing.

UNDEF***

Symbol is a label referenced by the program but
not defined. For example, GO TO TAG! in a program
where TAGl does not appear as a label.

decimal

field

name

decimal

This error is output to the line printer even
though the /N option is in effect. The line
number displayed is the line where the symbol is
first used.
DIM

Column 4 contains the array dimension (number of fields)
of
the alpha or decimal symbols. The column is meaningless for
other types of symbols.

SIZE

Column 5 lists the size of the symbol in characters.
The
size of a RECORD symbol is the number of the characters of
all its symbols plus 2.

ORIGIN

Column 6 gives the 15-bit (octal) byte address where the
symbol appears in core and for most users should be ignored.
Octal addresses are shown for labels.

The number of symbols used, number of errors detected, core required
and free locations are listed at the bottom of the symbol map.

4--6

Maximum number of symbols allowed in an BK system is 365;
larger systems, 511.

4.2

12K

CONDITIONAL COMPILATION PROCEDURE

The Conditional Compilation Procedure (CCP) is a feature which permits
the DIBOL programmer to include statements in a source program which
will be compiled by the compiler only if the programmer elects for
those statements to be compiled.
Statements included in

program

for

conditional

compilation

are

enclosed with angle brackets as in the following example.
RECORD A
Bl,
DS
Cl,
A4
PROMPT, Dl
RECORD N
NAME, AG
PROC
<PROMPT
XMIT(8,"ENTER NAME:")

XMIT ( 7 ,N)
STOP
END

Notice that the left angle bracket is followed by a control variable,
in this case PROMPT.
Until the control variable is turned "on"
somewhere in the program before the left angled bracket is encountered
by the compiler, those statements between the angled brackets will be
ignored. Notice too, that the right angled brackets appears on a line
by itself. The command to turn on a control variable is as follows:
=control variable
The example above shows a program that requires the operator to type
in a name on the keyboard. If the program is recompiled with the
control variable PROMPT on, it produces a DIBOL program which first
displays a message to the operator.
RECORD A
Bl,
05
Cl,
A4
PROMPT, Dl
RECORD N
NAME, A6
PROC
=PROMPT
;TURN ON PROMPT
<PROMPT
XMIT(B,"ENTER NAME:")

XMIT ( 7 ,N)
STOP
END

4-7

Conditional compilation can also
be
used,
for
example,
to
conditionalize debugging statements into a source file. Once the
program has been tested, the control variable can be shut off by
deleting the command to turn it on,· effectively removing the debugging
routines from the source.
CCP can also allow you to combine several similar (but not
source programs into one source.

identical)

If the control variable used in a CCP statement is undefined, the
compiler will automatically set aside space for it. This, however, is
wasteful of space and it is to the programmer's advantage to use, for
CCP symbols, symbols that are already being used for some other
purpose.
The CCP value of a symbol (on or off} is independent of
ordinary DIBOL value.

the

symbol's

If a CCP variable is used in the middle of a record definition (in the
record section of a DIBOL program} it must be a variable that has
previously been defined, because if not, the compiler will allocate
space for the variable in the middle of your record (where not
desired}.
CCP sections may be nested to any depth. Any sections inside a CCP
section that is off will be completely ignored by the compiler. This
includes other CCP statements. To indicate that certain statements
are being ignored (conditionalized out), the compiler listing will not
print the line number for that statement. There must be a matching
for each <used.
If this condition is not met, the compiler generates
a CCP ERROR message. This error is fatal if angled brackets do not
match by the end of the program.

4.3

ERROR MESSAGES

Most Compiler error messages are printed on the source listing
directly after the line in which the error occurs. A caret CA> in the
error message points to the approximate location of the error.
Other
errors are listed in the symbol map.
The Compiler error messages are:
Message

Explanation

BAD ALPHA VALUE

Initial
value
in
an
alpha
data
specification statement did not begin or
end with a single quotation mark.

BAD DECIMAL VALUE

The initial value for a decimal data
specification was incorrectly formed.

BAD PROC #

The number in a PROC statement was not a
digit from 0 to 7.

BAD RELATIONAL

An illegal relational occurs in an IF
statement. For example, a .GX. instead
of a .GT.

4-8

Message

Explanation

CCP ERROR

Matching
missing.

COMMA MISSING

No comma
expected.

DATA INITIALIZATION MISSING

No data initialization followed a comma
in a data specification statement.

EXPECTED LABEL IS MISSING

A required label is missing.

EXPRESSION NOT ALLOWED

A complex expression or bad character
occurs to the left of an = or where only
a variable is allowed.

EXTRA CHARS AT ST.MNT END

Extra characters occur at the end
legal statement.

FIELD TOO LARGE OR 0

In a data description statement, the
dimension was 0 or more than 3 digits
long, or the field size was 0 or larger
than 511.

ILLEGAL OPERATOR

A bad character was encountered in an
expression where an operator would be
expected.

ILLEGAL STMNT

The
statement
was
not
a
data
manipulation statement (it had no =) nor
did it start
with
a
recognizable
keyword.

INITIAL VALUE WRONG SIZE

The
initial
value
in
a
data
specification statement had a length
different from the field size specified.

LABEL NOT ALLOWED

A symbol in an expression was not of
type alpha or decimal, or a record or a
symbol which had been redefined was
used.

MISSING CLOSE PAREN

angled

bracket

appeared

where

one

No close parenthesis occurred where

was

one

was expected.

MISSING OPEN FAREN

No open parenthesis occurred
was expected.

MISSING OPERAND

A
bin~ry
operator
occurs
in
an
expression with no operand following it;
or no expression at all occurs where one
is expected.

MISSING OR BAD DEVICE

The device in an INIT statement was
missing
or started with an illegal
character.

4-9

where

one

Message

Meaning

MISSING QUOTE

The statement contained an odd number of
quotes (').

MISSING RELATIONAL

No relational
statement.

NAME PREVIOUSLY DEFINED

The name used was previously defined and
this statement tries to redefine it.

NOT A OR D

A character other than A or D occurred
in a data specification statement where
A or D was expected.

NOT LABEL

A symbol which was not
a
'label'
occurred where a label was required.

PROGRAM TOO BIG

Binary output too big for the binary
scratch area.
Remedy: Run PIP with /n
option, (refer to Chapter 5).

RECORD TOO BIG

A named record exceeded
in size.

STMNT TOO COMPLEX

The statement generated too much code
and
overflowed
the Compiler's code
buffer, or it had too much nesting and
overflowed
the
Compiler's push-down
stack. Remedy: break up the statement
into smaller parts.

SUBSCRIPT ERROR

No conuna or close
after a subscript.

SUBSCRIPT NOT DECIMAL

The type of a subscript was not decimal.

TOO MANY SYMBOLS!

A fatal error message. Only 365 symbols
allowed in symbol table in 8K system,
and only 511 symbols allowed in larger
systems.
The compiler stops compiling;
no storage map can be produced.

TOO MUCH DATA

Program's
bytes.

UNDEFINED NAME

A name is used which was
in the data section.

WRONG DATA TYPE

Mixed modes occurred in an expression;
or an argument which was supposed to be
decimal was not or one of the three
arguments
in
a
data
manipulation
statement was of the wrong type.

4-10

Data

appeared

510

characters

parenthesis

division

occurs

exceeds
never

32K

defined

CHAPTER 5
PIP

PIP (Peripheral Interchange Program) handles file tranfers between two
logical devices, i.e., moves data, source and binary files, and system
programs between two file-oriented devices; copies the contents of one
device to another, and consolidates files to remove free blocks. PIP
can also be used to allocate more space to the binary scratch area.

5.1

OPERATING PROCEDURES

Type the Monitor command:
.:.RU PIP [ /n]
and the CR key to load PIP. /n is optional and is a number 0-9 which
indicates the number of segments to allocate to the binary scratch
area. The /n switch is used in conjunction with the E option.
Refer
to section 5.1.4 for further information.
PIP responds to the RUN command with:
PIP 2.1108 {or current version number)
QE1.::..
requesting the type of transfer and I/O devices. Answer with one of
the options, B, c, D, E, s, V, X, described on the following pages.
After the option is specified, PIP displays the IN and OUT questions
requesting the input and output files. If an incorrect character is
entered when responding to the OPT, IN or OUT questions, type CTRL/U
and the correct response. Type CTRL/C to return to the Monitor.

5-1

OPT- B

5.1.1

Transfer Binary File

Type B to move a binary program between two file-oriented devices.
Answer IN with the name of the binary program to be transferred and,
optionally, a comma, and the input device, DTO-DT7, DKO-DK3. If no
device is specified, the system device is assumed.
Answer OUT with the name to be assigned to the output file, and,
optionally, a comma, and the output device, DTO-DT7, DKO-DK3. If no
device is specified, the sys~em device is assumed.
If an attempt is made to transfer to or
device, the IN or OUT message is repeated.
Example:
OPTINOUTOP T-

TEST,DTl
TEST,DKl
(CTRL/C typed here)

5-3

from

non-file-oriented

OPT-

5.1.2

Copy Device

Type C in reply to OPT- to copy the contents of an entire device onto
another similar device, i.e., disk to disk or DECtape to DECtape.
Answer IN
DKO-DK3.

with

the

three-character

Answer OUT with one of the physical

physical

DTO-DT7,

as explained for IN.

Example:

c
OTO

DT7
(X or CTRL/C typed here)

5-5

device,

OPT- D

5.1.3

Transfer Data Files

Type D to move data files
OUT.
Answer IN with a
must be alphanumeric), an
by a slash or comma, or a

/R
,,.,
,.._
/K
/F

between the devices specified after IN and
label (up to six characters, first character
optional logical unit number (1-15) preceded
device switch:

for paper tape reader

for
for keyboard
BCD paper tape in high speed reader

If /R is specified as the input device, records stored are variable
length.
If /C, the record produced is the same size as the card even
if a smaller number of columns are punched.
If a label is specified without a logical unit number or if both
specified and a data file of another name is found, PIP asks
MOUNT

xxxxxx

are

#01 FOR INPUT:

answer with the appropriate logical unit number
(1-15).
Answer OUT
with a label, number or device as described for IN. Output devices
are:
/L
/P
/T

line printer
paper tape punch
terminal

If a label is specified without
message
MOUNT
is displayed.

xxxxxx

default

unit

specification,

the

#01 FOR OUTPUT:

Answer with the appropriate logical unit number (1-15).

If a file with a different name already exists, Monitor asks
REPLACE

xxxxxx#nn?

Type Y and the CR key to destroy old file and replace with new.
Type
N and the CR key to return to ~'le OPT- quest:!. on.
If output device
specified is a logical unit (1-15) and the unit is filled,
an error
message is produced.
When at the end of the input file or card deck, PIP asks
MORE?
Answer with the CR key if there is no more input, or
label, logical unit or device.

5-7

specify

another

NOTE
PIP transfers data in alphanumeric mode only.
A
negative number is treated as the letter which has
the equivalent code.
(Refer to Appendix A, COS
Character Codes.)
Examples:
Explanation
OPTINMOUNT
1
OUTMORE?
OPT-

OPTINOUTMORE?
IN_MORE?
OPT-

OPTINOUTMOUNT
10
MORE?
INMOUNT

EMPNAM
EMPNAM #01 FOR INPUT:

Dump EMPNAM
file on logical
unit 1 on the
line printer.
CTRL/C is typed to
return to the
Monitor.

/L
N

Combines 2
data files into
one output
file.
CTRL/C is typed
to return to the
Monitor.

HRPAY,2
PAYFIL/l
y

SALPAY,3
N

/C
LIS Tl
LIS Tl

#01 FOR OUTPUT:

/C
LISTl

#02 FOR OUTPUT:

MORE?
OPT-

Input from
card reader which
creates a
multi-unit
file.
Logical unit 10
is full.

CTRL/C is typed
to return to the
Monitor.

5-8

loPT- El
I
I
5.1.4

Eliminate Spaces in Directory

Type E to consolidate the free blocks of the file on the input device,
and store the consolidated files on the output device. In addition,
it is possible.to eliminate one or two of the types of files
(source,
binary, and system) during the consolidation. These free blocks are
shown in the file directory and occur when a file is deleted f rorn the
directory. The Monitor is not copied.
Answer IN with DTO-DT7, or DK0-DK3 and the CR key.
Answer OUT with DTO-DT7, or DK0-DK3 and the CR key.
When consolidating the system device onto itself, PIP
free space as shown below:

eliminates

the

SYS
f

before

files

r
e
e

files

r
e
e

files

logical
unit

logical
unit
10

SYS
after

files

free

logical
unit
10

However when consolidating a device {other than the system
onto another device PIP eliminates the free space as follows
DKO
before

files

f
r
e
e

files

f
r
e
e

logical
unit

DKl

after

files

free

Example:
OPTINOUT-

E
DTO
OTO

The message
TYPES OF FILES TO BE SKIPPED (S,V,B):

5-9

logical
unit
10

device)

is displayed. Answer with the CR key if all files are to be
transferred, or answer by typing one or a combination of two of the
characters s, V, B separated by a comma and followed by the CR key if
one or two of the types of files are to be eliminated.
For example:
~R

PIP)
PIP 2.1108
OPT- E)
IN- OTO)

OUT- DKO)
TYPES OF FILES TO BE SKIPPED (S,V,B): S )
This would consolidate free space and eliminate source files located
on OTO and store the consolidated files on DKO assuming there was a
directory on DKO.
NOTE

All control keys are ignored until
operation is completed.

the

option

Binary Scratch Area Modification:
As mentioned in section 5.1, the E option can be used in conjunction
with /n to change the size of the binary scratch area. The binary
scratch area, normally two segments (32 blocks) long, is adequate for
most applications but extremely large DIBOL programs (containing many
statements) may need more space.
A typical sequence for expanding the binary scratch area is:
.R PIP/2)
PIP 2.1108
~ E)
IN-

DTO)

OUT- DKO)
TYPES OF FILES TO BE SKIPPED (S,V,B): )
This would allocate 64 blocks to
doubling its original size.

the

binary

scratch

area

DKO,

In genera.1, _

If the output device is not the input device, the binary
scratch area on the output device equals 2+n segments.

If the output device is the input device, the binary scratch
area is set either to the size of the current area or to 2+n,
whichever is less. That is, compressing a device onto itself
can shrink the binary scratch area, if 2+n is less than the
size of the original area. The binary scratch area cannot be
expanded if a device is being compressed onto itself since
that would require writing over existing files.

If no /n is specified, the size of the binary scratch area on
the input device is assumed.

5-lQ

OPT- S

5.1.5

Transfer Source Files

Type S to transfer source files between two file-oriented devices.
Answer IN with the name of the source file to be transferred and,
optionally, a comma, and the output device DTO-DT7, or DKO-DK3.
If no
device is specified the system device is assumed.
Answer OUT with the name to be assigned to the output file and,
optionally, a comma, and the output device, DTO-DT7, or DKO-DK3. If
no device is specified, the system device is assumed.
If an attempt is made to transfer to or
device, the IN or OUT message is repeated.

from

Example:
OPTIN-

S )
TEST, DTl)

OU'T'---

'T'PC:l'J"I
~ .... ....,~ t~~·-

OPT-

mn )

(X or CTRL/C typed here)

5-11

non-file-oriented

OPT- V

5.1.6

Transfer System Program

Type V to move a system program between two file-oriented devices.
Answer IN with the name of the system program to be transferred and,
optionally, a comma, and the input device, DTO-DT7 or DKO-DK3. If no
device is specified, the system device is assumed.
Answer OUT with the name to be assigned to the output file, and,
optionally, a comma, and the output device, DT0-DT7 or DK0-DK3. If no
device is specified, the system device is assumed. If an attempt is
made to transfer to or from a non-file-oriented device the IN or OUT
message is repeated.
Example:
OPT-

V )

INOUTOPT-

SORT, DTl )
SORT, DT4 )
( X or CTRL/C typed here}

5-13

OPT- X

5.1.7

Return to Monitor

Type X in response to OPT- and
return to the Monitor.

the

system

will

terminate

PIP

and

This feature is particularly useful when PIP is included in a string
of Monitor commands in a BATCH program. The X option signals the end
of the PIP program and the next Monitor command in the BATCH program
is executed.

5-15

5.2

ERROR MESSAGES
Message

Explanation

BAD DIRECTORY

Attempt to reference or store a file on a
device with no directory
(or a directory
that has been destroyed).

ILLEGAL DEVICE SWITCH

A switch was specified that was not /R, /C
or /K for input or /P, /L or /T for output.

NO ROOM

Attempt to store a file on a full device.

5-16

CHAPTER 6
BUILD

BUILD is a utility program used to create data files containing fixed
length records. Creating data files requires careful consideration of
their proposed usage.
(Designing BUILD records is discussed in
Appendix D.) BUILD allows input from the keyboard, card reader,
high-speed paper tape reader, or system device.
A BUILD control program must be written before BUILD can be run. This
control program provides the field description and order for input and
output of the data to be used to create the file. Be sure to prepare
a set of instructions on input preparation so data input will be
according to specifications in the control program.
During execution, BUILD checks each data input line to be sure the
data entered conforms exactly to specifications in the control
program. If no errors are found, the input line is written on the new
data file and maybe printed on the line printer.
If an error is
detected, the record is not written to the data file.
BUILD outputs a listing of the control file and data records created
on the line printer, if desired. A summary of the number of records
written, number of records not written and any hash totals kept is
printed on the line printer for each file created.
6.1

BUILD CONTROL PROGRAM

The BUILD control program is created and written
device with editing commands.

mass

storage

This program conveniently falls into three sections, Field Descriptor,
INPUT, and OUTPUT. Briefly, the Field Descriptor section defines all
the fields to be used, the type and size and any initial values to be
assigned.
The INPUT section describes the order in which the data will be
received. The input format specified in this section must be followed
when preparing data for input. The OUTPUT section describes the order
of the output of data. There may be up to seven output formats for
one set of input so several different output files can be created with
one input file.
Figure 6-1 is an example of a simple BUILD control
explained in detail in the following sections.

program

which

Errors in the BUILD control program cause a return to the Monitor.
Spaces
and tabs may be inserted anywhere in the program for
readability. As with any COS program, corranents inserted on a line
must be preceded by a semicolon.

6-1

The control program may be stored as two separate source programs: the
Field Descriptor section as one and the INPUT and OUTPUT sections as
the other •
;AUTOMATIC LINE NUMBER COMMAND,
;START AT 10, INCREMENT OF 10

.!.LN 10,10
.0010
:-0020
:0030
:-0040
~0050

.£0060
.0070
.0080
:0090
~0100

.0110
:0120
:-0130
:0140
:-01so
~0160

.0170
:0100
70190

~0200

• 0210
:0220
:0230
~0240

..!.0250

DEFINE
Fl,D6
F2 ,A30
F3,D7
F4,D6
F5,D6
Tl,DlO
Cl,D2,25
C2,D6,500000
INPUT CDR
PART
Fl
F2
F5
DATE
F4
VENDOR 1
F3
OUTPUT 1 PRTFIL
F3 C
Fl C
F2
F5
F4
END

;PART NUMBER
;PART NAME
; VENDOR NUMBER
;DATE
;WHOLESALE PRICE
; TEMPORARY STORAGE
;CONSTANT FIELD
;CONSTANT FIELD
;INPUT FROM CARD READER
;KEYWORD
; PART NUMBER
;PART NAME
;WHOLESALE PRICE
;KEYWORD
;DATE
;KEYWORD
;VENDOR NUMBER
;OUTPUT TO PRTFIL FILE
;VENDOR NUMBER
; PART NUMBER
;PART NAME
;WHOLESALE PRICE
;DATE

Sample input lines for above control program
:PART 126789 'HEX BOLT' 200
:DATE 30772
'i"VENDOR 14 76116
lPART 176117 WASHER 100
,;.VENDOR 7346177
:PART 176118
:VENDOR 3177615
IPART 12678·8- I #6 MACHINE SCREW' 300
:VENDOR 2061311

Figure 6-1.

Sample BUILD Control Program

6-2

Records created according to control program in Figure 6-1 would
like:
1476116126789HEX BOLT
7346177176117WASHER

6.1.l

22 spaces
24 spaces

look

000200030772
000100030772

Field Descriptor Section

The Field Descriptor section specifies all fields to be referenced in
later sections of the Control program. It gives their type (alpha or
decimal), size (1-510) and initial values, if any.
The format for this section is
DEFINE
fieldx, typesize, initial value

where
DEFINE

is the section heading and
statement in the program.

must

the

first

f ieldx

is an F, T, or C and a number.
The F, T or C
indicates data fields (F) , temporary fields (T) ,
and constant fields {C). These names are for the
convenience of the user and are treated equally by
BUILD. All F fields must be listed first, all T
fields second, and all C fields last. The number
(x) must be a positive, non-zero decimal number
not greater than 511 and may begin at any number.
Numbers may be skipped but within a field group
(F, T or C) must be in ascending order. Temporary
fields are used to accumulate
hash
totals.
Constant fields are used to enter a default value
or range check.
Remember, however, that Fl is assigned to the
first field, F2 to the second, etc. If the field
names assigned are, for example,
Fl
FS
F6
Blank fields will be left in core.
F2 would cause an error.

typesize

References

is A for alphanumeric fields or D for decimal
fields.
Use A to describe fields which contain
other than numbers. Size is an unsigned decimal
number in the range 1-510.

6-3

initial value

Initial values can be assigned to any field.
Alphanumeric values consist of any sequence of
legal COS characters (excluding single quotation
marks) and are enclosed in single quotation marks.
Initial decimal values consist of a sequence of
digits optionally preceded by a + or - sign or
followed by a - sign. These initial values must
be of the same size and type as the field for
which they are specified. The sign and quotation
marks do not count in the size of the initial
value. If an initial value is not specified, the
field is initially set to 0 if decimal or blank if
alpha.
Remember, there is a difference between decimal
and alpha constants consisting of the same digits
and if a constant is to be referenced as alpha and
decimal, it must be defined both ways.
For
example,
Cl,D3,234
C2,A3,'234'

Field description statements must be entered on consecutive lines and
each line is terminated with the CR key. Use the editing commands to
create this file.
Example:
.0010
:0020
-:0030
-:0040

:oo5o

-:-006 0

:0010

:ooao
:0090

6.1.2

DEFINE
Fl,D6
F2 ,A30
F3,D7
F4,D6
F5,D6
Tl,DlO
Cl,D2,25
C2,D6,500000

;STARTS BUILD CONTROL PROGRAM
;PART # FIELD, DECIMAL, 6 DIGITS
;PART NAME FIELD, ALPHANUMERIC, 30 CHARS
;VENDOR # FIELD, DECIMAL, 7 DIGITS
;DATE FIELD, DECIMAL, 6 DIGITS
;WHOLESALE PRICE
; TEMPORARY FIELD
;CONSTANT FIELD CONTAINING VALUE OF 25
;CONSTANT FIELD CONTAINING VALUE OF 500000

INPUT Section

The INPUT section specifies the device {i.e., card reader, keyboard,
and arrangements of the data to be input. The format of th-e
INPUT section is

etc.)

INPUT dev
keyword, output format#
f ieldx

where
INPUT

is the section header and must
statement of the INPUT section.

6-4

the

first

dev

is the 3
device:
KBD CDR RDR SYS -

character

designation

for

the

input

terminal keyboard
card reader
high-speed paper tape reader
a file on the system device

In the case of CDR, RDR and SYS, the input data is
prepared in advance and stored on cards, paper
tape, or the system device.
keyword

( conuna)

is the name assigned to a group of fields which
will make up one input line to the BUILD program.
Keyword is an alphanuneric character string of any
length but only the first six characters are
significant.
Field names cannot be keywords.
Keywords provide the connection between the data
entered at BUILD run time and the data format
specified in the control program.
is optional and specifies comma as the field
delimiter
on
input lines.
This allows the
alphanumeric data input for that keyword
to
contain embedded spaces and tabs.
In most cases,
however, spaces are suggested as delimiters.
For example,
if control program specifies
PART,
then the input line is
PART,126789,HEX BOLT,200
However, if the control program specifies
PART
then the input line is
PART 126788 '#6 MACHINE SCREW' 300

output format#

is a number (1-4 for 8K systems or 1-7 for 12K
systems) which specifies the output format to be
used for this input line. This number corresponds
to an OUTPUT section nu.~ber (described in section
6.1.3).
If an output format number is
not
specified, there is no output for fields under
this keyword.
If an output # is specified, no
output
takes
place
until
input
for
the
corresponding keyword fields is completed.

fieldx

is one of the previously described fields
(F, T,
or C)
and a number. These fields must be listed
in the order in which they will be input.

6-5

To enter the BUILD input line without keywords,
(*) in place of a keyword. For example,
INPUT

specify

asterisk

listed

before

CDR

*l
Fl
F2
All field descriptions beginning with the
those beginning with keywords.

must

NOTE
BUILD input lines without keywords must be entered
in the exact order specified in the INPUT section
of the control program because there are no
keywords to relate an input line to the fields
described in the control program.
Data is inserted in

* fields in order of input.

The INPUT section is terminated by the OUTPUT
section of the control program.

heading

for

the

Example:

·:-0110

DEFINE
Fl,D6
F2,A30
F3,D7
F4,D6
F3,D6
Tl,DlO
Cl,D2,25
C2,D6,500000
INPUT CDR
PART
Fl
F2
FS
DATE
F.4
VENDOR 1

.0180

.0010

~0020

~0030

.0040

:-ooso

:0060
:-0010

:-ooso
70090
:0100
:-0110
:-0120
70130
:0140
:-01so

:0160

;INPUT IS FROM CARD READER
;!ST INPUT LINE HAS KEYWORD PART
;!ST FIELD OF PART LINE IS PART NUMBER
;2ND FIELD OF PART LINE IS PART NAME
;3RD FIELD OF PART LINE IS WHOLESALE PRICE
; 2ND INPUT LINE HAS KEYWORD DATE
; lST FIELD OF DATE LINE . IS DA~.E ..
; 3RD INPUT LINE HAS KEYWORD VENDOR
;AN,D ALL INPUT IS.TO BE OUTPUT
;ACCORDING TO FO.RMAT 1
;IF VENDOR IS MISSING; NO OUTPUT OCCURS
;!ST FIELD IS VENDOR NUMBER

OPTIONS
The INPUT section of the control program also allows assignment of
default fields, skips, or errors for missing fields, flag fields,
range checks, hash total fields, and checkdigits.

6-6

These options may be specified for any field and have
format shown:

the

order

and

SKIP ~

fieldx =dfldx [flag] [RANGE ( [lfldx], [ufldx])] [HASH (hfldx)] [CHECK]
ERROR
Spaces may be used to separate the options for readability.
Any
combination of options can be used but they must appear in the order
shown above.
Use one of the SKIP, default {dfldx) or ERROR options to specify an
alternative for a missing field in a BUILD input line. SKIP causes
the value previously stored in fieldx to be used again, assuming it
was not cleared.
=dfldx assigns the current value of the default field
(dfldx)
to
fieldx. The default field (usually a C field) must have the same type
and size as fieldx. ERROR causes an error message on the terminal at
BUILD run time.
If none of these is specified, SKIP is assumed when
fieldx is missing from the BUILD input line.
A constant field
(C)
cannot be used as fieldx.
Use the flag option to set the flag field to 1 when f ieldx is present
and zero when not present. The flag field can be checked later in the
user's DIBOL program. The flag field must be a decimal field.
Use the RANGE option to specify a high and low limit for the data in
the destination field. The destination field and the fields specified
as the limits must be decimal fields.
The low limit field is
specified first and must be less than or equal to the size of the
destination field.
If the low limit field is omitted, there is no
lower limit.
The high limit field is specified second and must be
larger than or equal to the size of the destination field.
If the
high limit field is omitted, there is no high limit check. The value
of the data entered at run time must be within the range specified.
Use the HASH option to request a hash total on the data in the
destination field.
The destination field and the hash total field
(hfldx) must be decimal. Since the hash total field
(usually a T
field)
is used to store an accumulating sum of the destination field
data, it cannot be smaller than the destination field and should be
larger.
If the hash total exceeds the storage field , the high order
(leftmost) digits are lost. The field used to store the hash total
should not be used for any other purpose within the BUILD conl:rol
program.
Use the CHECK option to request a checkdigit calculation on the data
entered in the destination field.
The destination field must be
decimal. The checkdigit is the rightmost digit and is not stored.
For example, if the field is a D6 and the number 1234561 is eritered,
the rightmost digit (1) is the checkdigit and is not stored.
Using
the same field size, D6, if the number 12345 is entered, 5 becomes the
checkdigit. The checkdigit formula is calculated on 001234 and 5 is
not stored.
If the checkdigit is not valid, an error results. The
formula for calculating the checkdigit is described in Appendix E.

6-7

Example:
0010
0020
0030
0040
0050
0060
0065
0070
0080
0090
0100
0110
0120
0130
0140
0150
0160
0170
0180

DEFINE
Fl,D6
F2,A30
F3,D7
F4,D6
F5,D6
F6,Dl
Tl,DlS
Cl,D2,25
C2,D6,500000
INPUT CDR
PART,
; PART INPUT LINE HAS COMMA AS DELIMITER
Fl ERROR
; ERROR MESSAGE IF PART NO. NOT ENTERED
F2 F6
;F6=0 IF F2 NOT PRESENT
F5 RANGE (Cl,C2) HASH Tl
;WHOLESALE PRICE MUST BE
DATE
;WITHIN RANGE SHOWN IN Cl AND C2. STORE
F4
;HASH TOTAL IN Tl
VENDOR 1
F3 CHECK
;DO CHECKDIGIT ON VENDOR NUMBER
EXPLANATION
RANGE check on amount $.25 to $5,000.00 HASH total
on amount.
If there is not a part number, output an error.
Do a checkdigit on vendor number. Data input for
keyword PART must be separated by commas.

6.1.3

OUTPUT Section

The OUTPUT section specifies the name and sequence of the data to be
stored on a mass storage device. The OUTPUT section is composed of
OUTPUT blocks, each of the following form
OUTPUT output format# label/default unit
fieldx

where
OUTPUT

is the heading of the last section of the control
program and must appear even if no output is
desired. Be sure that OUTPUT is not used as a
keyword in the control program.

output format#

is the number (1-4 for 8K systems, 1-7 for 12K
systems) which links a keyword in the INPUT
section to a particular OUTPUT format.
This
number must be in ascending order (beginning with
l} and must appear after OUTPUT.

6.-S

label

is the name to be assigned to the newly created
data file~ The name can be anv unique sequence of
characters (only the first six-are significant).

default unit

is an optional entry specifying the logical unit
(1-15) where the output file is located.
(rf the
number specified exceeds 15, it is taken modulo
16.)
If the file is not found on the default unit
when the control program is executed, a MOUNT
message is displayed.

f ieldx

The F, T or C fields in the order

which

they

are to be output.
The OUTPUT section ends with the specification of another
section (up to 4 for 8K, 7 for 12K) or an END statement.

OUTPUT

If another OUTPUT section is specified, its format number must be the
next number in sequence. Different OUTPUT format sections can contain
the same fields so that several data files can be created from one
input. All fields between OUTPUT and END or OUTPUT and another OUTPUT
form one record; record length must not exceed 510 characters.
OPTIONS
The OUTPUT section of the control program has two
which are specified as follows:

options,

OUTPUT 1 datal
FlO C
TSO +

T75

+cs

Use C to clear a field after output of the record.
Decimal fields are
cleared to zeros, alpha fields to blanks. During input the contents
of fields not cleared with C options can be repeated in following
records by skipping fields in subsequent input lines.
Use + as a counter or sequence check. The field to be incremented is
usually a temporary field
(T)
and must be decimal. This field is
output as part of the record being created. + increments the field by
one; to increment by any other number specify an increment field
(+fieldxx). The increment field (usually a C field) must be decimal
and must not be larger than the destination field.
If the increment
sum exceeds the destination field, the high order digit is lost.
The
c and + o'ptions cannot be used on the same field.

6~-9

Example:
0010
0020
0030
0040
0050
0060
0070
0080
0090
0100
0110
0120
0130
0140
0150
0160
0170
0180
0190
0200
0210
0220
0230
0240
0250

DEFINE
Fl,D6
F2 ,A30
F3,D7
F4,D6
F5,D6
Tl,DlO
Cl,D2,25
C2,D6,500000
INPUT CDR
PART,
Fl ERROR
F2
F5 RANGE (Cl,C2) HASH Tl
DATE
F4
VENDOR 1
F3 CHECK
OUTPUT 1 PRTFIL
;OUTPUT FORMAT l; FILE NAME, PRTFIL
;VENDOR NUMBER
F3 C
;PART NUMBER
Fl C
F2 C
;PART NAME
;WHOLESALE PRICE
F5 C
F4
;DATE
END
EXPLANATION
Clears fields
written.

6.1.4

with

"C"

each

time

record

Storing the BUILD Control Program

When the BUILD control file is complete, it should be stored on a mass
storage device for future use.
Use the command
.WRITE BLDSPC
for example, to store the control program.

6-10

6.2

OPERATING PROCEDURES

To load the BUILD program, type
• RUN BUILD [, filel, ••• , f ile7] [/xx]
in reply to the dot displayed by the Monitor
where
filel, ••• ,file7

are the BUILD
control file

control programs.
If the BUILD
specified the INPUT device as SYS,

then the rest of the files are names of data files
on the system device to be input.
If no files are
specified, the file in the edit buffer is used to
create a master file.
Multiple files are concatenated and passed to
program as one large file.
are optional switches which affect the
printed by BUILD. The first switch is:

/xx

list the BUILD control program, or

do not list the control program.

the

listing

The second switch is:
/L

list the data records created

do not list the data records

If a second switch is not specified, a data
records list is not produced.
(/N is assumed.) If
the no list condition is
specified,
errors
encountered in the control program or data records
are listed on the line printer as is the summary
of hash totals, records created, etc.
Examples:
• RU BUILD,BLDSPC/NL
lists the data records and summary
program •

totals

but

not

• RU BUILD,BLDSPC/L
lists the control program and summary of totals, while
• RU BUILD,BLDSPC/LL
outputs both listings and the summary.

6-11

the

control

When BUILD is loaded, the message
MOUNT xxxxxx #01 FOR OUTPUT:
is displayed. xxxxxx is the output file name specified in the control
program and #01 is the sequence number.
Continuing the example, the message is
MOUNT PRTFIL #01 FOR OUTPUT:
Answer with the number (1-15) of the logical device where the output
file is to be stored. When BUILD is complete, the Monitor DIRECTORY
command can be used to obtain a label of the data file.
The next
steps depend on the device (KBD, CDR, RDR, SYS) specified in the INPUT
statement of the control program. If the data output exceeds the
storage capacity of the logical unit, the following message is
displayed if default messages were not specified:
MOUNT xxxxxx #02 FOR OUTPUT:
where xxxxxx is the label of the output file. Mount the new output
device if necessary.
Type the number of the logical unit where the
rest of the output is to be stored.
BUILD can be aborted at any time by typing CTRL/C, to return to the
Monitor.
All output is lost. The BUILD program must not be run if
the line printer is inoperative.

6.2.1

KBD - Data Input from Keyboard

If the INPUT statement
command with

specifies

KBD,

BUILD

responds

the

RUN

DO YOU PREFER PROMPTING?
Answer YES and a CR key to have the input line keywords printed
automatically by BUILD. To use prompting, the data input must consist
(if present) appear in
of successive input lines whose keywords
exactly the order in which they are presented in the INPUT section of
the control program with no omissions. Monitor displays

after each keyword to indicate it is ready for input.
Answer NO and a CR key to allow each input line keyword to be typed.
BUILD displays a colon (:) when it is ready to accept input. Input
lines can then be typed in any order and BUILD will match the input
line keyword to the keyword in the INPUT section of the control
program.
If the reply is YES or NO, the dialogue is over and
begin. BUILD assumes errors will be edited on line.

6-12

data

input

can

When all the input lines have been entered, type CTRL/Z to indicate
the end of data input.
BUILD closes the output files and prints
number of records written, records not written and any hash totals
requested and returns control to the Monitor.

6.2.2

SYS, CDR or RDR - Data Input from System Device,
Card, or Papertape

If the INPUT statement specifies SYS, CDR or RDR,
The question
DO

YOU WISH TO EDIT ON-LINE?

is displayed. Answer NO and the CR key and future errors are printed
on the line printer and all fields in the bad input line will become
undefined. Records with undefined fields are not written to the
output file.
They remain undefined unless cleared in the output
section or redefined with subsequent input.
Answer YES and the CR key to have an error message displayed on
terminal whenever an error is encountered.
BUILD displays
incorrect input line, the error message and the message

the
the

RETYPE:
Type the correct input line. For details on data input refer to
Section 6.3.
If just the CR key is typed, the field in error becomes
undefined.
When BUILD reaches the end of the paper tape in the reader, it asks

Answer NO and the CR key if there is no more input.
Otherwise, put
additional paper tape in the reader and type Y and the CR key to
continue reading data.
When input is complete, BUILD closes the output data files and prints
the number of records written and records not written and any hash
totals requested and returns control to the Monitor.
When BUILD reaches the end of cards in the reader, the terminal alarm
sountls.
If ttrere are more cards put them in card reader and type any
key to resume operation. Otherwise type CTRL/Z to terminate input.

6.3

BUILD INPUT LINE

The data input line is free form, i.e., specific data fields need not
be in prespecified column positions.
However, the fields must be
input in the same order within a keyword line as specified in the
INPUT section.
Data input lines entered at the keyboard may or may
not contain keywords depending on the answer to the prompting
question.
Data input lines prepared on paper tape or cards must
contain keywords, unless an asterisk was specified in the control
program.

6-13

The following conventions must be observed when entering input lines.
1.

The first field in an input line may be preceded by spaces or
tabs if desired.

Each input line must be terminated with the CR key.

The fields, if more than one is entered per· line, must be
separated from each other by spaces, tabs, or commas. For
example, an input line described in the control program as
DEFINE
Fl,D6
F2,A30
F5,D6
INPUT KBD
PART,
Fl
F2
FS
can be entered as
PART,126789, HEX BOLT, 500
or if described as:

PART
Fl
F2
FS
can be entered as

A field can be omitted by actually omitting it if there are
no flel.ds to the right or by typing two consecutive commas if
there are fields to the right. For example
PART,126789,HEX BOLT
PART,, HEX BOLT,100
Omitting a field may cause an error depending on the options
(SKIP, default field or ERROR) specified for the field in the
control program.

Decimal data must be entered in one of the following forms:
digits
+digits
-digits
digits-

6-14

The plus sign is optional and minus can be placed at the
beginning or end of the data.
Decimal data cannot have
embedded blanks, i.e., 12 34 is illegal.
The number of
digits must be less than or equal to the field size specified
in the control program in the corresponding field.
If the number entered is smaller than the field size,
right justified and zeros are inserted on the left.
number is too long, an error results.
6.

Alphanumeric data must be entered in
forms:
'data'

one

the

it is
If the

following

data

In the first form ('data') any legal COS character (including
spaces and tabs) is permitted except a single quotation mark
(apostrophe). BUILD treats tabs inside quotes as single
spaces, but echoes them as backslashes on the terminals.
In the second form any legal COS character except spaces,
tabs, and commas is permitted. The first character, however,
must not be a single quotation mark.
If the keyword PART did
not have a comma, then input would be
PARTL-1175L-1WASHER....,lOO
The number of alphanumeric characters must be less than or
equal to the size of the corresponding field specified in the
control program.
If the data is smaller than the field,
it
is left justified and padded with blanks.
If the data is too
large, an error results.
7.

If comma was specified as the field delimiter in the INPUT
section of the control program, then alphanumeric data not
enclosed in single quotation marks may have embedded spaces
and tabs but no commas and must be separated from the next
field by a comma.
For example, if the control program
specified:
DEFINE
Fl,A30
F2,A30
F3 ,A30
INPUT KBD
NAME ,1
Fl
F2
F3
the input line would be typed as
NAME, JOHANSEN' s MARKET, 2 MAIN

s·r. , I MAYNARD, MASS. '

Extra text at the end of a data record is illegal except
comments which must be preceded by a semicolon.

6-15

for

The arrangement of * and keyword records is immaterial at run
time.
BUILD inserts records in the asterisk fields by the
order of input and in keyword records by matching keywords.
For example,
Descriptor

Input Lines

*
eyword2 •••

* •.•

etc.

There is a current asterisk record pointer in BUILD
keeps track of the next asterisk record to be filled.
inputting the last * record, BUILD goes back to the
one.

6.4

which
After
first

ERROR MESSAGES

BUILD error messages are output to the line printer as the errors
encountered.

are

BUILD checks each section of the control program and
for errors.

data

the

input

If editing is being done on line, the line in error, error message,
and RETYPE messa<Je are displayed on the terminal. The keyword is
displayed if prompting is in effect. Otherwise, the keyword must be
retyped.
If no input editing is being done, data input errors are
logged on the line printer in the form
input line

****

error message
Failure to correct a field in error causes the field to become
undefined. The record containing an undefined field is not written on
a storage device.
Error messages followed by an explanation point (!) indicate fatal
errors.
The error messages for the Field Description section of the
control program are:

Ei-16

Message

Meaning

BAD DIGIT IN DECIMAL INITIAL VALUE
Alpha character
value.

decimal

initial

Maximum = 960

DATA TABLE OVERFLOW!

Too
much
characters.

data.

DESCRIPTOR TABLE OVERFLOW!

Too many F, T
Maximum = 160.

and

EXTRA CHARS AT STMNT END

Characters not relating
appear with statement.

FIELD NUMBER MISSING OR 0

Field number or default unit number is
missing, is 0 or greater than or equal
to 512.

INITIAL ALPHA VALUE DOESN'T BEGIN WITH QUOTE
Beginning quotation
initial alpha value.

fields

defined=

mark

statement

missing

for

INITIAL VALUE TOO BIG

The initial value specified
than the field size.

larger

INITIAL VALUE TOO SMALL

The initial value specified
than the field size.

smaller

specified

where

MISSING CLOSE QUOTE ON ALPHA INITIAL VALUE
Quotation mark
necessary.

not

MISSING INITIAL VALUE

Comma was inserted after type and size
but initial value was not specified.

NO COMMA AFTER FIELD NAME

No comma or a character other than comma
was specified after the field name.

NOT A OR D

A character other than A or D occurred
in a data specification statement where
A or D was expected.

NOT BUILD FILE!

File

did

not

start

with

DEFINE

statement.
NOTHING AFTER FIELD NAME

Field type and size are not
after field name and conuna.

specified

NUMBER REPEATED OR OUT OF ORDER
A field sequence number is used more
than once or is out of ascending order
sequence.

6-17

Error messages for the INPUT/OUTPUT sections of
are:
Message

the

control

program

Meaning

BAD END STATEMENT

Missing or misspelled END statement.

BAD DEFAULT FIELD

Default field does not agree
or type of object field.

BAD DEPICTOR

OUTPUT format in error.

BAD DEVICE

Not KBD, SYS, RDR or CDR.

BAD FIELD NUMBER

Field number specified is
2047.

BAD INPUT STATEMENT

Statement missing or misspelled.

BAD OPTION

Something other than C or + is specified
in OUTPUT section.

BAD OUTPUT FORMAT i

Format number is wrong
.(1 or 7) •

BAD RANGE CONSTRUCTION

In RANGE portion of statement an open or
close parenthesis or comma is missing.

C-FIELD NOT PERMITTED

c-type fields cannot be specified in the
INPUT section.

with

size

greater

than

out-of-range

DESTINATION FIELD NOT DECIMAL
The destination field specified
defined as decimal.

not

hash

bigger

than

FIELD NOT DECIMAL

Field not defined as decimal.

HASH FIELD MISSING

No field specified for storage
total.

INCREMENT TOO BIG

Length of increment field
length of output field.

MISSING FIELD NAME

Field name is missing.

OUT OF ROOM!

Too many descriptors in INPUT section.
Maximum
keywords, approximately 121.
Approximate maximum fields
for
all
keywords = 192.
Too many fields in
OUTPUT section. Approximate maximum =
128.

OUTPUT RECORD TOO BIG

Record exceeds 510 characters.

6-18

Message

Meaning

UNDEFINED FIELD

Field referenced was not defined in
Field
Descriptor section of control
program or a decimal field was being
equated to an alpha field, or an alpha
field was specified for a flag.

UNDEFINED FORMAT

Format number referenced but not defined
(error given on END statement).

UNEXPECTED END OF FILE

Missing END statement.

Error messages for data input are:
Message

Meaning

BAD ALPHA FIELD

Illegal entry in alpha field.

BAD CHECKDIGIT

Checkdigit calculated by BUILD does
match the one entered.

BAD DECIMAL FIELD

Illegal entry in decimal field.

BAD DELIM

Fields are separated
delimiter, i.e., by a

by an
or!.

BAD KEYWORD

Keyword entered not
control program.

described

EXTRA CHARS AT END OF DATA

More data was entered than was defined
in the control program, i.e., 4 fields
entered when only 3 were defined.

FIELD OUT OF RANGE

Data entered is not within the
specified in the control program.

range

NECESSARY FIELD MISSING

Field specified as necessary in
program was not entered.

control

TOO BIG

Data entered is greater than the size of
the field as defined in the control
program.

6-19

not

illegal
in

the

CHAPTER 7
SORT

SORT arranges COS data files in order according to a key specified in
the SORT control program. The key consists of certain fields or parts
of fields which appear in each record of the file being sorted.
The
data records can be sorted in ascending or descending order based on
the key.
Before SORT can be run, a SORT control program must be written.
The
control program defines the fields of the records to be sorted,
specifies labels for input and output files, and the key to be used in
the sort.
SORT handles one file per run and there must be a separate SORT
control program for each file sorted. Files to be sorted must contain
fixed length records.
If a multi-reel file is to be sorted, the SORT program sorts each reel
separately.
A separate merge pass of the same SORT program must then
be done to combine the file. To sort a multi-reel DECtape file on
disk without going through a merge phase, PIP the data file onto disk
before sorting.
(See PIP option D.)

7.1

SORT CONTROL PROGRAM

The SORT control program is created and written on a mass storage
device with editing commands.
It consists of a file Descriptor
section and an INPUT/OUTPUT section.

7.1.1

File Descriptor Section

The File Descriptor section has the form:
DEFINE
Fxyt.ypesize

where:
DEFINE

is the section heading and
statement in the program.

7-1

should

the

first

is the first field in the record to be sorted.
All fields in the record to be sorted must be
defined in the File Descriptor section in the
order they appear in the record. The number (x)
must be a positive, non-zero decimal number not
greater than · 511 and should begin at 1 with no
numbers skipped.

typesize

is an A for alphanumeric fields or a D for decimal
fields and an unsigned decimal number in the range
1 - 510.

File Descriptor statements must be entered on consecutive lines and
each line is terminated with the CR key. Use the editing commands to
create the control program.
NOTE
The BUILD Field Descriptor written previously to
create the data file can be used as the SORT File
Descriptor if the input and output fields are
exactly the same and if it was stored as a
separate program. T and C fields and initial
values are ignored by SORT.
Example:
0010
0020
0030
0040
0050
0060

DEFINE
Fl,D6
F2,A30
F3,D7
F4,D6
F5,D6

This example is based on the
program.
(See Chapter 6.)

7.1.2

;PART NUMBER
;PART NAME
;VENDOR NmmER
;DATE
;WHOLESALE PRICE
file

created

with

the

BUILD

control

INPUT/OUTPUT Section

This section specifies the input and output file labels and the number
of work units available for the SORT.
These work units become part of the
SORT run time.

7-2

MOUNT

message

displayed

during

Format of the INPUT/OUTPUT section of the control program is:
INPUT filnam/d,filnam/d
SORT n/wl,w2 ,w3
KEY fieldx(n,m)-,fieldx(n,m) •••
OUTPUT f ilnam/d
END
where
INPUT filnam/d,filnam/d

is the name of the file containing the
records to be sorted.
If no name is
specified, SORTIN is assumed.
/d is
optional and is the unit (1-15) to check
for the named file.
If the control
program is used for a separate merge
operation, the additional filnams are
the files to be merged.

SORT n/wl,w2 ,w3

is the number (n) of logical units to be
used as scratch areas during the sort.
n is any number from 3 to 7. There must
be at least three units to execute a
sort.
If the SORT statement is not
present 4 units are assumed.
It is
recommended that the size of the scratch
units be as large as one "reel" of the
input file.
/wl,w2,w3 are
optional
default
work
units to be used if
available. Using default units bypasses
the mount work unit message.

KEY fieldx(n,m)-, •••

f ieldx is the field name and number of
the field to be used as the SORT key.
(n,m) are optional and specify the nth
through mth characters of the field.
If
no characters are specified, the entire
field is used as a SORT key. - requests
a SORT on that field in descending
order.
If - is not specified, ascending
order is assumed (+ may also be used to
indicate ascending order). Up to eight
fields or parts of fields
can
be
specified for the sor-t key. There is no
restriction on the size of the sort
field as long as the total size of the
fields which comprise the key is not
larger than 510 characters. The sort is
done left to right on the key using the
COS codes shown in Appendix A. The
leftmost key is most significant, and
the leftmost character in each key field
is
most
significant
for
sorting
purposes. n and m are unsigned integers
with n less than or equal to m. Each rn
must be less than or equal to the length
of the record.

1-3

OUTPUT f ilnam/d

is the data file name to be given to the
sorted records.
If this statement is
missing SORT assigns the name SRTOUT to
the output.
For multi-reel files the
names $TMPOO - $TMPnn are used.
/d is
the default unit to check for the output
file.
If the file is not there, the
mount message is displayed.
NOTE
For a multi-reel input file,
SORT sorts the first reel onto
the output file and then asks
for the second reel of input.
For example, when sorting a
3-reel input file, the output
files (which are to be input
to MERGE)
are called $TMPOO,
$TMP01, and $TMP02.

END

Terminates control program.

Example:
0010
0020
0030
0040
0050
0060
0070
0080
0090

DEFINE
Fl,D6
F2 ,A30
F3,D7
F4,D6
FS,D6
INPUT PRTFIL
SORT 4
KEY Fl-,F4

0100
0110

OUTPUT PRTFIL
END

7. 2_

;DATA FILE NAME
;WORK UNITS
;SORT KEYS = PART
;NUMBER IN DESCENDING ORDER
;DATE IN ASCENDING
;ORDER
;SORTED DATA FILE NAME

SORT OPERATING PROCEDURES

SORT is called via the Monitor command:
.RU SORT[, filel, ••• ,file7]
where
filel, •• ,file7 are the SORT control program which can be stored
as more than one file.
If none are specified, the
file in the edit scratch area is used.

7-4

Mount the input file and n-1 (where n is the number of logical units
to be used for SORT) of the scratch units. The Monitor outputs the
message:
MOUNT

filnam iOl FOR INPUT:

where f ilnam is the input file name specified in the SORT control
program.
Type the number of the logical unit (1-15) where the data
file is stored. Monitor outputs the messages
MOUNT $WORK2 #01 FOR OUTPUT:

MOUNT $WORKn #01 FOR OUTPUT:
Type the number (1-15) of the logical unit where all but the last of
the SORT scratch units are located. The Monitor checks the labels of
the files currently on these units.
If a label exists, Monitor asks:
REPLACE xxxxxx #nn?
Type YES and the CR key to use the unit as a scratch area and destroy
what is currently stored there. Type NO and the CR key to save the
present label and specify another unit for a work area.
When the first sort phase is complete, the input is no longer
Monitor displays the message:

needed.

MOUNT $WORK1 #01 FOR OUTPUT:
In a tape SORT the input tape can be dismounted and the last work unit
mounted or the input file can become the last scratch area.
(Monitor
will ask REPLACE? if this is done; type YES and the CR key).
Type the number (1-15) and the CR key of the logical
last work area is located.

unit

where

the

When the main part of the SORT is complete, Monitor displays:
UNIT xx IS FREE
MOUNT filnam #01 FOR OUTPUT:
xx is one of the units previously assigned as a work unit.
SORT no
longer needs the unit and it May be used for the output unit if
desired.

Mount the output file.. Type the logical unit number (1-15) where the
output is to be stored.
SORT reads the sorted records from the work
units and stores them on the specified output unit.
If a default output unit is specified in the OUTPUT line of
control, the MOUNT ••• FOR OUTPUT message may not appear.
NOTE
Sometimes the SORT program does not request that
the last work unit be mounted - it types UNIT xx

7-5

the

SORT

IS FREE where xx is the unit where the input file
is mounted, and asks that the output file be
mounted. Do not be alarmed
SORT has decided
that
the
file
is
sufficiently
small and
well-ordered that the main part of the sort is
unnecessary.

7.3

RUNNING SORT AS PART OF AN UPDATE PROCEDURE

If the SORT control program has a line the first six non-blank
characters of which are UPDATE, then the contents of the INPUT and
OUTPUT parameter lines are ignored and the SORT program uses the label
$UPDOO for both its input and its output file.
{See Chapter 8 for
UPDATE.)

7.4

MERGE OPERATING PROCEDURE

To run SORT as a merge only operation, type:
.RU SORT[, filel,file2, ••• ,file7]/x
where
filel, ••• ,file7

are the control program (possibly stored in two or
more segments).

is one of the following switches:
/A

specifies that names of files to be merged are to
be entered from the terminal in answer to the
INPUT FILE LABELS: message. The output data file
and default unit name are specified in the OUTPUT
clause of the control program.

specifies that names of files to be merged are
listed in the
INPUT line of the SORT control
program. This option bypasses the INPUT FILE
LABELS: message.

specifies that the name of the files to be merged
is listed in the INPUT line of the SORT control
program. Checks for n files of the same name on
n(l-6) default units as specified.
If the number
of units specified is more than the number of
units shown on the SORT control INPUT line, a
MOUNT message is displayed for those files not on
the INPUT line.

can be used with any of the above
lists
the SORT control program
printer.

7-6

switches and
on the line

wnen called with the /A option, SORT requests the names
files to be merged with the message:

the

data

INPUT FILE LABELS:
Enter the 1-6 character data file names separated by commas up to a
maximum of six and default units if desired. At least two names must
be entered or the error message NO INPUT is displayed.
For example:
PRTFIL/10,FILEl/3
If default units are not specified, Monitor outputs the message:
MOUNT filnam #01 FOR INPUT:
(where filnam is the data file name) for each data file specified.
For example:
MOUNT PRTFIL #01 FOR INPUT:
Type the logical unit number (1-15) where the data file is located.
Example:
Sample Control Program called PAYKEY

0100
0110
0120

0130
0140
0150
0160
0170
0180

DEFINE
Fl ,A6
F2 ,D5
F3,Dll
INPUT PAYROL/4, PAYl/2
SORT 3/1,2,3
KEY F2
OUTPUT PAYROL/6
END

To SORT the data file described in the control program PAYKEY,
.!. RUN

SORT , PAYKEY

The input data file. is PAYROL located on logical unit 4.
There are
three sort work units, logical units 1, 2 and 3. The output data file
is PAYROL and the sorted file is to be put back on unit 6.
The /A option in SORT is used as follows:
RUN SORT, PAYKEY /A)
INPUT FILE LABELS:
PAYROL/4, PAYl/2, PAYO)
MOUNT PAYO #01 FOR INPUT:
10)

..!..

7-7

The input data file names to be merged are:
PAYROL on logical unit 4
PAYl on logical unit 2
PAYO on logical unit 10
The output data file, PAYROL, is put on logical unit 6.
The /M option in SORT is used as follows:
.RUN SORT, PAYKEY/M)
The input data file names to be merged are found in the INPUT line
the control program:

PAYROL on logical unit 4
PAYl on logical unit 2
The output data file, PAYROL, is put on logical unit 6.
The /n option in SORT is used as follows:
RUN SORT, PAYKEY/2 )

..!..

The input data files to be merged are:
PAYROL on logical unit 4
PAYROL on logical unit 2
7.5

RULES FOR USING DEFAULT UNITS

If you are using default units in your control programs for SORT,
BUILD, etc., the data files should be mounted on those units. If the
Monitor cannot find the input file on the default unit the message
? MOUNT xxxxxx #xx FOR INPUT:

is displayed. In the above MOUNT message the ? indicates that
input data file "xxxxxx" was not on the default unit specified.

the

MOUNT PAYROL #01 FOR OUTPUT:
6)
If output of a data file using a default unit is attempted and another
file already exists, the REPLACE message is displayed:
REPLACE LIST #02 ?
Y)
If output of a data file onto a logical unit
used for input or output is attempted,
displayed.
IN USE
MOUNT PAYROL #01 FOR OUTPUT:
10)

that is already being
the message IN USE is

7.6

ERROR MESSAGES

Message

Explanation

File descriptor errors:
BAD DIGIT IN DECIMAL INITIAL VALUE
Alpha character
value.

decimal

initial

BAD WORK UNIT COUNT

Number of work units not in range 3-7.

EXTRA CHARS AT STMNT END

Characters not relating
appear on statement line.

FIELD NUMBER MISSING OR 0

Field number is missing or
greater than or equal to 512.

ILLEGAL SORT KEY

Bad syntax on KEY statement, KEY
complex, or KEY statement missing.

ILLEGAL UNIT

Default unit is 0 or greater than 15.

INITIAL ALPHA VALUE DOESN'T BEGIN WITH QUOTE
Beginning quotation
initial alpha value.

mark

statement
is

missing

or
too

for

INITIAL VALUE TOO BIG

The initial value specified
than the field size.

larger

INITIAL VALUE TOO SMALL

The initial value specified
than the field size.

smaller

specified

where

MISSING CLOSE QUOTE ON ALPHA INITIAL VALUE
Quotation mark
necessary.

not

MISSING INITIAL VALUE

Comma was inserted after type and size
but initial value was not specified.

NO COMMA AFTER FIELD NAME

No comma or a character other than comma
was specified after the field name.

NOT A OR D

A character other than A or D occurred
in a data specification statement where
A or D was expected.

NOTHING AFTER FIELD N~~~

Field type and size are not
after field name and comma.

specified

NUMBER REPEATED OR OUT OF ORDER
A field sequence number is used more
than once or is out of ascending order
sequence.
TOO MANY FILES

Merge only,
specified.

7-9

than

input

files

Explanation

Message
UNRECOGNIZABLE LINE

Parameter line did not start with a qood
keyword.

SORT Errors:
File contains
length.

NO INPUT

Input file is null or not enough
files specified for a merge.

OUTPUT ERROR

Indicates a system malfunction.

UNIT xx IS FREE

NOT AN ERROR - merely an informative
message to aid the operator. xx is a
COS logical unit number.

7-10

records

variable

BAD RECORD SIZE

input

CHAPTER 8
UPDATE

COS UPDATE is a file maintenance program
insertions and deletions to data files.

which

makes

changes,

UPDATE must also have a control program before i t can be executed.
UPDATE operates in two passes, -passl reads the control program and
UPDATE commands and checks for errors.
It then outputs a modified
version of these commands to a scratch file for the user to sort.
Pass2 uses the UPDATE commands read from the scratch file to modify
the input file, and creates a new output file.
The UPDATE commands are listed on the line printer as they are
executed.
A summary of the number of records read, changed, inserted
and deleted is printed when the UPDATE is complete.

8.1

UPDATE CONTROL PROGRAM

The control program consists of a file dscriptor section followed
UPDATE, INPUT, SORT, KEY, OUTPUT and END statements in that order.

8.1.1

File Descriptor Section

This section defines the fields as they appear in the file to be
updated.
The restrictions are the same as for the SORT program. T
and C field names are accepted but are ignored by UPDATE.
Initial
values are retained unless specifically changed by an UPDATE command
(see section 8.2). Be sure not to skip any numbers when assigning
field names.
The file descriptor section must start with a DEFINE
statement.
Example:

O<HO
0020
0030
0040
0050
0060

DEFINE
Fl,06
F2,A30
F3,D7
F4 ,D6
FS,D6

;PART NUMBER
;PART NAME
;VENDOR NUMBER
;DATE

;WHOLESALE PRICE

This file descriptor section is based on the file created
in Chapter 6.

8-1

with

BUILD

8.1.2

UPDATE Statement

The form of this statement is:
UPDATE filnam[/unit]
where
filnam

is the 1-6 character name of the data file to be
updated.
This name may not contain spaces, tabs,
or commas.

/unit

is optional and specifies the logical unit {1-15)
which contains the file.
If the number specified
exceeds 15, it is taken modulo 16.
If the unit is
omitted or if the file is not on the unit
specified, Monitor outputs
the
usual
MOUNT
message.

Example:
DEFINE
Fl,D6
F2 ,A30
F3,D7
F4 ,D6
F5,D6
UPDATE PRTFIL/3

0010
0020
0030
0040
0050
0060
0070

8.1.3

;DATA FILE TO BE UPDATED IS
;PRTFIL LOCATED ON LOGICAL
;UNIT 3.

INPUT Statement

The form of this statement is:
INPUT dev
where
dev

is the physical device {KBD, RDR, CDR, SYS) to
used to input the UPDATE commands
KBD RDR CDR SYS -

terminal keyboard
paper tape reader
card reader
system device file

8-2

Exa."nple:

0010
0020
0030
0040
0050
0060
0070
0080

8.1.4

DEFINE
Fl,D6
F2 ,A30
F3,D7
F4,D6
F5,D6
UPDATE PRTFIL/3
INPUT SYS

:UPDATE COMMANDS
;ARE STORED
;IN A FILE ON THE
;SYSTEM DEVICE.

SORT Statement

The form of this statement is:
SORT n
is the number (3-7) of logical scratch units which
are to be used if the updated records must be
SORTed. This is an optional statement.
If the
UPDATE commands are sorted, no SORT statement is
needed. There must be at least three units to
execute
a
SORT.
If no scratch units are
specified, 4 is assumed. If no SORT statement is
specified, UPDATE executes a combined passl and
pass2, reading one UPDATE command at a time and
performing the update on the input file directly.
No scratch unit is needed in this case.

where n

Example:
0010
0020
0030
0040
0050
0060
0070
0080
0090

DEFINE
Fl,D6
F2 ,A30
F3,D7
F4,D6
FS,D6
UPDATE PRTFIL/3
INPUT SYS
SORT 4

; THERE ARE FOUR SCRATCH

;UNITS FOR THE SORT

8.1.5

KEY Statement

The form of this statement is:
KEY fieldx(n,m)-[, ••• ,fieldx(n,m)-]
where

8-3

fieldx(n,m)-, •••

is the SORT key used to sort the data file.
(See
SORT control program.) There can be up to eight
fields or parts of fields in the SORT key.

Example:
DEFINE
Fl,D6
F2 ,A30
F3,D7
F4,D6
F5,D6
UPDATE PRTFIL/3
INPUT SYS
SORT 4
KEY Fl-,F4

0010
0020
0030
0040
0050
0060
0070
0080
0090
0100

8.1.6

; THE KEY FIELDS
;ARE PART NUMBER IN
;DESCENDING ORDER
;AND DATE
;IN ASCENDING ORDER.

OUTPUT Statement

The form of the statement is:
OUTPUT filnam[/unit]
where
filnam

is the 1-6 character name to be assigned to the
updated file.
If the statement is omitted the
output file has the same name as the input file.

/unit

is the logical unit (1-15) where the output will
be stored.
If no unit is specified, a MOUNT
message is displayed.

Example:
0010
0020
0030
0040
0050
0060
0070
0080
0090
0100
0110

DEFINE
Fl,D6
F2 ,AJO

F3,D7
F4,D6
FS,D6
UPDATE PRTFIL/3
INPUT SYS
SORT 4
KEY Fl-,F4
OUTPUT PRTFIL/3

;CALL UPDATED
;FILE PRTFIL AND
;STORE ON UNIT 3

8.1.7

END Statement

The form of the statement is:
END
END is the terminating statement of the
must be present.

UPDATE

control

program

and

Example:
0010
0020
0030
0040
0050
0060
0070
0080
0090
0100
0110
0120

8.2

DEFINE
Fl,D6
F2,A30
F3,D7
F4,D6
F5,D6
UPDATE PRTFIL/3
INPUT SYS
SORT 4
KEY Fl-,F4

OUTPUT PRrFILi3
END

;TERMINATING STATEMENT OF
; CONTROL PROGRAM

UPDATE COMMANDS

The UPDATE conunands can be input from tne keyboard at UPDATE run time
or can be prepared previously on paper tape, cards or the system
device.
UPDATE commands have the general form:

{~).....,key [Fn=value], •••
where
I,D,C

denote Insert, Delete, and Change

key

represents the data in the fields ttsed to sort the
record.
The character string typed must be
exactly the same length as specified for the SORT
fields. The string is continuous; no spaces, etc.
can separate fields if more than one is in the
sort key.
If the data file is sorted on an alpha
key (Al2) all trailing spaces must be typed on an
UPDATE conunand line, for example,
CL-IJONES JOHNL-IL...1L...1F6= •••

is a field name and sequence number as specified
in
the
UPDATE File Descriptor.
Fields are
processed from left to right and Field numbers may
appear in any order.

8-5

=value

is the data to be inserted in the field named
(Fn).
This data must agree in type with the type
specified for Fn and must be less than or equal in
size.
If the value specified is smaller than the
field size, the alpha field data is left justified
and padded with blanks; decimal field data is
right-justified and padded with zeros.

On an insert, missing

fields are set to their
initial value if one was specified, otherwise the
field is set to blanks
(alpha)
or
zeroes
(decimal).

Decimal values have the format

[~]digits H
Alpha values have the format
'characters'
No single quotation marks may appear within the
alpha value.
If no character is a space, tab or
comma then the surrounding quotation marks need
not be present.
A command which is too long for one UPDATE command line can be
continued on the next line. Type a comma at the end of each line to
be continued. There can be any number of continue lines.
Example:
IL-19988773771761 Fl=001122,)
F2='SCREW,#ll',F3=100,F4=7771777,F5=32072
An error in an UPDATE command line causes the line to be ignored and
an error message is output to the line printer. Execution continues.

8. 3

UPDATE EXAMPLE

A-file called ACCT has a mistake in the zip code of an employee
employee number is 63411.
0010
0020
0030
0040
0050
0060
0070
0080
0090
0100
0110

DEFINE
Fl ,A40
F2,DS
F3,A30
F4,A5
F5,D8
UPDATE ACCT
INPUT KBD
KEY F2
OUTPUT ACCT
END

whose
The following control file is prepared:

;NAME
;EMPLOYEE NUMBER
;STREET ADDRESS
;ZIP CODE
;SALARY

8-6

·1·n1s file describes tne fields or "Cne ri.Le Acc·i·, specifies "Cnat 1"C is
sorted on field2
(employee number) and says that input commands are
from the keyboard. No SORT statement is included because only one
update command is needed and so no sorting is necessary. If more than
one change is to be made in the file and if the changes are typed with
their keys not in order then a SORT statement would be necessary.
The user then types:
• RUN UPDATE/lL
to run passl and get a listing of the control file.
If no SORT is required, UPDATE asks that the input and output files
{both named ACCT) be mounted. If a SORT statement had appeared, the
mount message would have requested the scratch tape $UPDOO.
(ACCT
would not be mounted until pass2). The single UPDATE command can then
be typed at the keyboard:
CL-163411L...IF4=11691

specifying the new zip code of 11691 for the record whose key is
63411.
All other fields remain the same. The user then types CTRL/Z
to specify the end of the input from the keyboard.

8.4

OPERATING PROCEDURES

UPDATE

is called with the Monitor co~.mand:
..!...RUN UPDATE [, filel, ••• , file7] [/s (1) s (2)]

where
filel, ••• file7

are the names of the files which contain the
UPDATE control program (possibly stored in two
segments, FDR and I/O) and the UPDATE conunands {if
previously stored on the system device).
If no files are specified, UPDATE uses the file in
the edit buffer.

/s(l)s(2.)

are switches
The first switch is 1 or 2 specifying the UPDATE
phase.
1
read control program and UPDATE
commands and check for errors, put in scratch
area.
2
read scratch area after sorting and
perform UPDATE. If neither 1 nor 2 is spec.if ied,
passl is assumed.
The second switch is
/N

for no listing (default)

for a listing of UPDATE conunands, {pass2) or
control program (passl).

8-7

If no errors are found in the control program and the UPDATE commands
are sorted, UPDATE continues automatically to pass2. Otherwise, SORT
must be used to sort the UPDATE commands into the same order as the
file.
Type RU SORT,UPCON,UPCOM (control program). SORT asks that
$UPDOO be mounted. When the SORT is complete, call the second phase
of UPDATE. For example,
.RU UPDATE,UPCON,UPCOM/2
NOTE
If the input device specified was KBD, there is no
indication from UPDATE that it is ready to accept
input from the keyboard.

8. 5

ERROR MESSAGES

indicates a fatal error.
Message
BAD DEVICE

Explanation
Illegal device
statement

specified

BAD DIGIT IN DECIMAL INITIAL VALUE
Alpha character
value.

input

decimal

initial

BAD INPUT STMNT

Input statement is
order.

BAD KEY STMNT

Key in KEY statement has bad syntax.

BAD UPDATE COMMAND

Bad syntax, probably missing
comma.

BAD UPDATE STMNT

Update statement is incorrect.

BAD VALUE

Value in update
incorrect.

DATA TABLE OVERFLOW!

Too much data, maximum = 960 characters.

DESCRIPTOR TABLE OVERFLOW!

Too many F, T
maximum = 160.

EXTRA CHARS AT STMNT END

Characters other than A or D occurred in
a data specification statement where A
or D was expected.

FIELD NUMBER MISSING OR 0

Field number or default unit
missing or is 0 or > = 512.

I RECORD ALREADY EXISTS

Tried to
present.

8-8

missing

command

and

insert

too

fields

record

out

long

defined,

number

already

Message

Explanation

INITIAL ALPHA VALUE DOESN'T BEGIN WITH QUOTE
Beginning quotation
initial alpha value.

mark

missing

for

INITIAL VALUE TOO BIG

The initial value specified
than the field size.

larger

INITIAL VALUE TOO SMALL

The initial value specified
than the field size~

smaller

specified

where

MISSING CLOSE QUOTE ON ALPHA INITIAL VALUE
Quotation mark
necessary.

not

MISSING INITIAL VALUE

Comma was inserted after type and size
but initial value was not specified.

MISSING OR OUT OF ORDER

On C or D, record was
Probably because scratch
sorted.

not
there.
file was not

MISSING UNIT

Unit number specified is a
16.

multiple

NO COMMA AFTER FIELD NAME

No comma or a character other than comma
was specified after the field name.

NO DEFINE!

Control
DEFINE.

file did not start with
a
Returns to Monitor immediately.

NO END STATEMENT

END statement
specification.

missing

after

OUTPUT

NONEXISTENT FIELD

Tried to update Fn where no Fn
in file description section.

appeared

NOT A OR D

A character other than A or D occurred
in a data specification statement where
A or D was expected.

NOT I,D, OR C

Bad first character on update conunand.

NOTHING AFTER FIELD NAME

Field type and size are not
after field name and comma.

specified

NUMBER REPEATED OR OUT OF ORDER
A f ieid sequence number is used more
than once or is out of ascending order
sequence.
OUT OF ORDER

SORT statement
missing
commands are not in order.

UNEXPECTED END OF FILE

Missing END statement.

8-9

and

update

CHAPTER 9
BOOT

BOOT is used to bootstrap the system from one device to another, i.e.,
if the system has been moved from DECtape to disk, BOOT may be run to
start the system on disk.
9.1

OPERATING PROCEDURES

The command to run the BOOT program has the form:
~RUN

BOOT/xx

and the CR key:
where
is the switch which indicates the device where the
system to be bootstrapped resides.

/xx

/RK

is the RKOB or RKBE disk unit O.

/RF

is the RF08 disk.

/DT

is the TC08 DECtape unit o.

/TD

is the TD8E DECtape unit o.

/LT

is the LINCtape unit O.

If no switch or an illegal switch is specified, BOOT
error message NO and control returns to the Monitor.

displays

Do not try to bootstrap onto a device which is not ready or
exist.

does

the
not

Other options are available for experienced programmers:
/PT

loads ~11.e high-speed RIM loader and BIN paper tape
loader into field O.

/ZE

zeroes all of core in

BOOT may also be used to bootstrap OS/8.

9-1

CHAPTER 10
PATCH

PATCH is used to fix (or patch) system programs or the Monitor on a
COS system tape. The data to make the changes is provided by Digital
Equipment Corporation as patches in the form of dialogue with the
~a~~tl
progr~n with the user input underlined.
It would be necessary,
therefore, to type only the underlined data exactly as shown and to be
sure that the computer output corresponds with what is illustrated.
If everything matches, the patch is correctly made.
To better
understand the workings of PATCH, it is important to note that both
system programs {files) and the Monitor consist of blocks of numerical
information
(program code; machine language instructions)
on the
system device.
Each block consists of words of 4-digit numbers.
These words are numbered from 0 to 377 (base 8). PATCH reads one of
these blocks, allows user examination and/or change
(patch)
of
individual words in this block and writes the new block back out to
the system device.

10.1

OPERATING PROCEDURES

Be sure the system tape has the current version of the progr~m
patched. The command to run the PATCH program is of the form:

.RU PATCH
and the CR key.
PATCH replies:
COS PATCH SYSTEM VERSION 11.08 (or current version
number)

FILE NAME:
Respond with one of the following:
(a)

The name of the file to be patched.

{b)

/N and/or a carriage return, to indicate a patch for
the Monitor not for a program. The system responds
PATCHING MONITOR.

(c)

CTRL/Z or /X. To indicate end of PATCH operation. The
message EXIT is printed and control returns to the
Monitor (just as if CTRL/C had been typed.)

When typing in letters or digits, a typing mistake can be corrected
(before typing the CR key) by typing CTRL/U and retyping the entire
line correctly.

10-1

The next PATCH question is:
BLOCK:
Type the block number specified on the patch information sheet as sent
out by DEC. This specifies which BLOCK is to be patched. Typing END
or CTRL/Z at this point causes a return to the FILE NAME question. If
at any time before typing END or CTRL/Z it is discovered that the
wrong BLOCK number was typed, type R (for RESTART) and PATCH returns
to the FILE NAME question.
Reenter the file name, then type the
correct BLOCK number and continue from there.
PATCH displays:
LOCATION:
Type in the number provided by DEC. This specifies which word in
block is to be patched. PATCH responds with:

the

OLD VALUE: nnnn
where nnnn is the four-digit old (current) value of this location.
It is very important to verify that this value is the one that DEC
said it should be. If it is not, do not proceed; something is wrong.
a.

Check that everything previously typed was letter perfect.
If the wrong BLOCK number was typed, type R and restart at
FILE NAME.
If the wrong LOCATION was typed, type the CR key
in answer to the NEW VALUE: question. This makes no change
to the location specified. Then type the correct location
number in answer to the LOCATION question:

If everything typed was correct, perhaps an old version of
the Monitor or the system program in question is being used.
Before starting PATCH, the version of the program in question
should be verified.

If everything seems correct but the dialogue doesn't agree
with what DEC predicted, save all output and consult your
software specialist.

When the program asks:
NEW VALUE:
enter the new value as supplied by DEC. This new value replaces the
old value in core.
The block is not changed on the system device
until all patches are made to this block.
If the wrong new value is typed, retype the location, verify that its
current (old) value is the one typed in by mistake, and then enter its
new corrected value.

10-2

If a new value is entered for the wrong location, reenter that
location and as the new value, type its original value (determined by
looking back on the listing for the OLD VALUE at the time this
location was first referenced). If a scope display is used and the
original value cannot be determined, type R and redo the entire block.
Answer NEW VALUE: with the CR key or CTRL/Z if the location
examined and no new value is to be entered.

was

just

When all patches to a block are complete, type CTRL/Z or END to the
LOCATION: question.
At this point, PATCH asks for the RELATIVE
CHECKSUM (similar to a hash total). Type the relative checksum, as
supplied to you by DEC, and the CR key. If every thing is correct,
the new corrected block is written on the system device, and the
message:
NEW BLOCK PATCHED OK
is displayed.
The question BLOCK: is displayed again to allow
patching of another block. Type in the next block number in this file
to be patched or type END or CTRL/Z to terminate patches for the file.
In the latter case, PATCH displays the message:
nn BLOCK(S) PATCHED IN THIS FILE
FILE NAME:
If an error is detected, the message:
BAD CHECKSUM
LOCATION:
is displayed and the faulty block is not written to the system device.
The new block as changed is still in core. Review the numbers typed
in and possibly examine the locations changed to see if they are
correct. If the error is found, fix it and then type END or CTRL/Z to
the LOCATION: message. If an error is not found, type R to restart
the program and try patching the entire block again.

10-3

10.2

ERROR MESSAGES
Message

Explanation

BAD CHECKSUM

An attempt was made to write
which was incorrectly patched.

BAD DIRECTORY

Logical unit 0 was not a system device.

BAD NUMBER

A number with more than 4 digits, a
non-digit or containing the digits 8 or
9 (these are never used) was typed.

BLOCK TOO BIG

An incorrect block number was typed.

FILE NOT FOUND

The program with the name specified was
not found on the system device. It must
be there.
If it is on another tape, it
may be mounted on logical unit 0 when
PATCH first starts up; but thereafter DO
NOT CHANGE UNITS.

LOCATION TOO BIG

A location greater than 377 was typed.

NO CHANGE IN BLOCK

An attempt was made to write a block but

block

It
cannot be larger than the length of the
file being patched.

no changes were made in it.
If this was
on purpose, fine, otherwise make the
changes to the block again.

10-4

CHAPTER 11
DAFT

The DAFT (Dump and Fix Technique) program is similar in function to an
editor but is used for data records. DAFT provides the capability to
search for, examine and change records and to list records or parts of
records on the line printer or terminal. DAFT may be used to make
minor changes to a data file.
DAFT allows one input and one output file to be open at the same time.
These two files may be the same file when in update mode. At any
given time, there is a record from the input file in core known as the
current record.
This record may be modified by the CHANGE command
before it is written on the output file. An output file need not be
open if records from the input file are only being examined.
11.l

OPER~TING

PROCEDURE

The command to run DAFT has the form:
~RU

DAFT[, filel, ••• ,filenJ

and the CR key.
Where
filel, ••• ,filen

11. 2

are previously created files which contain a
string of DAFT commands to be used to dump or fix
the data file. These files are optional and if
not
specified, commands are entered via the
terminal. After the last command in the last file
is executed, additional commands may be entered
from the keyboard.

COMMANDS

DAF.X.. c . ommands.. begin with a keyword
(consisting. o£ any number of
non-blank characters, of which only the first one is significant).
Some commands take arguments, in which case the keyword is followed by
one or more spaces and the arguments.
These arguments may be
separated from
one or more spaces.

In Table 11-l, an expression in brackets
optional.

[ ]

indicates

that

The following symbols are used in Table 11-1 to show the form of
commands:
n

is
DAFT

represents an unsigned non-zero positive integer.
If it is optional in a command and is omitted, n=l
is assumed.

11-l

< a,b >

represents a key field consisting of character
positions a through b inclusive. a and b must
each be unsigned non-zero positive integers and b
must not be smaller than a. If this is optional
in a command, and is omitted, the key specified in
the KEY command is used.

label

is a data file name.

unit

represents an integer in the range 1-15. If this
is specified, the system looks for the specified
file on this logical unit before issuing a MOUNT
message.
If given, /unit must immediately follow
the file name.

indicates that before a record is read from the
input file, the current record in core (if there
is one) is written on the output file. The + sign
need not have a space before it unless it is the
only argument.

data

represents a piece of data.
form:

Alpha

data

has

the

' characters ••• '
and decimal data has the form:
[-]digits •••
Before data is used in executing the command, it
is adjusted to the same length as the key field.
If the data is smaller and alpha, it is left
justified in the field and padded with spaces on
the right.
If it is smaller and decimal, it is
right justified and padded with zeroes on the
left.
If data is larger and alpha,
excess
characters on the right are ignored. If it is
larger and decimal, excess characters on the left
are ignored.
TABLE 11-1.

DAFT COMMAND SUMMARY

Command
Advance [n]

Function
[+]

Advances the input file n records.
is omitted, 1 is assumed.

If n

Backspace [n]

Backspaces n records if the input file
was opened with the UPDATE command. If
n is omitted, 1 is assumed.

Change [<a,b>] data

Replaces the data in the current key
field of the record currently in core
with the data specified.
If the key,
<a,b>, is used in this command, it
temporarily overrides the key specified
by the KEY statement.

11-2

T~..BLE

11-1.

DAFT COMMAND SUMMARY (Cont'd)

Command

Function

Display [n]

Sets the width of the print line of the
listing
device
(terminal
or
line
printer) ton characters (maximum 130).
If n is omitted, this conunand turns on
the grid if it was off and turns it off
if it was on.

'R ; +.....
x ... -

Returns control to the cos
Monitor
providing no output file is open.

Fini [+]

Closes the output file.
If
+
is
specified, the current record and the
remainder of the input file are first
copied to the output file.

Goto n [+]

Makes record n the current record.

Help

Displays a summary of DAFT

commands

the terminal.
Input label [/unit]

Key a,b

Opens the specified file
(label)
for
input.
The first record is read and
becomes the current record.
Sets

the

current

key

character

positions a through b inclusive.

List [n]

[<a,b>]

[+]

Lists on the line printer, n consecutive
records
beginning
with the current
record. The key, <a,b>, if specified,
represents those consecutive characters
in the record that are to be considered.

Output label [/unit]

Opens the
output.

Put [n]

Writes n copies of the record
in core on the output file.

Query

Lists information about the
current
status of the program on the terminal.
This includes the names of the input and
output files, the units where the files
are located, the record in core and the
version number of the DAFT program.

Rewind

Rewinds the input file.
becomes the current record.

11-3

specified

file

(label)

for

currently

Record

TABLE 11-1.

DAFT COMMAND SUMMARY (Cont'd)

Command

Function

Search [<a,b>] data [+]

Searches the current record and then
succeeding records for an occurrence of
the specified data appearing in the
current key field.

Type [n] [<a,b>] [+]

Same as list except
terminal.

Update label [/unit]

Opens the specified file {label} for
updating.
This command can only be
specified for a file with fixed length
records since direct access I/O is used
to move records.

Version

Displays the version number
program on the terminal.

Write [n]

Performs same function as ADVANCE [n] +,
i.e., n-1 consecutive records, beginning
with the current record are written on
the output file. The nth record (after
the original one) becomes the current
record.

Lists the record number and size of the
current record on the listing device
(either
terminal
or
line
printer
depending on whether the last record was
listed via a TYPE or LIST command). The
line printer is the initial listing
device.

11.3

output

the

DAFT ERROR MESSAGES
Message

Explanation

0 NOT ALLOWED

In some commands, 0 is not a perrnissable
argument.

BAD DIGIT IN DATA

In a CHANGE or SEARCH
command,
a
character other than digits or a minus
sign is contained in a numeric data
field (a field without quotes).

CANT BACKSPACE PAST BEGIN OF FILE
Attempted to backspace past beginning of
file.
The first record becomes the
current record.
CANT BACKSPACE WITH SEQUENTIAL INPUT
Attempted to backspace
input.

11-4

with

sequential

Message

Explanation

END OF INPUT FILE AT RECORD nnnn

An attempt was made to read past the end
of file mark on the input file. This is
not necessarily an error. nnnn was the
last record read.
The input file is
closed.
EXCESSIVE GRID SIZE

The grid

{printer

width)

may

not

after

the

greater than 130 characters.
EXTRA CHARS

Extra characters were found
logical end of a command.

ILLEGAL RECORD - CLOSING FILE The file being updated contains a bad
record (one not the same size as record
1).
Only fixed length records
are
permitted on such files. The file is
closed.
KEY ENTIRELY PAST END OF RECORD
The key specified in a LIST or TYPE
command began with a character greater
than the record size.
KEY EXTENDS PAST RECORD END
A change or search was attempted with a
key that extends past the end of a
record. However, a list with such a key
is possible.
In such a case the list is
terminated at end of record.
KEY TOO BIG

The range of the key cannot
characters.

exceed

100

NO DATA

Data must be specified in
SEARCH command.

CHANGE

NO INPUT FILE

The command requires an input
one is not open.

file

but

NO LABEL NAME

The filnam was omitted in
OUTPUT or UPDATE command.

NO OUTPUT FILE

The command requires an output file but
one
is
not open.
The command is
terminated at the point just prior to
writing the current record on the output
file.

OUTPUT FILE ALREADY OPEN

A request was made to open an output
file while one was already open. Only
one output file may be open at a time.
The request is ignored.

11-5

INPUT,

Message

Explanation

OUTPUT FILE STILL OPEN

An EXIT cannot be made when the output
file is still open. The output file can
be closed with the FINI command or
CTRL/C may be typed.

PUSH DOWN OVERFLOW

The program will abort with this message
when too many errors are made.

11.4

DAFT OUTPUT

Records may be listed with a grid above them. This grid consists of
two lines of numbers showing the character positions. The lower of
the two lines represents the units digits of the column counts.
The
upper line represents the tens digits. The tens digits are printed
for the first and last character in the record (or part of the record)
listed or whenever it changes.
If there is a hundreds digit, it is
printed in colunm 1 or whenever it changes.
Refer to Figure 11-1 for a sample of a DAFT operation.

11-6

COS MO~ITOR 2.0606
= R DAFT
*HELP
ADVA;-.JCE N+
BAC?<SPACE N
CHA~GE <A, B> DATA
DI SPLAY N
EX:! T
FINI +
GOTO V+
HELP
I ~PUT LABEl../tNI T

LI ST N '{E'f+
0 UTPUT LABEL/lNI T
PUT \J
QUERY
RE\'I~D

SEARCH <A, B> DATA +
TYPE 'J <A, P>+
UPDATE LABEl./lNI T

VERSI O'l
WRITE "J

* VERSI O"J
DAFT VERSIO\J 6·06
*V
DAFT VERSIO~ 6·06
*VERY
DAFT VERSIO~ 6·06
*I )JPUT MAI LNG
M Qt.NT MAIL'.\JG #01
FOR

I 'JPUT:

1
*TYPE 1
RECORD 000001 OF FILE ~AIL~G,

RECORD 1..E\JG1d=204 CHARACTE&S

DIGITAL EQUI PME'JT CORP·

D·

12-3

146 "!AI 1\1 ST·

"1AY\JARD

F·

PAv'LOC:{

MA017540Sl~-1

0012345Alllll

Figure 11-1.

DAFT Sample

2003SUPFRVISOR

11-7

06217

*DI SPLAY 70'\

*T 1
RECORD 000001 OF FILE MAIL\JG1
DIGITAL EQUI P'.'lf E"J T CORP•
146 '.'tfAI \1
111
*D

RECORD L E\JG11-i=204 CHARACTERS

D· F• PPVLOCK

st.

12-3

MAY\JARD

MA017540S/3-1

0012345All

062172003SUPERVISOR

*T 2
RECORD 000001 OF FILE '.'t1AIL~G1
0

RECORD LE~Gl"1=204 cqAf.ACTEfiS
3

1234567g9Q123456789012345678901234567g9Q1234567~9012345673901234567~~0

DIGITAL EQUIPMENT CORP·

D· F• PAVLOCK

12-3

g
9
10
1
2
3
4
1234567890123456739012345678901234567d9012345673J012345679901234567890
146 ~AI~ ST·
MAY'lARD
~A017540S/3-1
0012345A11

1234567g90l234567q901234567g90l2345673901234567~901234567890123

111

062172003SUPERVISOR
RECORD LE~GTq=204

RECORD 000002 0 F FILE MAI L\JG1
0

c~rnRACTEHS

1234567690123456789012345673901234567890123456789012345678901234567d90
DIGITAL EQUI PME'lT CORP· K • RI c:.r ER
12- 3
7

12345678901234567890123456789012345678901234567~9012345678901234567890

146 MAI~ ST•
14

MAYNARD
6

MA01754COS 300 0001972! 3
8

123456789012345678901234567~90123456789012345678901234567890123

053172002WRITER

* T 2<251 50>

RECORD 000002 OF FILE MAIL~G1
2

RECORD LE'.'JGH=204 CiiARACTERS

5678901234567~901234567890

K· RI C>-IER

Figure 11-1.

DAFT Sample

11-8

(Cont'd)

RECORD 000005 OF FILE MAIL'JG,
DEC
*Q
I \lPUT FILE: MAIL\JG
U!\JI T: 00
0 UTPUT FILE: 1"11 O\J El
UNIT: 00
:< EY=< 001, 050>

RECORD LENGTi=204 CiARACTEES

S· G• WEl.LCOME
OPE\l

RECORD 000005 OF FILE ~AIL'JG,

RECORD LE'\JGT-1=203 C-IARACTERS

DAFT VERSIO\l 6·06

*Q
INPUT FILE: ~Ail.'JG
U\JI T: 00
0 UTPUT FILE: /\JO\JE/
U\ll T: 00
KE'f=<OOl,510>

OPE'J

RECORD 000005 OF FILE MAivrn,

RECORD LE:-JGTi=2C4 CiARACTERS

DAFT VERSION 6·06
; SHOULD CAUSE A\l ERIWR MESSAGE
CA~T BACKSPACE WITH SEQUE"JTIAL I'JPUT

.Figure 11-1.

DAFT Sample

11-9

CCont' d)

RECORD 000003 OF FILE MAILl\JG,

RECORD LENGTH=204 CHARACTERS

2
3
4
5
56789012345678901234567890
S• RABP.JOWI TZ

*T 2
RECORD LE!'lGT~=204

RECORD 000003 OF FILE MAIL:\JG,
DIGITAL EQUIPMENT CORP·

S· RABI NOWI Tl

146 MAIN ST·
345

C~ARACTERS

12-3

MAYNARD

MA01754COS 300 0099999C12

062172009COMPILER

RECORD 000004 OF FILE MAILNG,
DIGITAL

RECORD LENGTri=204 CHARACTERS

12-3

R•' LAR'f

146 MAIN ST

MA01754

MA'f\JARD

061272999 SORT

*D
*KEY 1, 50

; SET PERMA'.°'JFNT :<E'f

* T2

RECORD 000004 OF FILE "'!AI L'JG,

RECORD LEVGT'-1=204 CiARACTERS

123T5678901234567~901R3456789012345678901234567390

R· LARl'

DIGITAL

*T 2

RECORD 000004 OF FILE MAIL\JG,

RECORD LE\JGT'1=204 C:.tARACTERS

i-23456789012345678901234567~901234567139012345678-JO

DI GI TAL

R· LAR'f

RECORD 000005 OF FILE MAIL\JG,

RECORD LEr>JGTfi=204

C~ARACTERS

12345679901234567890123456739012345673901234567890
DEC
S· G· WELLCOME

*D
*T

Figure 11-L

DAFT Sample

11-10

(Cont'd)

0000000

CHAPTER 12
COS-OS/8 CONVERTER

The COS-OS/8 Converter program, CONVEX, provides the capability to
convert source or data ASCII files created with OS/8 to COS source or
data files.
COS data files can be converted to OS/8 ASCII files.

12.1

OPERATING PROCEDURES

The command to run CONVEX has the form:
.RUN CONVEX
and the CR key:

CONVEX replies with the progra~ version nti'!lber:
VERSION 12.28

(or current version number)

And the message:
INPUT MODE-

Answer with the letter code:
A
D

(OS/8 ASCII file mode)
(COS Data file mode)

and the CR key.
It then displays the message:
FILE NAMEThe answer has two forms depending on the mode.
1.

For mode A (OS/8 ASCII) type the OS/8 file name (one to six
characters optionally followed by a period and up to a two
character extension) followed by a comma and then the de~1ice
name which is one of:

DTn
DKn
where n is a single digit referring to the physical unit #
(DTO-DT7 or DKO-DK3).
If no device is specified, CONVEX uses
the system device.

12-1

Examples:
ABC,DTS
RICH.TM,DKO
2.

For mode D (COS data) type the COS data file name (one to six
characters) optionally followed by a slash and the logical
unit number (1-15). If a logical unit is specified, the
system first checks that logical unit for the specified file
before issuing a MOUNT message.
Examples:
CUST/5
LIST/l
PROS PT
Illegal:
CLIENT ,DTl

If everything is ok, the program displays the message:
OUTPUT MODE Type one of the following one character codes:
A
D
S

(OS/8 ASCII file)
(COS Data file)
(COS Source files)

The next question is:
FILE NAME If the option
explained in
CONVEX asks:

specified was
steps l and 2.

A or D, reply following the rules
If an existing file is to be destroyed,

REPLACE?
Type Y for YES, N or any other character for NO and the CR key.
If the option specified was s, type the filename, optional
COS device e.g. GROUPA,DT3.
This device must be physically the same as the
tape to tape or disk to disk.

system

comma

and

device,

i.e.,

specified.

Type

When the conversion is complete the message:
INPUT MODE is displayed again and another
CTRL/C to return to Monitor.

conversion

12-2

can

NOTE
If instead the message:
OUTPUT MODE is displayed again after successful output has
occurred (MODE A) , CONVEX needs more space to
continue. Specify an additional output file.
NOTE
OS/8 files are always generated so as to be
multiples of 16 blocks long. For this and other
reasons, the resulting OS/8 files may be longer
than necessary.
To recapture the wasted space,
use OS/8 PIP to copy the file back to itself using
the /A option.

12.2

Error Messages

In some cases, illegal or bad syntax on input causes the
be reasked.

question

Other errors generate one of the following:
Message

Explanation

BAD CHAR

Warning message logged for each attempt
to convert an OS/8 character you tried
to convert to COS which is not in COS
character set.
Characters such as t,
CTRL/Q, etc. Note: + is trans lated to
<, tabs are converted to proper number
of spaces. Vertical tabs, line feeds,
form feeds,
rubouts etc. are ignored.
Lower case characters are converted to
upper case.

BAD DEV

Specified an output device with
-s
option and output device was not same
type physical device as system device.
e.g.
FILE,DKl when system device is
OTO.

BAD DIRECTORY

Attempted to reference or store a
on a device with no directory
directory that has been destroyed.

FULL

Specified an output device with
-s
option and source file being created was
filled.

NO END

Entire OS/8 ASCII input file has been
read but no end of file (CTRL/Z) was
found.

12-3

file
or a

Explanation

Message
NO ROOM

Directory of output device was full;
room for specified output file.

NONCE ERROR

User tried to use
a
feature
not
currently S'Upported or documented {but
might be implemented in some future
release).

NOT FOUND

File with specified name was not found.

12-4

CHAPTER 13

FORMAT PROGRAMS

There are four format programs provided with
mark disk or DECtape for use with cos.

the

COS

package

which

These programs are:
RK8MRK

Formats an RK08 disk.

RKEMRK

Formats an RK8E disk.

TDMARK

Formats a DECtape on a TD8E drive.

DTMARK

Formats a DECtape on a TC08 drive.

These programs are described in detail in diagnostic documents
(RK08
Disk Formatter, DEC-08-DSKB; RK8E Disk Formatter, DEC-08-DHRKD-A-D;
TD8-E DECtape Formatter, DEC-8E-EUZC-D; TCOl - TUSS DECtape Formatter,
DEC-08-EUFB-D)
provided with the system.
The procedures in the
following paragraphs cover the use of these programs with the COS
system.

13-1

IRKBMRK

13.1

RK8MRK

Mount the system tape or disk and the disk to be formatted.

13.1.1

Operating Procedure

The command to run RK8MRK is :
~RUN

RK8MRK

and the CR key.
RK8MRK replies with the question:
DRIVE NO.?
Type the unit number of the RK08 disk to be formatted.
Be sure that
the unit specified is not a good system device since formatting
destroys the contents of a disk.
DRIVE NO.?

RKBMRK displays the question:
WRITE STANDARD FORMAT? (TYPE "Y" FOR YES)
Type Y to format the disk for use with cos. Any other reply
a non-standard format which can not be used with cos.

provides

When formatting is complete RK8MRK displays the messages:
DISK FORMATTED
and
DRIVE NO?
At this point specify another drive to be formatted or type CTRL/C
return to the COS Monitor.

13.1.2

Error Messages

All error messages displayed contain the track
number where the error occurred. For example:
TRACK 0104 SECTOR 02
PARITY ERROR

13-3

number

and

sector

Message

Explanation

CONTROL BUSY ERROR

Disk IOT issued when control
which would effect operation.

DATA RATE ERROR

The processor was busy and did not
respond to a data break request within
the 13 micro seconds required.
The
transfer is terminated immediately.

EXEC WORD ERROR

error occurred while
EXEC. word of the sector.

reading

the

FORMAT ERROR

reading

the

FORMAT PARITY ERROR

A parity error occurred in either the
Header word or Exec word while reading
the sector.

PARITY ERROR

A bit in data, parity or timing has been
picked up or dropped on read. Transfer
continues to the end of the sector where
the error occurred. Word count, current
address information can be used
to
identify the error.

SECTOR NO GOOD

The program attempted to read or write
data on a sector whose header words
indicated a bad sector. The transfer is
terminated immediately.

TIME OUT ERROR

The control
did
not
complete
operation after 32 revolutions.

TRACK ADDRESS ERROR

Track, surface or sector address read
from the disk did not agree with the
address count registers or the disk
drive
electronics
indicated
track
position 000 and the track counter did
not agree.
The transfer is terminated
immediately.

error occurred
header word.

while

was

busy

TRACK CAPACITY EXCEEDED ERROR
The program attempted to read
beyond sector 17.
WRITE LOCK ERROR

write

The program attempted to write a section
that was write protected.
The write
operation is terminated immediately.

Error Recovery
If the program halts with the program counter = 1337 indicating
non-existent drive or device not adjusted correctly. If CTRL/C does
not work, restart the COS system.
(Refer to Appendix B.) If an error
reoccurs, the disk may be bad.

13-4

At the end of each track in which an error occurred on the
read operation the message:
RE-WRITE TRACK (Y FOR YES)?
is displayed.

Type Y to rewrite the track.

13-5

write

and

RKEMRK

13.2

RKEMRK

Mount the system tape and the disk to be formatted.

13.2.l

Operating Procedure

The command to run RKEMRK is:
.R RKEMRK
and the CR key.
RKEMRK replies with the following:
RK8E DISK FORMATTER PROGRAM
FOR ALL QUESTIONS, ANSWER Y FOR YES OR N FOR NO.
The following questions are asked:
FORMAT DISK 0?
FORMAT DISK l?
FORMAT DISK 2?
FORMAT DISK 3?
ARE YOU SURE?
After answering Y or N to each of the above questions the
formatted and the following message is printed:

disk(s)

RK8E DISK FORMATTER PASS COMPLETE
FORMAT SAME DISK(S) AGAIN?
Answer Y or N. At this point each of the FORMAT DISK n? questions are
printed again. Exit to Monitor by typing CTRL/C or continue procedure
for formatting disks.

13.2.2

Error Messages

The following error messages may be displayed
disk for COS:

when

marking

RK8E

DISK DATA ERROR
READ STATUS ERROR
WRITE STATUS ERROR
RECALIBRATE STATUS ERROR
After the error header mentioned above is displayed, the terminal will
print some of the following error information pertaining to the
failure.

13-7

PC:

PROGRAM LOCATION OF FAILURE

CD:

EXPECTED INFORMATION

ST:

CONTENTS OF THE STATUS REGISTER

CM:

SOFTWARE COMMAND REGISTER

DA:

SOFTWARE CYLINDER, SURFACE, AND SECTOR REGISTER

AD:

ADDRESS OF DATA BREAK

DT:

DATA FOUND DURING DATA BREAK

After the error information is displayed, the terminal will display
one of the following questions asking the error recovery desired.
1.

If the error was a recalibrate error, the following
will be displayed:

question

TRY TO RECALIBRATE SAME DISK AGAIN?
Typing a Y for yes will result in a repeat of the recalibrate
sequence on the disk in error. Typing N for no will result
in progressing to the next available disk.
2.

If the error was a write error, the following
be displayed:

question

will

TRY TO FORMAT SAME CYLINDER AGAIN?
Typing Y for yes will result in a repeat of the write
sequence on the current cylinder.
Typing N for no will
result in progressing to the next sequential cylinder.
3.

If the error was a READ or
question will be displayed:

check

error,

the

following

TRY TO CHECK SAME CYLINDER AGAIN?
Typing a Y for yes will result in a repeat in the READ and
check sequence on the current cylinder. Typing a N for no
will result in progressing to the next sequential cylinder.

13-8

TDMAR..T{

13.3

TDMARK

TDMARK records the timing and mark tracks on a DECtape mounted on
TD8E transport.

13.3.l

the

Operating Procedures

The command to run TDMARK is:
.RUN TDMARK
and the CR key.

TD.MARK replies with the question:

UNIT?
Before replying to this question, mount the DECtapes to be marked with
just enough turns on the right hand reel to provide a grip. Make sure
that no two tape units are set to the same unit number; set the WRITE
ENABLE/WRITE LOCK switch to WRITE ENABLE and the REMOTE/OFF/LOCAL
switch to REMOTE. Set the WTM (Write Timing Mark) switch to on (in
down position).
This red switch is located on the upper right hand
side of Module 868
(printed circuit card with wires) within the
processor (the black box behind the lights and switches).
Answer the UNIT question with one or two unit numbers, corresponding
to the units of the tapes to be marked. For instance, if tapes are
mounted on units 0 and 1, type:
0 1)

Spaces are ignored, so it makes no difference if spaces are typed
between the unit numbers.
Only one specification of a unit is
significant, i.e., typing 000111 has the same effect as typing 01.
Only tapes on units 0 and 1 may be formatted.
If an error is detected
in the response to this question, the question is repeated.
TDMARK
replies with:
FORMAT?
Type MARK and the CR key and the program assumes 201 words 2702 blocks
{standard PDP-8 format). Any other reply produces a format which is
not compatible with cos.
TDMARK displays the message:
0201 WORDS,2702 BLOCKS.OK?(YES OR NO)
If NO is typed, the program reverts to the FORMAT? question.
the program displays:
SET SWITCH TO WTM

YES

Type the CR key and if the switch is set, the tape on the
specified begins to move.

first

unit

Once all of the tapes specified have been marked, the message:
SET SWITCH TO OFF
is displayed. Reset the WTM switch to normal (up position)
and type
the CR key to start the second pass. Note that during the second pass
with multiple DECtape units, as soon as one tape stops, the next tape
starts; the first tape is completed and may be replaced with a fresh
tape in preparation for formatting additional tapes. When formatting
is completed, the message:
FORMAT
is displayed. Typing SAME and the CR key repeats the entire process
with the same format.
The new DECtapes must be mounted and ready
before the CR key is typed in response to the SET SWITCH TO WTM
message.
When all tapes are formatted, type CTRL/C to return to the
COS monitor.

13.3.2

Error Messages

The following error messages may be displayed when marking a tape

cos.

for

Error messages for response to SET SWITCH TO WTM:
Explanation

Message
SETUP

Indicates an error in the DECtape setup.
One of the units specified is in WRITE
LOCK position, not selected, or the
write flip-flop is unable to be set, or
there may be a timing error.
(After
UNIT message is displayed.)

SWITCH NOT SET TO WTM OR SINGLE LINE FLAG
SWITCH TO WTM.

FAILED

SET.

SET

Indicates the switch on the M868 module
is not set to the WTM position or the
timing generator for writing the mark
and timing tracks is not setting the
single line flag.
If the switch was not set to
WTM
position, set the switch and type the CR
key.
If the switch was set to WTM position
and this error occurred, try again or
examine the timing generator circuit.

13-10

Error messages for marking and verifying a tape:
PC
PC
PC
PC
PC
PC

xx xx
xx xx
xxxx
xx xx
xx xx
xxxx

BLOCK NUMBER ERROR PHASE n
CHECKSUM ERROR PHASE n
DATA ERROR PHASE n
MARK TRACK ERROR PHASE n
TIMING ERROR PHASE n
WRITE ERROR PHASE n

where xxxx equals the program counter {PC) at the time of the failure.
n equals the pass in which the error occurred.
(See Recovery=)
Recovery:
Although an error should cause doubt concerning the entire process. a
restart may be made (except in phase 0) by typing RETRY.
RETRY
causes the program to go back to phase 1. Type RESTART
to return to
the UNIT? question.
Phase 0:
Phase 1:
Phase 2:
Phase 3:
Phase 4:
Phase 5:

Write timing and mark track forward.
Reads mark track reverse.
Write data, forward block and reverse block
numbers forward and writes the checksums.
Displays block numbers in AC (accumulator)
reverse.
Reads data, forward block and reverse block
numbers forward and calculates the checksum.
Reads reverse block numbers in reverse.

The entire program may be restarted at 0200 any time.

13-11

jDTMARK

13. 4

DTMARK

DTMARK records the required timing and mark tracks on a DECtape
mounted on the TC01-TU55 DECtape unit. Up to eight DECtapes may be
formatted at a time, assuming the system has eight tape transports.

13.4.1

Operating Procedures

The command to run OT.MARK is:
..!..RUN DTMARK
and the CR key.

DTMARK replies with:

OTA?
Mount the DECtapes to be marked on the tape transports, with just
enough turns of tape on the right hand reel of each transport to
provide a grip. Make sure that no two tape units are set to the same
unit number.
Set the WRTM (WRite Timing Mark} switch located on the
hardware frame to the down position {on). Set the WRITE ENABLE/WRITE
LOCK switch to WRITE ENABLE and the REMOTE/OFF/LOCAL switch to REMOTE.
Reply to the DTA? question with a unit number or series of unit
numbers, corresponding to the DECtape units where the tapes are
mounted. For instance, if tapes are mounted on units 2 and 3, type
2 3 and the CR key. Spaces are ignored, so it makes no difference if
spaces are typed between the unit numbers.
O~ly one specification of a unit is significant, i.e., typing
has the same effect as typing 2 3) •

2 2 3 3)

The program displays:
DIRECT?
Type MARK and the CR key to format the DECtape for use with cos.
Any
other reply causes a format which is incompatible with cos.
(Refer to
TC01-TU55 DECtape Formatter Writeup.)
The program assumes 201(8)
words, 2702(8)
blocks
(standard PDP-8 format)
and displays the
message:
0201 WORDS,2702 BLOCKS.OK? (YES OR NO)
If a NO) is typed, the program reverts to the DIRECT? question.
YES), the tape on the first unit specified begins to move.
Once all of the tapes specified have been marked, the message:
SET SWITCH TO NORMAL

13-13

is displayed.
Return the WRTM switch to normal (up position) and type
the CR key to start the second pass. Note that during the second pass
with multiple DECtape units, as soon as one tape stops, the next tape
starts.
The first tape is completed and may be replaced with a fresh
tape in preparation for rerunning the program to format additional
tapes.
When DTMARK has formatted the tapes specified, the message:
DIRECT?
is displayed again. Type SAME and the CR key to repeat the entire
process with the same format.
The new DECtapes must be mounted and
ready before SAME) is typed.
(0 means DECtape unit 8).
When all tapes are
Monitor.

13.4.2

formatted,

type

CTRL/C

return

the

COS

cause

Error Messages

Errors typed in response to the DTA? and
return to DTA? or DIRECT? respectively.

DIRECT?

questions

Error messages for response to OK? (YES OR NO) :
Message

Explanation

SETUP?

Indicates an error in the DECtape setup, unit in
WRITE LOCK position, nonselectable unit or switch
not in WRTM position.
After display of this
message, the DTA? question is displayed.

Error messages which may occur during marking and verifying a tape:
xxxx SHOULD BE yyyy BLK ERROR PHASE n
xxxx SHOULD BE yyyy DATA ERROR PHASE n
END TAPE ERROR PHASE n
LAST INT NOT END ZONE
MARK TRACK ERROR PHASE n
PARITY ERROR PHASE n
SB!iE€T- -E-ROOR PHASE- n

TIMING ERROR PHASE n
Recovery from error messages:
Although an error should cause doubt concerning the entire process,
restarts may be made (except when in phase 0), by typing RETRY and the
CR key.
RETRY causes the program to return to PHASE 1. Type RESTART
and the CR key to return to the DTA? question.
PHASE 0:
Mark track write.
PHASE 1:
Writing last reverse block number forward.
PHASE 2:
Writing block numbers and data in reverse.
PHASE 3.
Reading and checking block numbers and data.
If it is necessary to restart the DTMARK program, type CTRL/C to
return to the COS Monitor then type RUN DTMAPJ<.
If CTRL/C does not work, rebootstrap the
B) and then call DTMARK.

13-14

cos system (refer to Appendix

CHAPTER 14
CREF

CREF (Cross REFerence) is a utility program.
Primarily an aid to
program development, it provides an alphabetical listing of all
symbols used in a DIBOL program, along with the line number where each
symbol is defined and all the line numbers where each symbol is used.
CREF requires BK of core and can handle any reasonable BK program.
If
12K of core is available CREF expands its tables to make use of the
extra space. CREF expects the source program to be free from errors.
Only a very minimal amount of error checking is preformed by CREF and
no attempt should be made to CREF programs with compilation errors.
If CREF finds a line it cannot handle it will print:
xxxx IS BEING IGNORED
where xxxx is a line number.
For an example of CREF operation, consider the listing in Figure 14-1.
It is a short program that computes the square roots of numbers from 1
to 99.
(Output of program is shown in Figure 14-3.} Its CREF table,
Figure 14-2, illustrates the basic functions of CREF.
The columns of the CREF Table are:

14.1

NAME

Name of the symbol.

DEF

Line number
defined.

program

where

symbol

REFERENCES

Line numbers
referenced.

program

where

symbol

OPERATING PROCEDURES

Type:
~RUN

CREF[,filel,file2, ••• ]

and the CR key.
Where
filel,file2, •••

are the parts of a DIBOL program, just as
in running COMP (maximum 7).
If no files
are specified, the Editor scratch area is
used.

CREF reads the DIBOL program, produces its table on the
and returns control to the COS Monitor.

14-·l

line

printer

cos 01Bo~

01•MAR-73 THUR
COMPI~ATION ~1ST1NG
v1212e PAGE ~1
DATA DtyIS!ON
S Q U A R E R 0 0 T S U B R 0 U ! I N t.

0010
0020

0030
0040
0050
0060
21070
e~s0

H90
"~00

0U0
0~20

START
A. ,

s,
x,.

SGIRTX,
TEMPX,
N•

s Q UAR E

R 0 0

s ue R 0 uT I NE

RECORD PRINT
A15
Ali5

RECORD
015

D15
015
03,001

Figure 14-1. Square Root Subroutine
(Sheet 1 of 3)

14-2

COS DlSO~
~1•MAR•7~ THUR
PROCEDURE OIV!SlON

coMelL.A!lON LISTING

PROC l.
~OOP1

tNtTC~4L,PT)
X!N•1~0~001

CAL.L SQRT
ARN

S • SQRTX 1 fXXX~XXX~

INeR N

XM!TC~ 1 PRlNT>

(N~~t~99)

STOP
SQRT,

SQ~OOP 1

GOTO ~OOP

TtMPXeX/2
ITR1A~ VA~UE
SQRTX,CT£MPX+(X/TtMPX))/2
1N£WTON1S ME!HOO
Ir(SQRTX•TtMPX ,EQ 1 0) RETURN
TE:MPX1SQRTX

ENO

CO'O SQ~OOP

Figure 14-1. Square Root Subroutine
(Sheet 2 of 3)

14-3

co~
#

H11

01fiOL.

01 111 MAR11t7~

RECORD

AL.PMA
Al..PMA
OECMAL.
OECMAI..

01
01

'• ', l

DE:CMAL.,
OECMAI..

01
llJ1
"1

• '• 1000

OE:CMAL,

HJS5

SQRTX

0012
002,3

011,
00~'

HU
HV
HZI

STORAGE MAP L.ISTtNG
DIM SUE
ORIGIN

TYPE

0H~
H0~

H5'6
H217
H10

THUR

NAME:

TEMPX

OECMAf.

L.Aen.

L,OOP
SQRT

·, ·• xxx·.
.• 99

L.ABF:l.

I rt

AL.PMA
OECMAI,.

·• '2

OE'.CMAL.

SQL,OOP

·, ·.

L.ABEL.
OE:CMAI.

u
u

32
15
15
15

200021
200212

201.00

15
"~
u
01
e1
n

01
00

v1212e

PA C.Jf: 03

20021
2"042
202161

20117
2"'122
10871

20U3
1013~

20132

2'2
"1
01
01

20143
UH,41
20144

2~141

H16 SYMBOl.S

NO tRRORS o£T£cT£o.

me K oORE REQYlREP C3042l f'RE:E L.OcS Cl 14 slJrfERSj

Figure 14-1. Square Root Subroutine
(Sheet 3 of 3)

14-4

Co$3H Ci:lEF'

V1l 108

OEF'

REP'ERENCES

STMSOii.
A

421

LiOC>P

160
110
30
50
260

§9~00P

SQRT

SQRTX

i'tMPX
x~

25i1

1U
220
160

THUR

0hMAR•73

180

200

22e

270
260
260

26~
27~

21~-

19'11
29"

170

9(ll

19~

2~0

10"
80

25Z

260

1(>21

Figure 14-2.

250

280

CREF of Square Root Subroutine

14--5

pAGE 1

1
2
3
4
5
6
7

1,00"
1~414

l,'132
2~000

2·. 236
2·, 449
2J645

2·1 a2e

3,~00

12
13
14
15
16

3~1,62

3,316
3,464

3,605
3,741
3,872
4.000

4~1,23

4·. 242

l,9

4,358

12'.~

?.1

?.2
23
~4

4,472

4;ss2
4ft690

4,795
4,898
5·, 000
5'~2l99

5', l 96

28
29

5,291

5·, 477
5,567

31
32
33
34
35

5,385
5R656
5,744
5,830
5,916

61el00

6,082
6·,16 4

3A
39
40

41
42
43
44
45
46
47
48
49

6 -. 244
6 ·, 324

e.403
6,480
6 •• 557

6", 633
6,108

6!182

6', 855
6,928

52
53

7,2100
, .• 071
1·, 141
7' ~11
7·, 280

54
55

7, 348
1·, 416

Figure 14-3.

Output of Square Root Subroutine
(Sheet 1 of 2)

14-6

,,,...,

s:;;6,

57
58
59

7.463
7.549
'7, 615
7', 681

7, 745

1,810

62
6:5

64
65
66

7,974
7,937
8,000
R•el62

er.124

a,J.65

68
69
70

8', 246

a,306
8,366
8 1 426
8,485

8,544

74
75

s,~02

77
78

79
30
81
82
83
84
85
86
87

90
9 j,
92

93
94
95
96
97
98
99

A 1 660

a 111
1

a·ft 774

~,SH

e,aes

~r, 944

9,000
9,a55
9·, 110
9', 165

9,219
9~273

9,327
9,380
9,433
9.486
9,539
9,~91

98643

9~, 695

9,146
9,197
9,848
9 '• 899
9,949

Figure 14-3.

Output of Square Root Subroutine
(Sheet 2 of 2)

14 ......J

14.2

ERROR MESSAGES

Although CREF expects the source to be error free,
errors are detected.
xxxx IS BEING IGNORED

CREF has encountered a
interpret.

certain
line

types
it

cannot

In addition to the above message, CREF prints an informative message:
SYMBOLS DEFINED BUT NOT REFERENCED; xx
Symbols were found that were defined but
not referenced by the program.

14-8

APPENDIX A
COS CODES

Numbers are stored 2 characters per word in 6-bit binary form
{combinations of 1 and 0). Negative numbers are stored with the high
order bit of the low-order digit set to 1. For example, the number
1234- is stored as two words in the form

WORD 1
1

WORD 2
{with high order bit on)

and would be recognized as 123T. This means that any program where
the decimal to alpha conversion is neglected might produce negative
numbers with letters. Refer to Table A-1 for a list of equivalent
characters for negative

nu.~hers.

The sequence of the COS codes is the same as the output from the
utility.

A-1

SORT

TABLE A-1

CHARACTER

EQUIVALENT
NEGATIVE
NUMBERS

0
1
2
3
4
5
6
7
8
9
A
B

D
E
F
G
H
I
J
K

L
M

N
0
p

Q
R

x
y

z
null

-o
-1
-2
-3
-4
-5
-6
-7
-8
-9

COS CODES

COS CODE
21
22
23
24
25
26
27
30
31
32
42
43
44
45
46
47
50
51
52
53
54
55
5-6
57
60
61
62
63
64
65
66
67
70
71
72

73
00

A-2

CHARACTER

COS CODE

#
$
%
&

(
)

=
~

02
03
04
05
06
07
10
11
12
13
14
15
16
17
20
33
34
35
36
37

[

\
]
t

76
77

Leader/Trailer
LINE FEED
Carriage Return
Space
RUB OUT
Blank
Bell
TAB
FORM

APPENDIX B
LOADING COS

The COS system is supplied on DECtape (LINCtape for PDP-12 users).
It
is suggested that the user copy the COS tape on a certified DECtape
(or LINCtape) before using the system (Refer to Chapter 5.} The copy
can be used and the original stored in a safe place in case part of

the system should be inadvertently destroyed during snhsequent usee
The Monitor can be put into core memory in one of three ways. If the
computer is equipped with the automatic load feature, the Monitor is
loaded by pressing the INITIALIZE switch under the table or by raising
the SW switch on the computer console.
Otherwise the Monitor must be loaded into memory using the appropriate
Bootstrap routine.
The editing scratch area is not
start-up.

automatically

erased

during

system

If the system is equipped with an Analex printer, the printer must
turned off until the system is bootstrapped and re-configured.

1.0

AUTOMATIC LOAD

If one of the optional automatic bootstraps
MI8-EC
MI8-ED
MI8-EH
MI8-E

for
for
for
for

the
the
the
the

TC08
RK8
TD8E
RF08

was purchased, push the SW switch from down position to
Then return it to the down position.

2.0

position.

the

following

TC08 DECTAPE AS THE SYSTEM DEVICE

If your system uses DECtape as the system device,
procedures:

use

Mount the COS DECtape (DEC-08-0COSA-D-UO) on DECtape unit O.

Set the LOCAL/OFF/REMOTE switch to REMOTE and the WRITE
ENABLE/WRITE LOCK switch to WRITE LOCK. Set SWITCH REGISTER
to zero {all switches down} and press EXTD ADDR LOAD.

Enter the binary equivalent of the following locations and
instructions
(refer
to
Chapter 1 of Introduction to
Programming for information on binary numbers). To bootstrap
a TC08 system:

B-1

LOCATION

INSTRUCTION

7613
7614
7615
7616
7617
7620
7621
7622
7623

6774
1222
6766
6771
5216
1223
5215
0600
0220

7754
7755

7577
7577

Set the SWITCH REGISTER (SR)
to enter the first location
(7613)
and press ADDR LOAD. Set the switches to enter the
first instruction and push up on DEPosit switch.
Continue
entering instructions as described up to and including
instruction 0220. Then set 7754 in SR and press ADDR LOAD.
Set SR to 7577 and push up on DEP switch twice.
4.

Set terminal REM/LOC switch to REM.

Set SR to binary equivalent of 7613; press ADDR LOAD, CLEAR
and CONT keys in that order. The Monitor should respond with
the message
COS MONITOR 2.1108

(or current version
number)

DATE?
Enter the DATE command as described in Chapter 2.
At this
point the SYSGEN program can be used to change the system
unit number assignment (refer to SYSGEN, Chapter 3).
Set the switch to WRITE ENABLE.
6.

2.1

If the Monitor does not respond correctly, start
and reload the bootstrap routine.

step

SYSTEM RESTART ON DECTAPE

If the system halts or CTRL/C is inoperative, use the automatic
bootstrap switch to restart the system. Each time the Monitor is
reloaded, the old date is erased and the Monitor requests a new date.
Use the DATE command to enter the new date.
For users without a bootstrap switch, set SR to 7600 and press LOAD
ADDR, CLEAR and CONT, assuming the core resident monitor has not been
destroyed.

B-2

3.0

DISK AS THE SYSTEM DEVICE

If your system uses an RK08, RK8E, or RF08 disk as the system device,
the programs must be loaded on the disk from the DECtape supplied.
1.

Follow steps 1-6 described in section 2.0
sys tern device.

Enter the DATE command in response
message. Then enter the command:

the

DECtape

Monitor's

DATE?

.RU SYSGEN/C
in response to the
following questions:

Monitor's

period.

SYSGEN

asks

the

DO YOU DESIRE TO CHANGE THE SYSTEM DEVICE?
Type YES and a CR key.
WHAT IS THE NEW SYSTEM DEVICE
Type the 3 character designation for the disk
(DKn) and the CR key.

system

device

DO YOU WANT TO TRANSFER YOUR FILES?
Type YES and the CR key.
DO YOU HAVE AN ANELEX PRINTER?

Type YES and the CR key and SYSGEN skips to the last
question.
Type NO and the CR key and SYSGEN assumes an LP08
or LS8E printer and asks:
132 COLUMN PRINTER ?
Answer YES and the CR key if there is a 132-colurnn line
printer.
If NO is the answer, SYSGEN assumes an 80-colurnn
printer.
IS EVERYTHING CORRECT?
Type YES or NO and the
questions are reasked.

key.

The files are transferred to the disk specified.

disk

~ne

entered

the

To operate from

the

system device, refer to BOOT, Chapter 9, or restart the

system with one of the following bootstraps.
RK08 AND RK8E BOOTSTRAP ROUTINE

Location

RK08 Instruction

RK8E Instruction

0030
0031

6733
5031

6743
5031

B-3

Set the SWITCH REGISTER to 0030, press the ADDR LOAD, CLEAR, and
keys in that order.

CONT

NOTE
I f a PDP-12

PRESET
30.

is being used, execute an I/O
8-mode before starting at location

The Monitor should respond with the version number and the
message. Enter the DATE command as described in Chapter 2.

DATE?

RF08 BOOTSTRAP ROUTINE
Location

Instruction

7750
7751
7752
7753
7754

7600
6603
6622
5352
5752

Set SR to 7750; press ADDR LOAD, CLEAR, and CONT in that order.
The
Monitor should respond with the version number and the DATE? message.
Enter the DATE corranand as described in Chapter 2.

3.1

SYSTEM RESTART ON DISK

If the system halts and CTRL/C is inoperative, use one of the
bootstrap routines to restart the system. Each time the Monitor is
restarted the Monitor requests a new date. Use the DATE command to
enter the new date.
For users without the bootstrap switch if the core resident portion of
the Monitor is intact, set SR to 7600 and press LOAD ADDR, CLEAR and
CONT.

4.0

TDBE BOoTSTRAP

If your system uses TDBE's use the following bootstrap routine to load
the COS software.
1.

Location

Instruction

7300
7301
7302
7303
7304
7305
7306
7307

1312
4312
4312
6773
5303
6777
3726
2326

B-4

7310
7311
7312
7313
7314
7315
7316
7317
7320
7321

5303
5731
2000
1300
6774
6771
5315
6776
0332
1327
"'7C An

..,~..,..,

f,J''

/O'tV

7323
7324
7325
7326
7327
7330
7331
7332

5315
2321
5712
7354
7756
7747
7400
0077

Mount the COS 300 DECtape (DEC-8E-OCOSA-D-UO) on DECtape unit
o. Set the LOCK/OFF/REMOTE switch to REMOTE and the WRITE
ENABLE/WRITE LOCK switch to WRITE LOCK.
Set the switch
register (SR) to zero (all switches down) and press EXTD ADDR
LOAD.
Set the SR to the binary equivalent of the first location and
press ADDR LOAD. Set the switches to the binary equivalent
of the first instruction and push up on the DEP switch.
Continue entering the instructions as described, the location
is incremented automatically and need not be entered again.
2.

Set terminal REM/LOC switch to REM.

Set SR to binary equivalent of the first location; press ADDR
LOAD, CLEAR and CONT keys in that order. The Monitor should
respond with the message
COS MONITOR 2.1108

(or current
version}

DATE?

WRITE ENABLE unit 0 and enter the DATE command
in Chapter 2.

If Monitor does not respond correctly start
reload the bootstrap.

step

B-5

described
1

and

5.0

PDP-12 USERS

For a PDP-12 with LINCtape,
1.

Set left switches to 0700.

Set right switches to 0000.

Press I/O PRESET with mode switch set to LINCmode.

Mount system tape
LOCKED, REMOTE.

Press DO.

When LINCtape stops moving, press the START 20 switch.

(DEC-12-0COSA-D-UO)

drive

WRITE

Monitor displays
COS MONITOR 2.1108
DATE?

(or current version)

WRITE ENABLE unit 0 and enter the DATE command
in Chapter 2.

If Monitor does not respond correctly, start at
reload the system.

step

Any LINCtapes to be used with the COS system
MARK12 {DEC-12-SE2E-UO) (P option, 129 words).

B-6

must

described

marked

and
with

APPENDIX C
SIZE OF CODE PRODUCED
BY DIBOL COMPILER

DIBOL source statements are compiled by the DIBOL compiler and produce
code which can later be executed by the run time system (RSYS) •
Memory is also set aside for the constants and variables that are used
by the program.
Each variable (field name) takes up as much memory
(in bytes)
as
specified in its data specification statement.
For example, a
variable defined by a 6D3 specification requires 18 (decimal) bytes of
storage. This is 9 words since a computer word consists of 2 bytes.
Variables defined in an overlay record share memory with the variables
in the record being overlaid.
Each RECORD statement requires one additional word of storage.
This
word is reserved at the beginning of the record for storing the COS
word count during I/O operations.
Each record begins on a word boundary
(even byte address).
If
record has an odd length (in bytes), one byte of storage is wasted.

Each constant used in the procedure section of a DIBOL program
requires storage (in bytes) equal to the length of the constant. The
length of a decimal constant is equal to the number of digits in the
number, including leading zeros.
There are no negative constants.
The length of a character string is the number of characters in it.
Each distinct constant with a length of four or less characters or
digits appears only once in the space reserved for constants. Thus,
if the constant 32 appears three times in the program it will appear
only once in the reserved data area.
However, constants larger than
four characters or digits require space each time they appear in the
program.
The code produced by statements in the procedure section must start on
a word boundary.
Each statement requires an overhead of one word.
Each statement Tabel used require's on:e word.
Unlabeled stateniehts
with line numbers 1000 more than the previous line number require one
additional word each.
The number of words of code generated
determined by the following rule:

expression

can

Add together the number of variables and constants used.

Add in the number of operators which appear
,
unary + signs) • The operators include +,

not

count

Add one for each subscript reference (be it of the form
or V(A,B)).

V(A)

(do

- /, * , #.

C-1

The following table shows how much code is used by various DIBOL
statements.
E denotes the number of words required by the expression
E.
No. of Words of
Code Generated

Statement
ACCEPT (y,x)
CALL label
CHAIN chnum
DISPLAY {line,column,expr)
END [/list control]
FINI {channel)
FORMS (channel,skipcode)
GOTO label
GOTO (labell, ••• ,labeln) ,key
IF {exprl.rel.expr2)stmnt
INCR var
INIT (channel, dev)
INIT (channel, dev,filnam[,unit])
ON ERROR label
PROC [n] [/list control]
READ (channel,record,number)

.x+~+l

RETURN

1
0

1
chnum+l
line+column+expr+l
1

channel+l
channel+skipcode+l
1
key+n+2
exprl+expr2+3
var+l
Channel+2
channel+3+filnam+unit
1
0

channel+number+record+l

START [/list control]
STOP
[NO] TRACE

1
1
2
var+l
var+expr+l*
var+exprl+expr2+1
channel+number+record+l
channel+record+2

TRAP

var=
var=expr
var=exprl,expr2
WRITE (channel,record,number)
XMIT {channel,record[,label])

For the statement marked with an asterisk (*) in the previous table,
subtract 1 if the principal operator of expr is binary + or - and if
both types are decimal. For example,
D = 3+5 takes 4 words of storage, while
D = 3*5 takes 5 words. Similarly,
D=3+4+5 takes 6 words while
D=3*(4+5) takes 7 words.
Additional space is also required by the symbol table. This consists
of two words for each distinct variable, statement label, or constant
used.

c. . 2

APPENDIX D
DESIGNING A RECORD

The file design begins with the set of related information shown in
blank form. It is a simplified form which could be processed to print
out customer invoices.
DATE
PART#
PART NAME
WHOLESALE PRICE
VENDOR#
This information is used in a BUILD control program such as described
in Chapter 6.
First, the file designer notes the type and size of
each field in this record. Numerical fields on which calculations
will be performed later by a program are defined as decimal fields;
other fields are defined as alphanumeric fields.
0010
0020
0030
0040
0050
0060

DEFINE
Fl, D6 ;PART #
F2, A30 ;PART NAME
F3, D7 ;VENDOR #
F4, D6 ;DATE
F5, D6 ;WHOLESALE PRICE

The designer now assigns BUILD input line keywords. Here, he
group related items such as part #, name and wholesale price.
0100
0110
0120
0130
0140
0150
0160
0170
0180

should

INPUT
PART
Fl
F2
F5
DATE
F4
VENDOR 1
F3

A sample data entry form is now completed, showing the line names and
the size and type of each field. Th·±s can be used for reference when
data is entered. When this record design has been proven, it will be
more efficient to print these line names on the forms which are used
to collect the original data.

DATE (D6)~---PART# (D6)
NAME(A30) _ _ _ _ _ WHOLESALE PRICE (D6) _ _ _ __
VENDOR# (D7)~-------

D-1

APPENDIX E
BUILD
CHECKDIGIT
FORMULA

In most applications involving identification nurrbers, each number may

verified

for

accuracy

checkdigit; which is essentially a

redundant digit added to the normal number.
The checkdigit is
determined by performing an arithmetic operation on the number in such
a way that the usual errors encountered in transcribing a number are
detected. The checkdigit is determined as follows:
1.

Start with a number without the checkdigit ••• 5764.

Multiply the first digit and every other digit by 2 (left
right).
10 12

Sum the digits in the resulting numbers and the digits
multiplied.
1+0+7+1+2+4 =15

not

Subtract sum from next higher number ending in zero.

Add checkdigit to the end of the number.
57645
(This is the correct checkdigit if the number is entered in a
D4 field.)

20-15=5

Note that a checkdigit procedure is not completely error proof.
In
the example given above, 5764 or 5673 give the same checkdigit.
It is
unlikely, however, that transpositions of this sort will occur.
The
checkdigit does not guard against the possible assignment of an
incorrect but valid code, such as the assignment of a wrong valid
identification code to a customer.
If the number entered for a checkdigit calculation is shorter than the
field, the rightmost digit is used as the checkdigit and the remainder
of the number is right-justified and padded with zeroes on the left.
The zeroes are considered when the checkdigit formula is calculated.

E-1

APPENDIX F
HARDWARE
DESCRIPTIONS

The hardware components for COS are:
PDP-8/E

Computer with at least 8K of core

TD8E

DECtape drives,
RF08/RS08 disk}

VTOS

Terminal

PC8-E

Paper tape Reader/Punch (optional}

CR8-E
CM8-E

Card reader (optional) or
Card reader (optional}

LP08

Line printer or

LS8-E

Centronics line printer

The above hardware
paragraphs.

components

(or TC08 drives, or RK08, RK8E, or

are

described

the

following

Figure F-1 shows one of the possible system configurations.
The central processor unit {CPU) of the COS system is located under
the desk area of the system. There are two switches necessary for the
operation of the COS system.
OFF/POWER Switch

In the OFF position the switch disconnects
all primary power to the machine. In the ON
position all manual controls are enabled and
primary computer power is applied.

INITIALIZE Switch

Push this switch to
Bootstrap Loader.

initialize

the

system

Additional information on the CPU can be obtained from Introduction to
Programming.

F-1

Figure F-1.

Possible COS Hardware Configuration

F-2

1.0

VT05 TERMINAL

Basically, the VTOS terminal (Figure F-2) consists of a
CRT display.

keyboard

and

The basic function of the keyboard is to provide a convenient, on-line
means of transmitting ASCII characters to the computer to be processed
and displayed. An ASCII code is transmitted to the computer each time
a key is depressed.
The VTOS display allows direct viewing of all displayable characters
contained in the VT05 character set that are transmitted from and/or
received by the VTOS.
The VTOS displays up to 20 lines, 72 characters per line, or a maximum
of 1440 characters full screen.
A blinking cursor indicates the
position of the next character to be generated and moves to the right
automatically as each character is displayed.
The VTOS can be
operated in LOCal or REMote mode. In LOCal mode the terminal is off
line and there is no conununication with the computer. In REMote mode
(the normal operating mode) data is transmitted between the terminal
and the computer.

VTOS controls and keys and their respective functions

are listed in
Tables F-1 and F-2. Keys which are operative off-line only are not
covered here; refer to the hardware manual for the VTOS for a
description of their function.

F-3

TABLE F-1.

"i.lT05 CONTROLS

Control

Function

Power ON/OFF Switch

This switch is located on the right front of
the VTOS.
When the ON/OFF switch is turned
ON, power is applied to the complete system
and
the VTOS display refresh memory is
cleared. After approximately one minute has
elapsed,
the cursor (bright dash should
appear in the upper left-hand corner (HO.ME
position) of the screen.

REM/LOC Switch

This switch is located on the right front of
the VTOS.
In LOCal mode, the terminal is
off-line and data can be typed at
the
keyboard but not entered into memory. In
REMote mode, keyboard data is transmitted
from the VTOS to the processor.
NOTE
The CONTRAST, BRIGHTNESS, VERTICAL
and
HORIZONTAL
controls
are
adjusted in the same way as those
of
a
commercial
television
receiver.

CONTRAST Control

The CONTRAST control is located on
the
right-hand side of the VTOS and is used to
adjust for display contrast (clarity).

BRIGHTNESS Control

The BRIGHTNESS control is located on the
right-hand side of the VT05 and is used to
adjust for
CRT
display
brightness
or
intensity.
NOTE
To correctly
adjust
the
VTOS
character presentation, turn the
CONTRAST control counterclockwise
to
minimum,
then
adjust
the
BRI.GHTNESS control decreasing the
intensity
until
the
raster.
brightness is barely intensified
(just above the CRT cutoff point).
As a final
step,
adjust
the
CONTRAST control to the desired
level,
according
to
ambient
lighting conditions.

VERTICAL Sync Control

The VERTICAL sync control is located on the
right-hand side of the VT05 and is used to
properly synchronize the screen
in
the
vertical direction.

HORIZONTAL Sync Control

The HORIZONTAL sync control is located on the
right-hand side of the VTOS and is used to
adjust the
screen
picture
for
proper
synchronization in the horizontal direction.

F-5

TABLE F-2.

VTOS KEYS

Key

Function

Returns cursor to the left margin of screen.

Moves cursor down one line in same location.
If pressed when cursor is on the bottom line
(line 20), the display scrolls (top line
disappears and rest of lines move up one line
position).

RUBOUT

Works in REMote mode in conjunction with the
Monitor to erase characters.
Erases the
character above the cursor from memory and
echoes the deleted character on the screen,
i.e.,

TAB

Moves the cursor to the right to the next TAB
stop each time TAB is pressed.

ALT

Has no effect on the display and is ignored
by the terminal. It is included to provide
an alternate escape
character
(replaces
Carriage Return or Line Feed) for use in the
user program.

CTRL

Used simultaneously with character keys
perform special functions as follows:

CTRL/C

Causes an immediate return to the Monitor.

CTRL/N

Enables and disables imbedded numeric keypad.
See Table 2-1 in Chapter 2 for details.

CTRL/O

Suppresses terminal output.
If output is
already suppressed, enables terminal output.

CTRL/U

Erases the current input line.

CTRL/V

Suppresses Monitor version number message.
If output is already suppressed, enables
output of message.

CTRL/Z

Indicates no more input. When used in answer
to an ENTER message it forces the program to
use a default value.
When specified while
editing, it signifies end of paper tape or
card deck.

F-6

TABLE

'[;I_')

VTOS

..... -'-.

KEYS (Con' t)

Key

Function

SHIFT

Produces the character in upper case portion
of the key pressed simultaneously with SHIFT,
e.g., ! rather than 1.
Control characters
are "SHIFT inhibited" and alpha characters
are permanently set to upper case.

SHIFT LOCK

Enables

and

holds

the

shift function or
When enabled the

releases it if already set.
SHIFT LOCK key is lit.
Space

Produces a blank character position (wherever
the cursor is located) each time it is
pressed. Cursor moves one position to the
right.

Additional features:
The VT05 sou__nds a warning beeo when the line reaches character
position 65.
This alerts the user that there are 8 positions before
the end of the line. The cursor moves to the right automatically as
each character is displayed until it reaches position 72.

1.1

VTOS Start Up Procedures
1.

Press the OFF portion of the ON/OFF power switch.

For local
alignment
switch.

operation
purposes)

(used
press

For on-line operation,
REM/LOC switch.

for training, maintenance and
the LOCal portion of the REM/LOC

press

the

REMote

portion

the

Press the ON portion of the ON/OFF power switch.
The
blinking cursor
(bright dash) appears in the HOME position
(the first character position of the top line), and a beep is
emitted.

Allow approximately one minute for the CRT filament warmup.

If the cursor does not appear as specified,
key.

press

the

HOME

If the cursor still does not appear, check the BRIGHTNESS
control to be sure it is not set too low. Contact the local
DEC Field Service Office if these steps do not make the
cursor appear.
6.

In REMote mode all keyboard and control functions are
operational and the VTOS is ready for on-line operation. To
verify correct system operation, press the various controls
and displayable character keys individually and observe the
display.
In LOCal mode
all
keyboard
functions
are
operational. To verify this, press the keys individually and
observe the screen.

F-7

2.0

HIGH-SPEED PAPER TAPE READER AND PUNCH UNIT

The high-speed paper tape reader and punch unit shown in Figure F-3
performs paper tape input and output.
The high-speed paper tape
reader is used to input data into core memory from eight-channel,
fan-folded
(non-oiled), perforated paper tape.
The reader inputs
information photoelectrically at a rate of 300 characters per second.

Figure F-3.

High-Speed Paper Tape Unit

F-8

The reader controls are shown in Figure F-3
F-3.
TABLE F-3.

and

described

Table

HIGH-SPEED READER CONTROLS

Control

Function

ON/OFF

Press down on the ON portion of this switch
to supply power to the reader unit. The OFF
position setting disconnects the reader power
supply,,

FEED

Press the right side of this switch to
advance
tape through the reader without
recording by the photoelectric sensors.

Control knob

Turn this knob to the left to raise the tape
retainer lever so paper tape can be inserted
or removed. Turn the knob to the right to
lower the lever to hold the paper tape over
the sprocket wheel.

The reader is controlled by the computer, although the operator may
indirectly control the reader from the keyboard through the computer.
Paper tapes are manually positioned in the high-speed reader with
following steps.

the

Turn the control knob to raise the tape retainer lever.

Place the paper tape in the right-hand bin so
of the tape passes over the sensors first.

Place several folds of leader tape in the left-hand bin
the tape passing over the sprocket wheel.

Turn the control knob to lower the retainer lever over the
tape so the feed holes are engaged in the teeth of the
sprocket wheel.

Press the tape FEED button until
reader head.

6"'.

Tape·"

·±g

m:t~"'a:nced

am!

read

the

leader

tape

progra.r:nned

beginning

over

with

the

COihpUter

instructions.
Once the paper tape has been properly placed in the reader and the
leader/trailer has been positioned as outlined in the preceding steps,
the tape is normally read under control of system software.
The high-speed paper tape punch is used to record computer output on
eight-channel, fan-folded paper tape at 50 characters per second. All
characters are punched under program control from the computer.
The punch controls are shown in Figure F-3 and are described in
F-4.

F-9

Table

TABLE F-4.

HIGH-SPEED PUNCH CONTROLS

Control

3. 0

Function

ON/OFF

Press down on the ON portion of this switch
to supply power to the punch unit. The OFF
position setting disconnects the punch power
supply.

FEED

Press the right side of this switch to
punched tape for
advance
feed-hole-only
leader/trailer purposes.

DECTAPE TRANSPORT UNIT

The DECtape transport unit shown in Figure F-4 is a bidirectional
magnetic tape transport which reads and writes the 10-channel magnetic
tape. Tape movement can be controlled by programmed instructions from
the computer or by the manual operation of switches located on the
front panel of the transport. Data is transferred only under program
control.
The transport controls are described in Table F-5.
TABLE F-5.

DECTAPE CONTROLS

Control

Function

REMOTE

This switch position energizes the DECtape
transport
and
places
it under program
control. The REMOTE SELECT light comes on
whenever the unit is selected for use.

OFF

This switch
transport.

LOCAL

This switch position energizes the DECtape
transport
and
places it under operator
control from external transport switches.

WRITE ENABLE

This switch position enables the DECtape
search, read, and write activities.
WRITE light comes on when the switch is
in the WRITE ENABLE position.

WRITE LOCK

This switch position limits the
DECtape
transport to search and read activities only.
(This prevents accidental destruction
of
permanent data.)

F-10

position

disables

the

DECtape

for
The
put

TABLE F-5.

DECTAPE CONTROLS (Con't)

Control

Function

Unit Selector

The value specified by
dial (0-7} identifies
control unit.

this eight-position
the transport to the

On TDSE devices, the first controller is
hardwired to accept only O or 1 and the
second controller only 2 or 3.

With the transport in LOCAL mode, depressing
this switch causes tape to feed onto the
right-hand spool.
With the transport in LOCAL mode, depressing
this switch causes tape to feed onto the
left-hand spool.

DECtapes are mounted on the transport unit as follows:
1.

Set REMOTE/OFF/LOCAL switch to OFF.

Place DECtape on left spindle with DECtape label out.

Wind four turns on empty DECtape reel on right spindle.

Set switch to LOCAL.

Depress the
switch for a few seconds to make sure tape
is properly mounted.

Dial correct unit number on unit selector.

Set switch to REMOTE.
LOCK setting.

Select either WRITE

F-11

ENABLE

WRITE

Figure F-4.

DECtape Transport Unit

F-12

4•0

CARD READER

There are two types of card readers, the CR8-E which reads 12 - row,
80 column punched cards and the CM8-E which reads 12 - row, 40 column
mark sense cards or 12 - row, 40 column punched data cards with timing
marks.
The CRB-E card reader switches and indicators are described
F-6.
TABLE F-6.

Table

CARD READER SWITCHES lu~D INDICATORS

Switch/Indicator

Function

POWER (toggle circuit breaker and indicator)
All power to the
card
controlled by this switch.

reader

STOP (momentary-action pushbutton/indicator switch)
Actuation of the STOP switch immediately
overrides the PICK COMMAND and negates
the READY status. The card reader will
stop operation after the card currently
in the track is read completely.
Power
is not removed from the reader by this
action.
The red STOP indicator
is
illuminated as soon as the switch is
depressed.

RESET (momentary=action pushbutton/indicator switch)
Actuation of the RESET switch clears all
error
logic
and
initializes
all
counters.
Read Check (indicator)

The READ CHECK alarm indicator denotes
that the card just read may be torn on
the leading or trailing edges or have
punches in the 0 o.r 8 ls t col urnns • A
READ CHECK error will cause the reader
to stop.

Stack Check (indicator)

The STACK CHECK alarm indicator signals
that a card has failed to reach the read
station af±er .a PICK COMMAND has been
received.

Hopper Check (indicator)

The HOPPER CHECK alarm indicator denotes
that either the input hopper is empty or
the stacker is full.

F-13

Figure F-5.

Card Reader

F-14

5.0

LINE PRINTERS

5.1

LPOB Line Printer

The switches and lights described in Table F-7 are part
line printer, Figure F-6.

the

LP08

ON LINE/OFF LINE (switch)

Push up to put printer on line with
computer.

the

PAPER STEP (switch)

Push up to advance paper
disabled in on-line mode.

TOP OF FORM (switch)

Push up to advance paper to top-of-form:
disabled in on-line mode.

ON LINE (indicator)

Lights when ON LINE/OFF LINE
set to ON LINE.

READY {indicator)

Lights when printer is ready to
computer output.

T~.BLE

F-7.

LP08 SWITCHES

Switch/Indicator

~.ND

INDICATOR LIGHTS
Functions

one

switch

The ON/OFF switch controls the power to the printer.
There are two LP08 line printers:

80- and 132-column versions.

F-15

line;

receive

Figure F-6.

LP08 80-Column Line Printer

F-16

5.2

LS8-E Line Printer

The basic character set for the LS8-E (Figure F-7) contains the 10
numeric digits, 0 through 9; 26 upper case letters, A through Z; and
the 28 special characters shown below:

[
j
?

(
)

\
A

(space)

The LS8-E controls and their functions are shown in Table F-8.
TABLE F-8.

LS8-E CONTROLS

Control
ON/OFF (switch)

Function
Provides power to the printer and lights

when in t..~e ON position.
SELECT (switch)

Enables and halts printer and lights
the on position.

TOP OF FORM (switch)

Moves line printer paper to top of sheet
or proper location of the vertical forms
control tape
{Refer
to
Centronics
Manual} when installing forms.

FORMS OVERRIDE (switch)

Allows completion of form being printed
even
though
paper
out switch was
activated.

HARDWARE ALARM {light)

Signals whe.n head carrier has exceeded
the right margin stop. An audio alarm
also sounds.
If this should happen,
reset the printer and restart the job.

PAPER OUT (light)

Indicates printer is out of paper or
paper handling malfunction. Correct the
problem and restart the job.

Manual Controls
Paper Advance Knob

Located on left side
of
carriage;
provides
manual forward and reverse
paper positioning.

F-17

TABLE F-8.

LS8-E CONTROLS (Con't)

Control

Function

Paper Feed Level

Located on rear, right top of printer~
provides
initial
paper feed around
platen to allow positioning of paper in
the tractor feed mechanism.

Forms Thickness Control

Located on inside of printer, this dial
with
graduated
markings
provides
adjustment of clearance between platen
and face of print head. Adjustment must
be made according to the thickness of
the forms being used.

F-18

Figure F-7.

LS8E Printer

F-19

6.0

RK08 DISK

Figure F-8.

RK08 Disk

The switches, buttons and lights described in Table F-9 control the
operation of the RK08 Disk.
(See Figure F-8.) An RK8E disk is also
available.

F-20

TABLE F-9..

RKO 8 CONTROLS AND INDICATORS

Control

Function

ON (pushbutton and indicator)
Push ON button to supply power to the
disk.
ON button lights when power is
on.
OFF {pushbutton and indicator)
Push OFF button to halt the power supply
to

the

disk.

OFF

button lights when

power is off.
ATTN (pushbutton and indicator)
Indicates that the disk power supply has
a malfunction. Push the STOP switch to
reset the power supply.
NOTE
This button also lights when
system power is turned off
with
the
computer
ON/OFF
switch.
Push this button to
reset the power supply.
This
inter-locking action protects
the data on the disk from
spurious write signais while
computer and disk controls are
powered down.
DISK LOCKOUT 0-3 {pushbutton and indicator)
Press the appropriate button{s) (0-3) to
lock out the corresponding disk so data
cannot be written onto the disk platter.
The button lights when pressed. The
panel controls up to four disks.
SECT PROT

Prevents writing on protected sectors,
as specified by bit 0 of the second
header word of the disk sectors.
Press
to
protect
sectors
of
the
disk
previously coded for sector protect. by
settihg· bit u of the sectors 2nd header
word to 1.
(not implemented in COS}

START/STOP (switch)

Place in the START or up position to
initiate the disk starting cycle (only
if the disk cartridge is in place and
the disk cartridge receiver handle is in
the raised position). Place in the STOP
or down position, to initiate the disk
STOP sequence. Switch must be in the
STOP position to mount or dismount the
disk cartridge.

TABLE F-9.

RK08 CONTROLS AND INDICATORS (Con't)

Control

Function

SAFE (light)

The SAFE light comes on approximately 35
seconds after a STOP sequence has been
initiated. When SAFE is lit, the disk
cartridge access door interlock is off,
and it is safe to lower the
disk
cartridge
receiver
handle
for the
removal or insertion
of
the
disk
cartridge.

0,1,2 or 3
{unit #} (light)

The READY light (left-hand light) comes
on approximately 90 seconds after the
START sequence has been initiated. When
READY is lit the disk is ready to begin
communicating with the computer.

6.1

Mounting and Dismounting the RK08 Disk Cartridge

6.1.1

Mounting

Follow the steps below to initiate operation of the RK08 disk.
The DC power must always be ON before inserting a disk cartridge into
the disk.
If the disk cartridge is inserted without first turning the
DC power on, the read/write head carriage assembly may be in the way.
1.

Press the disk power supply DC ON switch.
on.

The ON light comes

When the SAFE lamp, on the front control panel of the disk,
comes on and the handle lock unlatches, the disk is ready to
receive the disk cartridge.

Pull open the. disk cartridge access door panel.
The
cartridge receiver is now raised into position for insertion
of the disk cartridge. The handle interlock will keep the
cartridge receiver handle locked in its closed position
whenever the SAFE lamp is turned off.

Slowly slide the disk cartridge into the receiver assembly
and make certain there is no internal resistance to its
insertion.
NOTE
If there is the slightest resistance to
the
insertion of the disk cartridge into the disk
receiver, the cartridge should be removed and the
interior of the disk receiver examined to determine
the cause of resistance.

F-22

Raise the disk cartridge handle to the up or loaded position.
Do not force the handle.

Close the disk cartridge access door.
NOTE
An interlock switch is activated when the disk
cartridge is properly inserted. However, if the
switch is not activated, the drive is not put in
the ready condition.
If the switch should transfer
to its inoperative position during an otherwise
normal operation, the disk will immediately stop
and the SAFE light will light.

Set the START/STOP switch to the START or up position to put
the
disk drive system into operation.
A disk handle
interlock activates to ensure that the handle cannot be
raised during normal operation.

After a 90 second delay the READY lamp comes
that it is ready for actual operation.

6.1.2

on,

indicating

Dismounting

Follow these steps to remove the disk cartridge.

6.2

Make certain the DC power is ON.

Set the START/STOP switch to the STOP or down position.
disk decelerates and the SAFE light comes on.

Open the disk cartridge access door and
receiver handle down.

pull

the

cartridge

Remove the
assembly.

cartridge

receiver

Close the disk cartridge access door.

disk

cartridge

from

the

The

Mounting and Dismounting the RK8E Disk Cartridge

6.2.1

Mounting

The following is the correct cartridge mounting procedure for the RK05
disk drive.
Any deviation encountered during this procedure will be
considered an error condition.
1.

Set switch labeled RUN/LOAD to the LOAD position.

Verify that light labeled PWR is on.

F-23

Wait for light labeled LOAD to come on.

Verify that lights labeled ROY, ON CYL, FAULT, WT, and RD are
off.

Open access door.

Insert cartridge.

Close access door.

Set switch labeled RUN/LOAD to the RUN position.

Wait for lights labeled ROY and ON CYL to come on.

10.

Press switch labeled WR PROT and
labeled WT PROT goes on and off.

11.

Press switch labeled WT PROT until light labeled WT PROT goes
off.

12.

Verify that lights labeled FAULT, WT, and LOAD are off.

6.2.2

verify

that

the

light

Dismounting
1.

Press switch labeled WT PROT and wait until light labelled WT
PROT goes on.

Set switch labeled RUN/LOAD to the LOAD position.

Wait for light labeled LOAD to come on.

Open access door and remove cartridge.

Close access door.

F-24

APPENDIX G
COS FILES

There are four types of files in the COS system: source, compiler
binary, system program, and data. Source, compiler binary and data
files have similar structure.
The fourth, system programs, uses
standard OS/8 SAVE format.

1.0

COS SOURCE FILES

All source files, whether control programs (e.g., for BUILD) or DIBOL
program sources must be input with the LN command or number commands.
This makes all source files look the same and makes them compatible
with cos. The source file format is:
line
number

word
count n

n-1 words, two -237 ASCII
characters per word

The first word contains the word count for that line.
as

It is

computed

n = (# of characters on line+l)
+l
2

The second word is the text line number, 0000-7777, base 8 (0000-4095,
decimal}.
The third and successive words contain stripped -237 ASCII
packed two per word (e.g., AB is 4243).

2.0

characters

COS DATA FILES

'l'he format of data files is exactly the same as
there are no line numbers.

source

files

except

A data file looks like:
word
count n

n words, two characters
per word

Word 1 of a data record contains its length in words.
The next and
successive words contain the text of the record, two stripped ASCII
characters per word.

3.0

COS DIBOL COMPILER BINARY

Although the interpretation of the contents is entirely different,
externally a file of compiler binary is structured exactly like a data
file.
That is, for each line of DIBOL source
program,
the
interpretive code for that line is stored as a word count followed by
the interpretive code to be used by RSYS.

4.0

COS SYSTEM PROGRAMS

All system programs {e.g., PIP, COMP, BUILD, etc.) are stored in OS/8
SAVE format.
The first block of the file is a core control block
indicating where the rest of the blocks of the file are to be loaded.
Each successive block is a 256 word core image. See the OS/8 Software
Support Manual for details.

5.0

SYSTEM AND DATA TAPE FORMATS

COS puts a label on all tapes and disks.
This label occupies the
first 256 words of each device, of which four are the actual label,
one is the date, and the rest may be a bootstrap. Files on DECtape do
not use the last two blocks of storage.
Data files are completely devoted to the storage of data, beginning at
block 1. Each logical unit holds one data file only. Labels on data
files are put on by the Monitor in conjunction with DIBOL programs or
by system programs such as SORT and BUILD.
System files have the COS Monitor, the system programs such as COMP,
and a directory in OS/8 format.
(There is a discrepancy of two years
between directories printed by OS/8 and those printed by COS.)
System
files can also be used to save the sources of DIBOL programs, BUILD or
SORT control specifications, or compiled DIBOL programs, but no data
files. Their bootstrap starts up the COS Monitor.
Figure G-1 illustrates the layout of the Monitor portion of the system
device.
As noted in the figure, COMP should be the first file in the
file area. The location of COMP is particularly relevant when the
binary scratch area is to be expanded.

G-2

BLOCK #
(OCTAL)
BOOTSTRAP

t
DIRECTORY

6000-7777
RESIDENT MONITOR

J_
10

13
14

EDITOR OVERLAY

17
20

0-5777
EDITOR

"34"""
RSYS LOADER

"Sr

40
SOURCE SCRATCH

60
RSYS

67
70
COMPILER
OVERLAYS

77
100
BINARY SCRATCH

AREA

FILES
(COMP SHOULD
BE FIRST)

Figure G-1.

Monitor Organization

G-3

The label on a system tape is unique and identifies it as a system
tape.
It is put on all system tapes automatically, and looks like
[SYS]. The index of a system file looks like
Word

Contents

(-) the number of dire~tory entries in this block.

The starting block number for file storage.

Link word to next directory block, or 0 if end.
There are seven directory blocks on all multi-file
devices.

0 (unused).

{-)
the number of
(always equals -1).

First two characters of name.

Next two characters of name.

Last two characters of name.

Two character extension.

Date.

Length of file {negative).

11 through 255

auxiliary

words

per

entry

Repeat of 5-10 for each file.

Pseudo-DECtapes are reserved areas of disk, set aside by the SYSGEN
program.
They are allocated beginning at the first free block
following any COS system on the disk and going to the end. The system
index, however, knows nothing about these pseudo-DECtapes inhabiting
the upper area of the disk. All it knows is that the available space
for file storage only extends to block n, where n is something less
than a full disk.
Access
Data tapes and pseudo-DECtapes (which may only be used as data tapes)
will be referenced by their logical unit numbers as assigned by
SYSGEN. What SYSGEN will actually do is set up a table like the
following:
Handler address
starting segment
length

/physical device #

handler address
starting segment
length

/device #
/parameters

G-4

The handler address is a pointer to the device handler to use for this
logical unit (either Disk or DECtape). The physical device number is
which disk or DECtape drive to reference. The starting segment gives
a 16-bit starting block number, which is where on the physical device
the space allocated for this logical unit begins.
This allows for
possible future expansion to larger disks. The length is the number
of segments reserved for this logical unit.
For example, if logical unit 14 is assigned to 32-block area on
the fourteenth series of entries in the table might look like:

DKl,

RKOB handler address + l
212
-40

/start at segment 212(8)
/go for 40(8) segments

So, any references to logical unit 14 would refer to
blocks
5240(8)-6237(8)
of DKl. Note that block 5240 would have a label for
that logical unit, just as any other data tape.
For DECtapes, one of these table entries would be essentially the
o:>GUm:o,
except that the starting block would be 0 a..11d probably extend
for 46 segments (one complete tape). Two logical units should not be
assigned to the same DECtape, as it causes an egregious number of
problems when two logical units occupy a dismountable device and a
file extends to multiple units.

G-5

APPENDIX H
DIBOL
DEBUGGING
TECHNIQUE
(DDT)

In addition to the TRACE feature described in Chapter l, more complex
program debugging techniques may be implemented by specifying the /D
option when executing a source program (see Chapter 4) or when running
a binary program (see RUN command, Chapter 2). The features of the /D
option include: breakpoint, variable examination, subroutine call
traceback, and iteration.

1.0

ENTERING DDT MODE

The command to run a binary program in DDT mode has the form
~RU

[pronam+chainl+ ••• +chainn,filnaml, ••• ,filnarnn]/D

There is an additional amount of space required because of the /D
option.
It is 768 words plus 3 words for each label in the program
that appears before the PROC statement.
When the program is loaded with /D by either of the above commands,
initial control is passed to DDT which prints an appropriate DDT
version number followed by a "-" to indicate DDT command mode. DDT is
now ready to accept commands.
Whenever a CHAIN is entered, control is passed to DDT.

2. 0

DDT COMMANDS
Command

Explanation

varnam=

The current contents of variable varnarn is
printed.
Varnarn may have single or double
subs"°ripts •

varnam=v

Set varnam equal to the value v, where
any legal alphanumeric string.

If v has more characters than defined for
varnam, the error message ERROR IN COMMAND is
printed.
Prints the
examined.
=v

contents

the

last

variable

Sets the contents
examined to v.

the

last

variable

H-1

Command

Explanation

$nnnn

Sets a breakpoint at line number nnnn.
Only
one breakpoint may be active at a given time.
Do not set a breakpoint at line o.
The
breakpoint occurs prior to executing line
nnnn.

The breakpoint is executed
occurrence of line nnnn.

the

nth

For example:
. -$300

->4
When the program starts to execute line 300
for
the fourth time, the breakpoint is
executed and control is transferred to DDT.
CTRL/Z

Starts execution of DIBOL program.
If a
breakpoint
was set in a $nnnn command,
control reverts to DDT when nnnn is reached
and the following message is printed:
BREAK!
The user may now type additional commands
response to the "-".

The lines from which
CALLs
were
made
(push-down list) during execution of the
DIBOL program is printed.
This command is
generally used after a breakpoint or system
error has occurred to trace the execution of
the program. This command is a caret (A) on
the VTOS keyboard.
NOTES
1.

If a DIBOL program running under DDT causes an
message such as ILLEGAL SUBSCRIPT or NUMBER
TOO LONG, while a breakpoint was pending, control
will be transferred to DDT and the DDT commands may
then be used for program examination.
If the error
was fatal, the DIBOL program cannot be restarted by
the CTRL/Z command.

error

Once a DIBOL program is running under DDT, DDT
cannot be restarted unless a breakpoint occurs, or
an error occurs
with
a
brakpoint
pending.
Therefore, if the user has no requirement for a
breakpoint but wishes to return to DDT for program
examination if an error occurs, it is necessary to
set a breakpoint to a non-existent line number.

H-2

APPENDIX I
DIBOL
STATEMENT
SUMMARY

Statement

Explanation

ACCEPT (terminator,receiverj

Accepts input from KBD stores it in
receiver
and
stores
terminating
character
as
terminal
code
in
terminator.

BLOCK

See RECORD.

CALL label

Causes control to branch
whose label is specified.

CHAIN decimal expression

Causes the specified CHAIN to be loaded.
Execution resumes with the first PROC
statement in the specified CHAIN.
See
Sections
1. 4. 3
and 2. 2. 6 for more
detail.

DISPLAY (line,colurnn,expr)

Display expression on screen at line and
column
specified,
(0,0
current
location) er perform one of following
operations.
0 = position cursor, 1 =
clear to end of scope, 2 = clear to end
of line, 25 = emit a bell or beep sound.
No carriage
generated.

line

statement

feed

END [/list-control]

An optional
statement
inserted
physical end of DIBOL program.

FINI (channel)

Closes file associated with channel (if
output file, writes end of file, file
length). Frees channel.

FORMS (channel,skip-code)

Line printer control statement. If code
is 0 go to top of next form: if 1-4095
skip number of lines specified.
-1 =
skip to channel 2 and advance paper n
lines as specified on vertical forms
control
tape
(LS8E).
-2 = print
enlarged characters
for
next
XMIT
statement (LS8E).

GO TO label

Unconditional branch to label.

I-1

return,

Statement

Explanation

GO TO (labell, ••• ,labeln),key
If value of key is K, with K greater
than 0 and less than or equal to n,
branch to labelK.
Otherwise execute
next instruction.
IF (expressionl.rel.expression2)stmnt
Execute stmnt if relationship is
• rel. is one of the f cl lowing:
.GT.
.LT.
.LE.
.GE.
.EQ.
.NE.

true

greater than
less than
less than or equal
= greater than or equal
equal
= not equal

stmnt may be one of:
GO TO label
CALL label
STOP
RETURN
ON ERROR label
[NO] TRACE
!NCR variable

Increments value of variable by 1.

!NIT (channel, device [,filnam] [,unit])
Associate channel with specified device
and initialize device. For mass storage
devices only, specify file name.
If
unit is specified, system looks for file
on this logical unit before issuing
MOUNT message.
The device is one of the following:
INPUT

mass storage input data file
mass storage output data file
UPDATE mass storage data file
(fixed
length records) for update
KBD
keyboard input
TTY
terminal output
LPT
line printer output
CDR
card reader input
PTP
paper tape punch (high speed)
RDR
paper tape reader (high speed)
SYS
input source file
(on system
device)
OUT

[NO] TRACE

Stops/starts trace facility.

ON ERROR label

Branch to label if next statement incurs
a non-fatal error.

I-2

Statement

Explanation

PROC[n] [/list control] [;comments]
Begins PROCcedure section and n= maximum
number of mass storage logical units to
be opened at one time.
If omitted, n=7
is assumed.
Causes top of form on
compiler listing.
The comments
are
printed on each page as headings.
READ (channel,record,rec#)

Does a direct access read of data record
into an area in core.

RECORD [rname] [ ,X]
{'alpha initial value'

[fldnam),[x)

A) [n)t[decimal initial
(

""\ { ' P }

value~

Reserves areas of core where records are
stored during processing and defines the

fields
rname

fldnarn

x
n

,c
RETURN

• •1..1.

the record •
labels record area.
overlay of
another
record
area.
record must be less
than or equal to the record
being overlayed.
names a field
within
the
record
repetition count (1-512)
for
an array
size
of
field
(1-510
characters) default = 1
insert
initial
value
at
execution
insert
system
date
at
execution
assigns fldnam to the value of
the
run time options
(see
Section 1. 4 .16).
clears record loaded as
a
result of a CHAIN statement
(see Section, 1 .. 4.16).

Returns control to next statement
last CALL.

after

START [/list control] [;comments]

An optional statement which
anywhere.
Causes
compiler listing.
STOP

top

may appear
of form on

Terminates program execution and returns
control to Monitor.

Statement

Explanation

TRAP label

Allows a DIBOL program to spool output
to the line printer while executing the
program.

WRITE {channel,record,rec#)

label is the name of the line
routine (see Section 1.4.21).

printer

Does a direct access write
record to a specified file.

data

XMIT {channel,record[,eof label])
Transfers data between record area and
device attached to channel.
If end of
file occurs for an input file, branch to
label if specified.

Data manipulation statements:
variable=

clears variable to zeroes or blanks

variable=expression

assignment,
types.

variable=expressionl,

variable

where:

alpha literal
alpha variable
formatting.

expression

D
A

A
A

conversion

depends

Types must be:

upon

A=D,A.

action
Assign. If variable is too small, drops
high order digits.
If variable is too
big, right justified.
Alpha converted to decimal then stored.
Assign. If variable is too small, drops
characters on right.
If too big only
left-most characters are changed.
Decimal converted to alpha.
Treated as A to A.

D = decimal variable
A = alpha variable
R = record

I-4

APPENDIX J
ERROR
MESSAGE
SUMMARY

Message

Program

Explanation
Attempt

assign

than

units.
BAD ALPHA FIELD

BUILD

Illegal entry in alpha field.

BAD ALPHA VALUE

BUILD
COMP

Initial value in an alpha data
specification did not begin or
with a single quotation.

BAD CHAR

CONVEX

Attempted to
convert
an
OS/8
character for which there is no COS
code.

BAD CHECKDIGIT

BUILD

Checkdigit calculated by BUILD does
not match the one entered.

BAD CHECKSUM

PATCH

An attempt was
block
which
patched.

BAD COMPILATION

Monitor

User tried to SAVE a
binary that had errors.

BAD DECIMAL FIELD

BUILD

Illegal entry in decimal field.

BAD DECIMAL VALUE

COMP

The initial value for a decimal
data specification was incorrectly
formed.

BAD DEFAULT FIELD

BUILD

Default field does not agree
size or type of object field.

BAD DELIM

BUILD

Fields are separated by an
delimiter.

BAD DEPICTOR

BUILD

OUTPUT format in error.

BAD DEV

CONVEX

Output
device
not
same
type
physical device as system device.

BAD DEVICE

BUILD

Not KBD, SYS, RDR or CDR.

UPDATE

Illegal device specified
statement.

J-1

made
was

end

to write a
incorrectly
compiled

with

illegal

INPUT

Message
BAD DIGIT AT LINE nnnn

Program

Explanation

Run Time

A character other than +, -, space,
or the digits 0-9 was encountered
in an alpha to decimal conversion:
nnnn is the line number in which
the error occurred.

BAD DIGIT IN DECIMAL INITIAL VALUE
BUI~D

SORT
UPDATE

Alpha character in a decimal
initial value.

BAD DIRECTORY

PIP
Monitor
PATCH
CONVEX

Attempt to reference or store a
file on a device with no directory
(or a directory that has been
destroyed). Also an attempt to get
a directory of a data file.

BAD END STATEMENT

BUILD

Missing
or
statement.

BAD FIELD NUMBER

BUILD

Field number specified
than 2047.

BAD INPUT STATEMENT

BUILD
UPDATE

Statement missing, misspelled, or
out of order.

BAD KEY STMNT

UPDATE

Key in
syntax.

BAD KEYWORD

BUILD

Keyword entered was not
in the control program.

described

BAD LABEL

Monitor

Tape has no label, or its
incorrect.

form

BAD NUMBER

PATCH

A number with more than 4 digits, a
non-digit, or 8 or 9 was typed.

BAD OPTI_ON

BUILD

Something otl'l_er :than _C or
specified in OUTPUT section.

BAD-- OUTPUT FORMAT :ft

BUILD

Format number is wrong
range ( < O or 7 ).

BAD PROC #

KEY

misspelled

END

greater

has

statement

bad

out

COMP

The number in a PROC statement
not a digit from 0 to 7.

was

Run Time

Attempting to run a binary program
which contains a compilation error.
Check compilation listing for error
flags. Correct flagged errors, and
recompile.

BAD PROGRAM AT LINE nnnn

J-2

Message
BAD RA..~GE CONSTRUCTION

Program

Explanation

BUILD

In Rlil~GE portion of statement an
open or close parenthesis or comma
is missing.

RECORD SIZE

SORT

File contains records
length.

BAD RELATIONAL

CO.MP

An illegal relational occurs in an
IF statement. For example, a .GX.

BAD

variable

instead

BAD SWITCH

SYS GEN

Not /T, /C, or /L.

BAD UPDATE COMMAND

UPDATE

Bad syntax, probably missing F,
or comma.

BAD UPDATE STMNT

UPDATE

UPDATE statement is incorrect.

UPDATE

Value in update command too long or
incorrect.

BAD WORK UNIT COUNT

SORT

# of work units not in range 3-7.

BLOCK TOO BIG

PATCH

BAD

VALUE

BLOCK NUMBER ERROR PHASE n
TDMARK
C-FIELD NOT PERMITTED

BUILD

CANT BACKSPACE PAST BEGIN OF FILE
DAFT

incorrect block
number
was
typed.
It cannot be longer than
file
being
the length of the
patched.
Refer to DEC-8E-EUZC-D.

C-type fields cannot be
in the INPUT section.

specified

Attempt was made to backspace
beginning of file.

CANT BACKSPACE WITH SEQUENTIAL INPUT
DAFT
Attempted
to
backspace
sequential input.

CCP ERROR

COMP

Matching angled bracket
missing.

CHECKSUM ERROR PHASE n

TDMARK

Refer to DEC-8E-EUZC-D.

COMMA MISSING

COMP

No comma appeared
expected.

CONTROL BUSY ERROR

RKSMRK

Disk IOT issued
busy.

J-3

with

( < or > )

one

was

control

was

where

when

past

Message

Program

DATA INITIALIZATION MISSING
COMP

Explanation
No data initialization followed a
comma
in
a data specification
statement.

DATA ERROR PHASE n

TDMARK

Refer to DEC-8E-EUZC-D.

DATA RATE ERROR

RK8MRK

The processor was busy and did not
respond to a data break request
within the 13 us required.
The
transfer is terminated immediately.

DATA TABLE OVERFLOW!

BUILD
UPDATE
SORT

Too much data.
characters.

DESCRIPTOR TABLE OVERFLOW!
BUILD
UPDATE
SORT
DESTINATION FIELD NOT DECIMAL
BUILD

Maximum = 960

Too many F, T and C fields defined.
Maximum= 160.

The destination field specified
not defined as decimal.

DIBOL FILE NUMBER IN USE AT LINE nnnn
Run Time In !NIT, the channel number is
already INITed to a mass storage
device.
DIBOL FILE NUMBER NOT INITED AT LINE nnnn
Run Time An attempt was made to XMIT, READ,
or WRITE with a channel number that
was not INITed.
DISK DATA ERROR

RKEMRK

Refer to DEC-08-DHRKD-A-D.

DT n?

Monitor

The expected tape is not
on a DECtape drive.

available

For example, if following a MOUNT
message the operator has specified
that a file should be written on
unit 3, and unit 3 is not selected,
or is WRITE LOCKed, this message is
output.
On LINCtapes there is no message.
To recover, set switch to ENABLE
the unit or set the dial selector
to the proper number.
Then type
any key to continue.

J-4

Message

Program

END OF FILE

Run Time

END OF INPUT FILE AT RECORD nnnn
DAFT

~xplanation

The last record of an input file
has been read and the end of file
mark encountered but no EOF label
was
specified
in
the
XMIT
statement.
Attempted to read past end of file
mark on input. Not necessarily an
error.

END TAPE ERROR PHASE n

DTMARK

Refer to DEC-08-EUFB-D.

ENTER "DATE MM/DD/YY"

Monitor

Typed an unrecognizable date.

ERROR IN COMMAND

Monitor

Miscellaneous.

EXEC WORD ERROR

RK8MRK

An error occurred while reading the

EXEC.
EXCESSIVE GRID SIZE

word of the sector.

DAFT

The grid printer width may not
greater than 130 characters.

COMP

A required label is missing.

COMP

EXPECTED LABEL IS MISSING
EXPRESSION NOT ALLOWED

EXTRA CHARS AT END OF DATA
BUILD

expression or bad
character
occurs to the left of an =.

More data was entered than was
defined in the control programo

EXTRA CHARS AT STMNT END BUILD
SORT
UPDATE
COMP

Characters not relating to
statement appear on line.

FIELD NOT DECIMAL

BUILD

Field not defined as decimal.

BUILD

Field number or default unit. nmnhsr
is missing or is 0 or greater than
511.

FIELD NUMBER MISSING OR 0
SORT
UPDATE
UPDATE
FIELD OUT OF RANGE

BUILD

Data entered is
range
specified
program.

FIELD TOO LARGE OR 0

COMP

In a data description statement,
the dimension was 0 or more than 3
digits long, or the field size was
0 or larger than 511.

J-5

not within the
in the control

Message

Program

Explanation

FILE NOT FOUND

Monitor
PATCH

The program with the name specified
was not found. After, for example,
FETCH FILEX.

FORMAT ERROR

RK8MRK

An error occurred while reading the
header word.

FORMAT PARITY ERROR

RK8MRK

A parity error occurred in either
the Header word or Exec word while
reading the sector.

FULL

CONVEX

Specified an output device with -s
option
and
source
file being
created was filled.

HASH FIELD MISSING

BUILD

No field specified for
hash total.

storage

I/O ERROR ON xx, RETRY?

Monitor

System failed in three attempts
read from or write to a device.

I RECORD ALREADY EXISTS

UPDATE

Tried to insert
present.

record

already

ILLEGAL CHAIN AT LINE 0000

CHAIN argument does not match
command.

ILLEGAL DEVICE

Device other
DK0-DK3.

SYS GEN

ILLEGAL DEVICE AT LINE nnnn
Run Time

than

DTO-DT7

.RUN
or

Attempt to WRITE on a file that was
not
INITialized
for UPDATE or
attempt to READ from a file that
was not INITialized or INITialized
for INPUT or UPDATE.

ILLEGAL DEVICE SWITCH

PIP

A switch was specified that was not
/R, /C or /K for input or /P, /L or
/T for output.

ILLEGAL OPERATOR

COMP

A bad character was encountered in
an expression where an operator
would be expected.

ILLEGAL RECORD # AT LINE nnnn
Run Time

ILLEGAL SORT KEY

SORT

Record number is O, past the end of
the physical unit or the length
specified in the record header word
does not match the length of the
XMIT block (records in data file
are all not the same length).
Bad syntax on
too
complex
missing.

J-6

KEY
or

statement, KEY
KEY statement

Message
ILLEGAL STMNT

Program
COMP

ILLEGAL SUBSTRING AT LINE nnnn
Run Time

?ILLEGAL UNIT

Monitor

Explanation
The statement was not
a
data
manipulation statement (it had no
=) nor did it
start
with
a
recognizable keyword.
A DIBOL PROCedure section statement
has attempted to access a data
field, FI(m,n), but m=O or m n.
The unit specified is not
or DTO-DT7; for example,

DKO-DK3

WRITE FILE,XY3
WRITE FILE, RK9
ILLEGAL UNIT

SORT

Default unit is 0 or

IN USE

Monitor

The unit specified in answer to the
MOUNT
message is already being
used.

INCREMENT TOO BIG

BUILD

Length of increment field larger
than length of output field.

INSUFFICIENT SPACE ON DEVICE
SYS GEN

Attempt to allocate

15.

segments

than are available on a device.
INITIAL ALPHA VALUE DOESN'T BEGIN WITH QUOTE
BUILD
Beginning quotation mark missing
SORT
for initial alpha value.
UPDATE
INITIAL VALUE TOO BIG

BUILD
SORT
UPDATE

The initial value specified is
larger than the field size.

INITIAL VALUE WRONG SIZE COMP

The initial value
in
a
data
specification
statement
had
a
length different from the field
size specified.

INITIAL VALUE TOO SMALL

The initial value specified is
smaller than the field size.

BUILD
SORT

KEYWORD MISSING OR TOO BIG
BUILD
LABEL NOT ALLOWED

COMP

A
symbol
in
an
arithnetic
expression was not of type alpha or
decimal, or a symbol which had been
redefined was used.

J-7

Message

Program

Explanation

LAST INT NOT END ZONE

DTMARK

Refer to DEC-08-EUFB-D.

LINE # TOO LARGE

Monitor

Greater than 4095.

LINE TOO LONG

Monitor

Greater than 120 characters.

LINE TOO LONG AT LINE nnnn
Run Time

An input line

LOCATION TOO BIG

A location > = 400 was typed.

PATCH

(record)
overflowed
the block into which i t was read.

MARK TRACK ERROR PHASE n TDMARK
DTMARK

The DECtape being used is bad.
Try reformatting the tape.

!*!MEMORY FAIL!*!

Monitor

A hardware error has
Restart the program.

MISSING CLOSE PAREN

COMP

No close parenthesis occurred where
one was expected.

occurred.

MISSING CLOSE QUOTE ON ALPHA INITIAL VALUE
SORT
Quotation mark misplaced or
UPDATE
missing.
BUILD
MISSING FIELD NAME

BUILD

Field name is missing.

MISSING INITIAL VALUE

BUILD
SORT
UPDATE

Comma was inserted after type and
size but initial value was not
specified.

MISSING OPEN PAREN

COMP

No open parenthesis occurred
one was expected.

MISSING OPERAND

COMP

A binary operator occurs in an
expression
with
no
operand
following it; or no expression at
all occurs where one is expected.

MISSING OR BAD DEVICE

COMP

The devic.e i.n an INIT statement was
missing or started with an illegal
character.

MISSING OR BAD OPTION

COMP

In a record statement, no option
appeared
after a comma or the
option which appeared was not X or

where

c.
MISSING OR OUT OF ORDER

UPDATE

On C or D, record
Probably because
not sorted.

MISSING QUOTE

COMP

The statement contained
number of quotes (').

J-8

was not there.
scratch file was
an

odd

Message

Program

Explanation

MISSING RELATIONAL

COMP

No relational
statement.

MISSING UNIT

UPDATE

Unit number specified is a multiple
of 16.

NAME PREVIOUSLY DEFINED

COMP

The name
used
was
previously
defined and this statement tries to
redefine it ..

NECESSARY FIELD MISSING

BUILD

Field specified as necessary in
control program was not entered.

BOOT

Device switch illegal or missing.

NO BUFFERS LEFT AT LINE nnnn
Run Time

appeared

Not enough core available for I/O
buffers.
An I/O buffer of some
multiple of 512 characters is set
up for each active mass storage
file. Another possibility: too few
files were specified in the PROC
statement.

An attempt

PATCH

was made to write a
block but no changes were made to
it.
If this was on purpose fine
otherwise make the changes to the
block again.

BUILD
SORT
UPDATE

No comma or a character other than
comma was specified after the
field name.

NO DEFINE!

UPDATE

Control file did not start
DEFINE.
Returns
to
immediately.

NO END

CONVEX

Entire OS/8 ASCII input file was
read but no end of file {CTRL/Z)
was found.

NO END STATEMENT

UPDATE

END statement missing after
specification.

NO FILE AT LINE nnnn

Run Time

No file specified in RUN statement
to satisfy INIT (SYS) command.

?NO FILE TO SAVE

Monitor

Nothing in the edit work area
WRITE command is issued.

NO INIT DONE

Monitor

Program attempted a read or write
on a device not opened by the
system program.

NO CHANGE IN BLOCK

NO COMMA AFTER FIELD NAME

J-9

with a
Monitor

OUTPUT

when

Message

Program

Explanation

NO INPUT

SORT
MERGE

Input file is null or not enough
input files specified for a MERGE.

NO ROOM

PIP
CONVEX

Attempt to store a file on a full
device.

NONCE ERROR

DAFT

Tried to
use
a
feature
not
currently supported or documented.

NONEXISTENT FIELD

UPDATE

Tried to
appeared
section.

NOT A OR D

BUILD
COMP
SORT
UPDATE

A character other than A or D
occurred in a data specification
statement where A or D was
expected.

NOT BUILD FILE!

BUILD

File did not
statements.

update Fn where no Fn
in
file
description

start

with

DEFINE

NOT ENOUGH ROOM FOR SYSTEM AND FILES
SYSGEN
Device specified was too small to
accommodate
system
program and
files.
NOT FOUND

CONVEX

File with specified name not found.

NOT I,D, OR C

UPDATE

Bad first
command.

NOT LABEL

COMP

A symbol which was
occurred
where
required.

character

update

not a
'label'
a
label
was

NOTHING AFTER FIELD NAME BUILD
SORT
UPDATE

Field type and size are not
specified after field name and
comma.

NUMBER REPEATED OR OUT OF ORDER
BUILD
SORT
UPDATE

A field sequence number is used
more than once or is out of
ascending order sequence.

# TOO LARGE

SYS GEN

Number
4095.

NUMBER TOO LONG

Run Time

A decimal field longer than 15
digits was used in a calculation.

OUT OF ORDER

UPDATE

SORT statement missing and
commands are not in order.

J-10

entered

was

greater

than

update

Message

Program

Explanation

OUT OF ROOM!

BUILD

Too many descriptors
in
INPUT
section.
Maximum
keywords
=
approx.
121.
Approx.
maximum
fields for all keywords = 192. Too
many fields in OUTPUT
section.
Approx. maximum = 128.

OUTPUT ERROR

SORT

Indicates a system malfunction.

OUTPUT RECORD TOO BIG

BUILD

Record
is
characters.

PARITY ERROR

RK8MRK

A bit in data, parity or timing has
been picked up or dropped on read.
Transfer will continue to the end
of
the sector where the error
occurred.
Word
count,
current
address information can be used to
identify the error.

PARITY ERROR PHASE n

DTl-'.iARK

Refer to DEC-08-EUFB-D.

PROGRAM TOO BIG

COMP

Binary output too big for
the
binary scratch area.
Remedy: Run
PIP/n with Option E to enlarge
scratch area.

PROGRAM TOO BIG AT LINE nnnn
Run Time

PUSHDOWN OVERFLOW AT LINE nnnn
Run Time

READ STATUS ERROR

larger

than

Binary program does not
Reduce
available
core.
size.

510

fit in
program

Either (1)
a statement is
complex
or
(2)
subroutines
nested to a depth greater than
or a combination of the two.

too
are
50,

RKEMRK

Refer to DEC-08-DHRKD-A-D.

RECALIBRATE STATUS ERROR RKEMRK

Refer to DEC-08-DHRKD-A-D.

RECORD TOO BIG

COMP

A named record exceeded
in size.

510

words

REPLACE?

Monitor

Duplicate file names.
replace.

Type

RETURN WITHOUT CALL AT LINE nnnn
Run Time

The program tried to execute a
RETURN, but there was no place to
go; there was no corresponding CALL
statement.

J-11

Message

Program

Explanation

SECTOR NO GOOD

RK8MRK

SELECT ERROR PHASE n

DTMARK

SETUP?

TDMARK
DTMARK

Indicates an error in the DECtape
setup. One of the units specified
is in WRITE LOCK position, not
selected, or the write flip-flop is
unable to be set, or there may be a
timing error.

STMNT TOO COMPLEX

COMP

The statement generated too much
code and overflowed the Compiler's
code buffer, or it had too much
nesting
and
overflowed
the
Compiler's pushdown stack. Remedy:
break up the statement into smaller
parts.

SUBSCRIPT ERROR

COMP

No cormna or
close
parenthesis
occurs after a subscript.

SUBSCRIPT NOT DECIMAL

COMP

The type of
decimal.

SUBSCRIPT TOO BIG AT LINE nnnn
Run Time

The program attempted to read or
write data on a sector whose header
words indicated a bad sector.
The
transfer is terminated inunediately.

subscript

was

not

Program attempted to destroy the
run-time system or itself by using
a large subscript; larger than that
defined
in the Data Definition
section. Note that the run-time
system does not detect all illegal
subscripts; only those which would
cause the user's program or the
system to be destroyed.

SWITCH NOT SET TO WTM OR SINGLE LINE FLAG FAILED TO SET
SET SWITCH TO WI'M.
TDMARK
Switch on the M868 module is not
set to the WTM position_ or the
timing generator for writing the
mark
and timing tracks is not
setting the single line flag.
SYMBOLS DEFINED BUT NOT REFERENCED; xx
CREF
Symbols were
referenced.
error.

J-·12

defined
but
not
Not necessarily an

Message

Program

SYNTAX ERROR

SYS GEN

Missing
etc.

TEXT AREA FULL

Monitor

Greater than 8,150 characters.

TIME OUT ERROR

RK8MRK

The control did not complete
operation after 32 revolutions.

TIMING ERROR PHASE n

TDMARK
DTMARK

Refer to DEC-8E-EUZC-D.
Refer to DEC-08-EUFB-D.

TOO BIG

BUILD

Data entered is greater than the
size of the field as defined in the
control program.

TOO MANY FILES

MERGE

More than 6 input files specified.

TOO MANY ITEMS

COMP

More items appeared in a
data
initialization than were specified
by the dimension.

TOO MANY SYMBOLS!

COMP

A fatal error message 365 symbols
allowed
in symbol table in 8K
system, and 511 symbols allowed in
larger systems. The compiler stops
compiling; no storage map can be
produced.

TOO MUCH DATA

C01"..P

Program's data division exceeds 32K
bytes.

TRACK ADDRESS ERROR

RK8MRK

Track, surface or sector address
read from the disk did not agree
with the address count registers or
the
disk
drive
electronics
indicated track position 000 and
the track counter did not agree.
The
transfer
is
terminated
immediately.

TRACK CAPACITY EXCEEDED ERROR
RK8MRK
UNDEFINED FIELD

BUILD

Explanation
comma,

extra

characters,

The program attempted to
write beyond sector 17.

read

Field referenced was not defined in
field descriptor section of control

program

decimal

field

was

being equated to an alpha field, or
an alpha field was specified for a
flag.
UNDEFINED FORMAT

BUILD

Format number referenced
defined
(errors
given
statement).

J-13

but
on

not
END

Message

Program

Explanation

SYNTAX ERROR

SYS GEN

Missing
etc.

TEXT AREA FULL

Monitor

Greater than 8,150 characters.

TIME OUT ERROR

RK8MRK

The control did not complete
operation after 32 revolutions.

TIMING ERROR PHASE n

TOM.ARK
DTMARK

Refer to DEC-8E-EUZC-D.
Refer to DEC-08-EUFB-D.

TOO BIG

BUILD

Data entered is greater than the
size of the field as defined in the
control program.

TOO MANY FILES

MERGE

More than 6 input files specified.

TOO MANY ITEMS

COMP

More items appeared in a
data
initialization than were specified
by the dimension.

TOO MANY SYMBOLS!

COMP

A fatal error message 365 symbols
allowed
in symbol table in 8K
system, and 511 symbols allowed in
larger systems. The compiler stops
compiling; no storage map can be
produced.

TOO MUCH DATA

COMP

Program's data division exceeds 32K
bytes.

TRACK ADDRESS ERROR

RK8MRK

Track, surface or sector address
read from the disk did not agree
with the address count registers or
the
disk
drive
electronics
indicated track position 000 and
the track counter did not agree.
The
transfer
is
terminated
immediately.

TRACK CAPACITY EXCEEDED ERROR
RKSMRK

comma,

extra

characters,

The program attempted t-o
write beyond sector 17.

read

UNDEFINED FIELD

BUILD

Field referenced was not defined in
field descriptor section of control
program or a decimal field was
being equated to an alpha field, or
an alpha field was specified for a
flag.

UNDEFINED FORMAT

BUILD

Format number referenced
defined
(errors
given
statement).

J.-14

but
on

not
END

Message

Program

Explanation

UNDEFINED NAME

COMP

A name is used which was
defined in the data section.

UNEXPECTED END OF FILE

BUILD
UPDATE

Missing END statement.

UNIT xx IS FREE

SORT

NOT
AN
ERROR
merely
an
informative
message to aid the
operator. xx is a COS unit number.

SORT

Parameter line did not start with a
good keyword.

WRITE ERROR PHASE n

TDMARK

Refer to DEC-8E-EUZC-D=

WRITE LOCK ERROR

RK8MRK

The program attempted to write a
section that was write protected.
The write operation is terminated
immediately.

WRITE STATUS ERROR

RKEMRK

Refer to DEC-08-DHRKD-A-D.

WRONG DATA TYPE

COMP

Mixed
modes
occurred
in
an
expression; or an argument which
was supposed to be decimal was not
or one of the three arguments in a
data manipulation statement was of
the wrong type.

Run Time

The program attempted to divide
zero.

CREF

The line
specified
interpreted.

L~~RECOGNIZABLE

LINE

never

ZERO DIVISOR AT LINE nnnn

xxxx IS BEING IGNORED

xxxx SHOULD BE yyyy BLK ERROR PHASE n
DTMARK
Refer to DEC-08-EUFB-D.
xxxx SHOULD BE yyyy DATA ERROR PHASE n
DTMARK
Refer to DEC-08-EUFB-D.

J-15

cannot

APPENDIX K
COMMAND SUMMARY

The commands summarized below are described in detail in Chapter 2
this manual.

1.0

COMMANDS FOR MONITOR FUNCTIONS
Command

Function

BAtch pronarn

Executes a string of
Monitor commands.

CTRL/C

Returns control to the Monitor.

CTRL/N

Turns on/off imbeddded numeric keypad at
terminal (see Table 2-1).

CTRL/O

Stops/starts terminal echo of
typed
input.
Also halts output from a LIST
command and the printing of the compiler
listing.

CTRL/U

Deletes the current input line.

CTRL/V

Stops/starts terminal echo of Monitor
version message COS MONITOR V2.1108.

CTRL/Z

Signals end of input and returns control
to the Monitor. Halts display of line
numbers from an LN command.

DAte mm/dd/yy

Stores the specified date for assignment
to new programs created and reports
printed.

DElete pronam[,dev]/x

Removes the named
system
program
directory.