Further Computer Appreciation T. F. Fry ACMA, AMBIM, FIDP Head of Department of Business and Secretarial Studies Cassio College, Watford, Herts. NEWNES-BUTTERWORTHS LONDON - BOSTON Sydney - Wellington - Durban - Toronto The Butterworth Group United Kingdom Butterworth & Co (Publishers) Ltd London: 88 Kingsway, WC2B 6AB Australia Butterworths Pty Ltd Sydney: 586 Pacific Highway, Chatsworth, NSW 2067 Also at Melbourne, Adelaide and Perth Canada Butterworth & Co (Canada) Ltd Toronto: 2265 Midland Avenue, Scarborough, Ontario, M1P4S1 New Zealand Butterworths of New Zealand Ltd Wellington: T & W Young Building, 77-85 Customhouse Quay, 1, CPO Box 472 South Africa Butterworth & Co (South Africa) (Pty) Ltd Durban: 152-154 Gale Street USA Butterworth (Publishers) Inc Boston: 10 Tower Office Park, Woburn, Mass. 01801 First published by Newnes-Butterworths 1977 Reprinted 1978, 1979 ©T. F. Fry, 1977 All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, including photocopying and recording, without the written permission of the copyright holder, application for which should be addressed to the Publishers. Such written permission must also be obtained before any part of this publication is stored in a retrieval system of any nature. This book is sold subject to the Standard Conditions of Sale of Net Books and may not be resold in the UK below the net price given by the Publishers in their current price list. ISBN 0 408 00239 5 Typeset by Butterworths Litho Preparation Department Printed in England by Butler & Tanner Ltd: Frome and London Preface The concept of Further Computer Appreciation is, on the one hand to follow on from some of the principles discussed in Computer Appre- ciation and, on the other hand, to include some new aspects for which there was no room in the first book. In the five years since Computer Appreciation was first published some examination syllabuses have been revised and expanded and it is hoped that this further text will provide students with a more com- prehensive cover for these syllabuses. Examinations I have particularly borne in mind are the Royal Society of Arts Computer Appreciation, Ordinary National Certificate and Diploma in Business Studies, The Institute of Administrative Accounting and the Data Processing papers of the Institute of Cost and Management Accountants and the Asso- ciation of Certified and Corporate Accountants. As with Computer Appreciation, I have tried to keep the text as simple and as non-technical as possible explaining perhaps more what is done than how it is done, and have orientated it towards the use of computers in business. May I express my thanks to those firms who have kindly supplied illustrations and systems descriptions—I.B.M., I.C.L., K & N Electronics Ltd., Kienzle Data Systems Ltd., National Cash Registers Ltd., and Westrex Company Ltd. Finally my thanks to Marian Bentley who with admirable patience deciphered my handwriting and most efficiently typed the manuscript. T. F. FRY 1 The Development of Computers little more than thirty years have passed since the commissioning of the first computer, designed and constructed at Harvard University in 1944. Over this short period of time have been developed the highly sophisticated and versatile machines in use today. Until thirty years ago, basic principles in the design of calculating devices had changed very little during the previous 250 years. Blaise Pascal's original concept of geared cog-wheels to perform addition and subtraction was improved upon to give a multiplication and division capability by Baron Gottfried Von Leibnitz and refinements developed and added over a number of years by a succession of innovators resulted in machines technically more efficient, smaller and more convenient to use. Limited areas of automation were exploited, particularly in the areas of Punched Cards and pre-set mechanical programming devices. Machines became powered electrically rather than by hand, but the cog-wheels still turned, the levers still pushed, the pins still probed, the relays still clicked. Little fundamental change had taken place. THE FIRST COMPUTERS Indeed, the machine completed at Harvard University in 1944 (the Automatic Sequence Controlled Calculator (ASCC)) did not, itself, get away from these basic mechanical principles and processes. While it is generally recognised as being the first working automatic computer, it was a far cry from machines as we know them to-day. It had no elec- tronic storage either for data or for program and the sequence in which instructions were recorded on paper tape had to be rigidly followed. It was slow, but within its limitations was a successful machine, performing the task for which it was designed for some fifteen years or so. In 1944 John von Neuman was largely instrumental in the design and development of the first full Automatic Electronic Computer known as ENIAC—Electronic Numerical Integrator and Computer. This immense machine completed at Pennsylvania in 1946 weighing thirty tons and containing some 18000 thermonic valves had a very limited storage capacity and no facility for storing its program. Instructions had to be conveyed to the machine through the medium of wired plug 1 2 The Development of Computers boards and by setting switches. However, compared with ASCC, opera- ting speed had decreased dramatically with a cycle time of 200 /is. Nevertheless, this machine marked the beginning of the development of truly electronic computers and, indeed, can be regarded as the first in what is often referred to as the first generation of these machines. It was during the construction of ENIAC, in 1945, that von Neuman published his report on EDVAC (Electronic Discrete Variable Automatic Computer). This introduced the design concept of an internal elec- tronically stored program and the idea of a machine with a storage capacity of 1000 words of 10 decimal digit capacity. However, EDVAC was not to be built until 1950, one year after the completion of EDSAC (Electronic Delay Storage Automatic Computer) —the first Automatic Electronic Stored Program Computer—at Cam- bridge University. This incorporated a mercury delay line store of sixteen 35-bit words and 3800 thermionic valves. Input and output media was 5 track paper tape reading and writing at 15 c.p.s. Itwas early in 1949 that Lyons, working in conjunction with Dr. Wilkes at Cambridge, decided to build a computer. Work on the machine commenced later on in the year, and incorporated a new type of storage medium, magnetic tape. The machine was known as LEO (Lyons Elec- tronic Office) having valves and magnetic tape storage. It was com- pleted in 1951. After trials the first application, bakeries sales evaluation, was implemented later on in the year, and in 1953 Lyons payroll was computerised on LEO I which coped with this work quite successfully until the machine was closed down in 1965. Up until 1951, the research and development of computers was generally the exclusive province of the Universities. From this point onwards, however, office machine and electrical equipment manufac- turers started to develop machines for commercial use, putting into practice the principles and techniques developed at the universities. This development included work by ICT—a combine of BTM (Hollerith) and Power Samas—to produce HEC I (Hollerith Electronic Computer) in 1951 and later, in 1953, HEC II, seven of which were marketed from 1954 onwards. The Ferranti Mark I was completed and delivered to Manchester University in 1951 and, indeed, was the first general purpose computer available commercially. Parallel with this development in hardware, intensive work was done on computer programming software. The mid-fifties saw the development of a new range of computers with new types of immediate access store— magnetic drum and magnetic core storage. At about this time a number of companies were involved in the construction and marketing of digital commercial machines, using these new developments. The Development of Computers 3 The fourth version of ICT's Hollerith Electronic Computer (HEC4) was marketed as the 1201 with 1 K of store, and 1202 with 4 K of store, a total of one hundred and twenty-four of these machines being sold. English Electric produced DEUCE (Digital Electronic Universal Calculating Engine) with both magnetic drum and mercury delay line storage. IBM introduced their 700 series and Honeywell their 400 and 800 computers. SECOND GENERATION COMPUTERS 1950 proved to be a landmark in computer technology, with the discovery of the planar silicon transistor and the incorporation of this device in the UNIVAC II machine. This point is generally recognised as the commencement of the second generation of computers. A year later in 1957, the ICT 1300 series of computers was introduced and the second edition of LEO, of which eleven machines were sold. Com- mercially, the most successful machine of this period was IBM's medium size machine, the 1400 series, with a virtually world wide acceptance. As hardware developed in power, complexity and sophistication, so it became necessary to develop programming software that eliminated 1st 2nd 3rd generation generation generation 19U THERMONIC VALVES j j PROGRAM PLUG BOARDS J MERCURY DELAY LINE STORAGE PUNCHED CARDS ! 1950 PUNCHED PAPER TAPE' ELECTRONICALLY STORED PROGRAM MAGNETIC TAPE! 1950 ITRANSISTORS j | MAGNETIC DRUM STORAGE FERRITE CORE STORAGE I TIMEJ SHARING SYSTEMS 1964 ' REAL TIME SYSTEMS MAGNETIC DISCS 1964 MULTI-PROGRAMMING FAMILY CONFIGURATIONS _i_ Figure 1.1. Computer generations the tedium and labour of program construction in machine orientated coding. Manufacturers had tended to develop their own programming codes in isolation directed to the use of their own particular machines and while a general standardisation of machine codes was impracticable due to the variance in design and construction of machines, the concept 4 The Development of Computers of high level languages in which programs could be universally expressed began to take shape. The first of these, FORTRAN, was introduced by IBM with their 700 series machine. The European Association for Applied Mathematics and Mechanics and the American Association for Computing, meeting in Zurich in 1958, began research into the development of a universal programming language and eventually produced ALGOL 60. A couple of years later the USA government in conjunction with computer manufacturers and users sought to develop a language specifically designed for commercial data processing; this resulted in the introduction of the COBOL language. New ideas were now taking shape in hardware development and use, and we find the terms Time Sharing and Real Time being introduced. IBM pioneered Time Sharing techniques by linking four control ter- minals to a 709 machine in the early sixties. This was known as Project MAC (Multiple Access Computers). The first machine with the capability of storing more than one program simultaneously and to switch from program to program according to processing demands was LEO II in 1961, from which was later developed LEO 326 and 360 models. It was the Sabre Reservation System for American airlines that represented the first large scale real-time system involving some twenty-five terminals covering a radius of 300 miles. THIRD GENERATION COMPUTERS In 1965/6 the third generation of computers were introduced. At this point, problems exercising the minds of companies involved in computer development can be summarised as: 1. To increase the versatility of configurations in the sense that advan- tage could be taken of any combination of the range of available peripherals by providing the capacity for interchange within the configuration. 2. To make greater use of the processing speed now provided by the central processor. 3. To provide hardware capable of meeting systems application demands for immediate and random access to records. 4. To develop means whereby a central computer could accept, process and transmit answers to a number of remote points simultaneously. 5. To provide input and output devices capable of fast and accurate transmission of information to and from the central processor. Up to this stage, computer configurations tended to inflexible. For instance a central processor of a given storage capacity with input and The Development of Computers 5 output peripherals of say, a card reader and a line printer, and with magnetic tape storage, would probably remain as such with modifications to the configuration unlikely. THE DEVELOPMENT OF PROCESSORS AND STORAGE SYSTEMS The development and introduction of a wider range of input, output and backing storage peripherals gave rise to the need to develop central processors with interfaces capable of accepting selections from within this range at the choice of the user. This in turn meant central processor design flexible enough to enable core store capacity to be increased as required. Compatability and versatility became the keynotes in those elements making up a computer configuration, so leading to the concept of what have been called family machines. The ICT 1900 series and the IBM 360 series are examples of these. As, in the early days of computing, input devices were no longer restricted to punched media, final output to a printed format or backing storage devices to magnetic tape. The demands of systems applications called for new input techniques, as for example magnetic ink character and optical character readers as well as direct remote on-line input. While expansion in the range of output devices was not as marked, the line printer was still generally the most suitable for commercial systems; never-the-less, for some specialised requirements visual display units or direct output to remote terminals were to play their part. In the realm of backing storage the need for large scale direct access stores had to be satisfied. This was mainly accomplished through the medium of magnetic discs. A further development centred around the need to take the greatest advantage of the speeds at which central processors could work. For all practical purposes, processing speed is governed by the rate that information can be conveyed to the central processor on the one hand and by the rate at which information could be output on the other. The comparatively slow speed of such input and output devices meant central processing units working at only a fraction of their capacity. Multi-processing became in part an answer to this with a central pro- cessing unit capable of storing two or more programs simultaneously and switching from one to another in order to decrease the amount of time the processor was standing idle awaiting work. While serial access backing storage devices were ideal for fast batch processing techniques, provided input records were sorted into an order compatible with those on the storage device, the time lag between the occurrence of an event and the updating of the records effected by the event was considerable. With the introduction of backing storage con- taining records, any one of which could be accessed in a very short 6 The Development of Computers time, came the concept of updating information as and when the event occurred. This necessitated the availability of an input device that could be used on the site of the occurrence. From this came the development of terminals, remote from the computer, but linked by direct line so that information transmitted could immediately be processed and, if required a response in the form of the results of processing immediately transmitted back to the terminal. THE PRESENT POSITION Computers in use today fall into a number of fairly well defined categories: 1. Small to medium sized configurations with core sizes around 16 K to 64 K supporting a range of peripheral devices. Backing storage of either serial or direct access type, that is magnetic tape or disc with input through punched media and output usually to a line printer. Generally used in a batch processing mode with a single stored Program. These are self-contained installations, accepting source data, performing data conversion procedures, processing files and out- putting reports in hard copy form. 2. Larger configurations capable of working in a multi-programming mode. With a larger core store to accommodate a number of prog- rams simultaneously and more backing storage to hold a wider range of files on-line. 3. Large time-sharing machines linked to users by remote terminals. These machines need high volume backing stores to hold programs that can be called in on demand and to hold a wide range of record files. 4. Real-time systems for up-dating records as the movement activity occurs, and for file interrogation purposes. Again these machines support remote terminals and need a large core store and high volume backing storage. 5. Machines using a data bank storage principle where all records are held in direct access devices available on-line for use in any pro- cessing system. 6. Dedicated machines. These are specialised machines designed for one specific application, and it has been the development of mini and later micro solid state electronics that has widened the range of applications possible with the microcomputer minicomputer. An example is the automatic recording of point of sale transactions either through the medium of till-keyboard depressions or light pen scanning processes of coded information on product labels. The Development of Computers 7 7. Visible record computers. These are sometimes known as mini- computers or desk type computers, and have popularity with the small and medium sized firm where requirements include a pro- gressively updated printed record of its transactions as well as electronically recorded data for analysis purposes and for reference. The type of machine has very much gained in popularity over the past several years. A detailed account of an application on a visible record computer is given in Chapter 13. FUTURE TRENDS An attempt to forecast where computers go from here can only be made in the light of current developments and trends. As far as machine sizes are concerned, the general tendency appears to be away from medium sized machines, towards very large central configurations serving a number of users remote from the machine but linked by data transmission lines. However, the other end of the spec- trum should not be overlooked. The increase in the use of small visible record computers has been dramatic over the past few years and for the fairly small business are becoming increasingly popular. Physically, central processors are tending to become smaller in size for equivalent power. This has been due to the development of mini circuitry followed by micro solid state circuitry, a process that is continuing. Research is going on into alternative kinds of immediate access store, as an alternative to ferrite cores. For example, semi- conductor memory units capable of holding a thousand or more bits are becoming generally available in solid state chips. Input and output devices of the future will tend more and more to eliminate the laborious process of converting data into a machine acceptable form. While magnetic tape character recognition and optical character recognition are already with us in limited form, the capture of data at source is the subject of a great deal of interest and research. New types of mass backing storage are the subject of research and experimentation. Will the conventional magnetic tape and magnetic disc be replaced eventually with such devices as for example mass data cartridge or electronic beam memory devices giving far larger on-line capacities, faster direct access rates, and higher data transmission speeds? Will acoustic input and output develop from its present very limited application and become a generally used media for communicating with the machine and indeed will sensory input by attaching electrodes to our head to pick up thought waves ever become a reality? Whatever happens, computers will continue to develop, input methods will become more direct, data banks will become more vast. Perhaps, in thirty years time computers will bear as little resemblance