The adoption of open source software by singaporean companies

1.1 Background to the Research The development and distribution of computer software has undergone an evolutionary process.. Open Source could be seen as a shift in the paradigm in whi

Trang 1

Jan 2009

Trang 2

Page i

Abstract

Based on collaboration rather than competition, Open Source Software provides a

new dynamic in the development and use of software systems As such it has the

potential to make a significant legal, social and economic impact on the industry

While its origins trace back to the start of the software industry, the recent success of

the Linux operation system, Apache web server or the Mozilla Firefox Internet

Browser provide impetus to the growth of interest in this movement However, while

a number of studies have been conducted on its development, few have provided

empirical evidence of its adoption within the South East Asian context This study

aims to investigate factors leading to the adoption of Open Source Software in

Singaporean Organisations

The research has found that the adoption of Open Source Software is driven by the

perception of a cost advantage The organisations interviewed have acknowledged

cost as being one of their biggest concerns and top priorities While costs were stated

to be of major concern to the organisations, objective measures of cost such as Total

Cost of Ownership (TCO) and Return on Investment (ROI) were rarely used by the

organisations studied This perception of cost saving is found to be led by industry

sources such as publications, conferences and websites

The next significant finding is the need for increased open source software skills in

the industry One of the major drivers of Open Source Software Adoption in the

organisations is that they posses pre-existing skills in Open Source Software use

This enables them to better mitigate risk and to lower their training costs

The final principal finding is that Open Source Software appears to be used mainly in

systems infrastructure applications Organisations reported a large degree of

satisfaction including increased stability, scalability and cost effectiveness Issues

remain with Open Source Software’s manageability, its quality of support and ease

of use

Trang 3

Page ii

Certification of Dissertation

I Edmund Koh Ker Yuan hereby certify that ideas, experimental work, results,

analyses, software and conclusions reported in this dissertation are entirely my own

effort, except where otherwise acknowledged I also certify that work is original, and

has not been previously submitted for any other award, except where otherwise

acknowledged

_ 16/01/2009

Trang 4

Page iii

Acknowledgements

I would like to take this opportunity to thank everyone who has provided me with

assistance and support for the duration of this study In particular, I would like to

thank the following people who have helped make this thesis a reality

I would like to thank my supervisors, Dr Dietrich Splettstoesser from the University

of Southern Queensland for the beginning stages my research, Dr Greg Timbrell and

Professor Alan Underwood from the Queensland University of Technology who

have help me focus on the task at hand I would also like to thank Ed Fitzgerald, Guy

Gable and Chan Taizan for their academic assistance I would also like to thank

Steve McWhirter and Zon Lim from Red Hat Asia Pacific for providing me with

access to their customers and insight into the operations of their organisation

Finally I would like to thank my family whose support and patience made this

research possible

Trang 5

Page iv

Table of Contents

Abstract i

Certification of Dissertation ii

Acknowledgements iii

List of Figures ix

List of Tables x

1 Introduction 1

1.1 Background to the Research 1

1.2 Justification for the Study 3

1.2.1 Size of the Industry 4

1.2.2 Literature Review 5

1.2.3 Benefits to Practitioners 6

1.3 Research Problem 6

1.4 Methodology 7

1.5 Principal Findings 8

1.6 Contribution 9

1.7 Outline of the report 9

2 Background 12

2.1 History of Propriety Software Development 12

2.1.1 The Birth of the software Industry 13

2.1.2 Bundled Software 15

2.1.4 Software Crisis 19

2.1.5 Standards and the need for interoperability 20

2.2 Emergence of Open Source Software 21

2.2.1 History of Open Source Software 22

2.2.2 Linux 27

2.2.3 Apache 29

2.2.4 Sendmail 30

2.2.5 Mozilla 31

2.3 Summary 31

3 Literature Review 34

3.1 Unbundled Proprietary Software and Property Rights 34

Trang 6

Page v

3.1.1 Patents 35

3.1.2 Copyright 36

3.1.3 Trade Secrets 37

3.1.4 Trademarks 38

3.2 Definition of Open Source 39

3.2.1 Berkeley Style Licenses (BSD-style) {Appendix A} 43

3.2.2 GNU General Public License (GPL) {Appendix B} 43

3.2.3 GNU Library General Public License (LGPL) {Appendix C} 44

3.2.4 Mozilla Public License (MozPL or MPL) {Appendix D} 45

3.3 Open Source Development Characteristics 45

3.3.1 Developers are users 46

3.3.2 Community 46

3.3.2.1 Balance of centralization and decentralization 46

3.3.2.2 Meritocratic culture 47

3.3.3 Motivation 47

3.4 Advantages 48

3.4.1 Costs 48

3.4.2 Customisation 50

3.4.3 Quality 51

3.4.4 Transparency 52

3.4.5 Reduced Vendor Control 53

3.5 Limitations 53

3.5.1 Development Guarantee 54

3.5.2 Publicity 54

3.5.3 Liability 55

3.6 Open Source Software Successes 56

3.7 Research on Open Source Software 56

3.7.1 Reasons for Open Source Initiatives by firms 58

3.7.2 Open Source Software diffusion 59

3.8 Research Question 60

3.9 Software Adoption 61

3.9.1 Technology Adoption Cycles 61

Trang 7

Page vi

3.9.2 Factors in Software adoption 66

3.9.3 Adoption Frameworks 67

3.9.3.1 Technology-push, need-pull 68

3.9.3.2 Information Systems Strategic Planning 70

3.9.3.3 Glynn, Fitzgerald and Exton Framework for OSS Adoption 74

3.9.5 Framework of Open Source Software Adoption 76

3.10 Conclusion 81

4. Methodology 83

4.1 Research Design 83

4.1.1 Research Paradigms 83

4.1.2 Qualitative Research 86

4.2 Case Study 87

4.2.1 Validity and reliability 88

4.2.1.1 Construct Validity 89

4.2.1.2 Internal Validity 89

4.2.1.3 External Validity 89

4.2.1.4 Reliability 89

4.2.2 Research design and primary data collection 90

4.2.3 Sampling and case selection 90

4.2.4 Case and interview protocol 91

4.2.5 Data collection and case analysis 93

4.2.6 Ethical considerations 95

4.3 Conclusion 96

5. Results 98

5.1 Interview Administration 98

5.2 Framework of Open Source Software Adoption 100

5.3 Pilot Survey 102

5.4 In - Depth Interviews 103

5.4.1 Organisation A 103

5.4.1.1 Internal Organisational Factors 103

5.4.1.2 External Organisational Factors 104

5.4.1.3 Internal Information Technology Factors 106

Trang 8

Page vii

5.4.1.4 External Information Technology Factors 107

5.4.1.5 Use of Open Source 109

5.4.2 Organisation B 112

5.4.3 Organisation C 119

5.4.4 Organisation D 127

5.4.5 Organisation E 135

5.4.6 Organisation F 143

Trang 9

Page viii

5.5 Conclusion 152

6 Data Analysis 155

6.1 Results from the Prior Study 156

6.2 Organisational Analysis 161

6.3 Cross Case Analysis 173

6.3.1 Internal Organisational Factors 173

6.3.3 External Organisational Environment 180

6.3.4 External Information Technology 183

6.3.5 Conclusion concerning the adoption of Open Source Software 186

6.4 Conclusion 195

7 Conclusions 198

7.1 Summary of study 198

7.1.1 Research Question 198

7.1.2 Research Design And Methodology 198

7.1.3 Framework of Open Source Software Adoption 200

7.2 Results from the Prior Study 202

7.3 Summary of Findings 207

7.4 Contribution of Research to Knowledge and Practice 209

7.4.1 Contribution to Research 209

7.4.2 Implications for Practitioners 210

7.5 Limitations of the Study 211

7.6 Directions for Future Research 211

7.7 Summary 213

References 215

Trang 10

Page ix

List of Figures

Figure 2.1: Vertically Integrated Mainframe and Minicomputer Industry Structure 15

Figure 2.2: PC Industry Structure 20

Figure 3.1: Open Source Framework 45

Figure 3.2: Overall TCO by Environment 50

Figure 3.3: Rogers Technology Adoption Cycle 62

Figure 3.4: Hype Cycle for Open-Source Software 64

Figure 3.5: Technology Push vs Needs Pull 68

Figure 3.6: Ward and Griffiths ISSP Framework 72

Figure 3.7: Glynn, Fitzgerald and Exton Framework 75

Figure 3.8: Framework of Open Source Software Adoption 79

Figure 4.1: Outline 88

Figure 6.1: Research Model 101

Figure 5.1: Revised Framework of Open Source Software Adoption 158

Figure 7.1: Research Model 201

Figure 7.1: Revised Framework of Open Source Software Adoption 204

Trang 11

Page x

List of Tables

Table 1.1: Interview Cases 7

Table 3.1: Comparison of Software Licences 39

Table 3.2: Requirements for Open Source Licence 42

Table 3.3: Price comparisons between proprietary and open source operating systems in US Dollars as at September 2005 49

Table 3.4: Existing literature on the research of Open Source Software 57

Table 3.5: Software/ Technology adoption models 66

Table 4.1: Paradigm comparison 84

Table 4.2: Advantages and disadvantages of realism and positivism paradigms for research 85

Table 4.3: Summary of interview protocol content 93

Table 5.4: Porter’s 5 Forces on Organisation A 105

Table 5.5: Open Source Software Used in Organisation A 110

Table 5.6: Comparison of Open Source and Closed Source Software in Organisation A 111

Table 5.7: Demographics of Organisation B 113

Table 5.8: Porter’s 5 Forces on Organisation B 114

Table 5.9: Open Source Software Used in Organisation B 118

Table 5.10: Comparison of Open Source and Closes Source Software in Organisation B 118

Table 5.11: Demographic of Organisation C 120

Table 5.12: Porter’s 5 Forces on Organisation C 121

Table 5.13: Information Technology Environment on Organisation C 122

Table 5.14: Open Source Software Used in Organisation C 125

Table 5.15: Comparison of Open Source vs Closes Source Software in Organisation C 126

Table 5.16: Porter’s 5 Forces on Organisation D 129

Table 5.17: Open Source Software Used in Organisation D 133

Table 5.18: Comparison of Open Source vs Closes Source Software in Organisation D 134

Table 5.19: Demographic of Organisation E 135

Trang 12

Page xi

Table 5.20: Porter’s 5 Forces on Organisation E 137

Table 5.21: Open Source Software Used in Organisation E 141

Table 5.22: Comparison of Open Source and Closes Source Software in Organisation E 141

Table 5.23: Demographic of Organisation F 143

Table 5.24: Open Source Software Used in Organisation F 151

Table 5.25: Comparison of Open Source vs Closes Source Software in Organisation F 151

Table 6.1: Results of Regression Analysis 159

Table 6.2: Respondents Role in Organisation 174

Table 6.3: IT Department Role 175

Table 6.4: IT Reinvestment 176

Table 6.5: IT Procurement Sources 177

Table 6.6: Necessity of Modifying Software 178

Table 6.7: IT Success Measures 179

Table 6.8: Issues Facing Industry 181

Table 6.9: Environment’s impact on IT Strategy 182

Table 6.10: State of Singapore IT Industry 183

Table 6.11: Top 5 Concerns 184

Table 6.12: Concerns occupying most time 185

Table 6.13: Priority Issues 186

Table 6.14: Products used and Importance 188

Table 6.15: Origins of Organisations’ Open Source Use 189

Table 6.16: Open Source Implementation Experience 190

Table 6.17: Expansion of use 192

Table 6.18: Open Source Software Adoption Factors 192

Table 6.19: Open Source Issues 193

Table 7.1: Interviews 199

Table 7.2: Results of Regression Analysis 204

Trang 14

Page xiii

Trang 15

Introduction Page 1

1 Introduction

This chapter provides a background to the research problem It begins with an

overview of the software industry and an examination of intellectual property

Open source software is then examined, highlighting its growth in recent

years The research question and its objectives are then examined The scope

of the research as well as the methodology is outlined The chapter also

includes a brief description of the organisation of the dissertation

1.1 Background to the Research

The development and distribution of computer software has undergone an

evolutionary process This saw it move from a largely customer-specific

product during the period from the mid 60s to the early 80s, to one that is

both highly generic in nature and protected by intellectual property laws

Prior to the 1960s, computer software was seen as mere algorithms and not as

processes or machines (The History of Software Patents n.d.) Computer

manufacturers distributed software free of charge as part of their customer

service policy (Johnson 1998) This changed when IBM, the leading

computer manufacturer at the time, announced in June 1969 that the pricing

of software would be separated from that of pricing hardware (Johnson

1998) This ‘unbundling’ of software is seen as the starting point of the

software industry and was further reinforced by the U.S Supreme Court

decision to grant a software patent in the case of Diamond vs Diehr (The

History of Software Patents n.d.) Computer software was thus recognized as

a form of intellectual property

Intellectual Property is defined as “intangible property that is a result of

creativity, such as patents, copyrights, etc.” (Oxford English Dictionary 1998

intellectual property) Computer software fits this definition and is covered in

most countries by some form of copyright law (Evers 2000) Copyright is

Trang 16

“the exclusive, legally secured right to publish, reproduce, and sell the matter

and form of a literary, musical, dramatic or artistic work” (Encyclopaedia

Britannica copyright) This grant the creator five exclusive rights over his or

her work (Evers 2000):

 Reproduction Right: The right to duplicate work in fixed format

 Modification Right: The right to modify work to create something

new The result is called ‘derived work”

 Distribution Right: The right to distribute the work to the public

 Public Performance Right: The right to play, dance, act or show the

work at public places, or to transmit it in public

 Public Display Right: The right to show a copy of the work at a public

place or to transmit it to the public

These rights enable software creators to license software to potential users for

a set fee and to generate revenue based on that fee A number of software

companies such as Microsoft and Oracle have benefited greatly through the

application of copyright law (Fortune n.d.)

Unfortunately, the restrictions on the use and distribution of software

imposed automatically by copyright law are undesirable to an ever-increasing

number of companies An alternative approach to software development has

emerged, namely Open Source Software As a rule, open source software is

peer reviewed and has demonstrated a level of reliability and robustness that

can match and even surpass commercial software (The Open Source Initiative

n.d.) Open source software is distributed freely and thus potentially reduces

the total amount spent on information technology within an organization

Trang 17

(TechRepublic 1999) The licensing, however, is quite different from

traditional software licensing

Open source software is defined by its attached licence which abandons the

essential rights granted to the original creator under copyright law This gives

anyone the opportunity to distribute and modify any received open source

software (Evers 2000)

One notable example of the growing significance of open source software is

the Linux operating system Linux has gained a 1% market share in the client

operating system market and a 27% market share in the server market

(Kusnetzky & Gillen, 2001) Other examples, such as the Apache web server,

mySQL database and sendmail email server, are expected to play an

increasingly important role in the Information Systems (IS) function of

organizations in the future

A number of studies have been done to analyse the adoption of Open Source

technologies (Evers 2000; Murray 2000; Pedhazur and McClure 1999) but all

these studies were explorative in nature and none of these studies have

specifically addressed the market in Singapore A study of this particular

market is indicated because of Singapore’s significance as a regional and

global economic hub and its relatively highly developed technological

environment

1.2 Justification for the Study

The proposed study can be justified on the following grounds

 The size of the industry involved

 Current gaps in the literature

 Possible benefits in outcomes for practitioners

Trang 18

1.2.1 Size of the Industry

Singapore currently has one of the largest computing industries in South East

Asia Seventy seven percent of all business sites have at least one standalone

personal computer present; ninety three percent for businesses with over ten

employees (iDA 2005a) It had an estimated information and communications

manpower of 111,400 in 2003 (iDA 2005b) There are an estimated

1,001,000 personal computers and 239,000 servers (Gartner DataQuest

2000) On 6 June 2000, Singapore announced an e-government initiative to

improve the IT used by the government, to perform more services online and

to use the Internet to offer new services (Gartner DataQuest 2000) Deputy

Prime Minister Tony Tan said that the Singapore government plans to spend

S$1.5 billion during the next five years to support this initiative (Gartner

DataQuest 2000) The growth in IT spending has steadily increased until, in

the financial year of 2008, the government announced its intention to spend

up to S$1 billion in Infocomm Tenders (iDA 2008) With a total population

of roughly 4 million residents and a total of more than 100,000 companies

(Singapore Department of Statistics 2005), affordable computer software is

critical to the Singaporean Economy

In Singapore and globally, there is an increasing interest in Open Source

technologies especially in Linux, which is one of the most notable examples

of this A leading analyst for Merrill Lynch, an investment bank, argues that

open-source can be “disruptive technology” that could topple industry

heavyweights as Microsoft and Sun (Economist 2001)

Disruptive Technology is a term popularised by Christensen in his book The

Innovator’s Dilemma (1997) Open Source could be seen as a shift in the

paradigm in which software is produced, distributed and used (TechRepublic

1999) and where ownership of the software no longer belongs to any single

entity Christensen (1997) believes that the main reason that successful and

apparently well-run and well established organizations lose market share and

sometimes go out of business is that they fail to recognize the distinction

Trang 19

between sustaining and disruptive technologies It remains to be seen whether

Open Source Software, such as Linux, could achieve this outcome

In October 2000 the Linux User Group (Linux Users' Group Singapore, 2001)

was formed, advocating the use of open source software Their objective is to

promote the growth and development of knowledge with respect to the

development, evolution, use and popularity of the Linux Operating System

and other associated software and hardware (Constitution of Linux Users’

Group Singapore n.d.) It has held several meetings, seminars, and hosts a

website (http://www.lugs.org.sg/) which is used as a forum promoting open

source technologies in Singapore It has not yet initiated any empirical studies

on the penetration of Linux and similar technologies

1.2.2 Literature Review

So far there have been only few studies exploring the adoption of open source

software in various countries Such a study has never been undertaken

specifically for Singapore This research will, therefore, add to the

Information Systems body of literature

Although the history of Open Source Software is considered to have started at

the dawn of computer technology (Economist 2001; Murray 2000; Evers,

2000), research on its impact on business has only begun during the late

1990s Very few empirical studies have been undertaken on the subject

despite its wide media coverage and a vast number of practitioners’ articles

Murray (2000) has addressed Open Source Adoption in Ireland and Evers

(2000) has addressed its impact on software development in Germany The

most recent work done in the adoption of Open Source Software done at the

time of this research is Mindel, Mui, and Verma’s (2007) paper on Open

Source Software Adoption in Association of South East Asian Nations

(ASEAN) Member Countries Unfortunately, this research is severely limited

as it only examines whether organisations in ASEAN member countries used

Trang 20

open source in their web server applications and not what factors had driven

them towards the use of Open Source Software

1.2.3 Benefits to Practitioners

This research offers several benefits for practitioners The research will

provide practitioners with a better insight into the Open Source market place

in Singapore It will help IT managers select and implement open source

solutions within their organizations Distributors such as Red Hat Enterprise

and Caldera International Inc will also benefit from the findings by

identifying ways of how to better tailor their products towards the

Singaporean marketplace

Similarly, software developers considering the Open Source model will also

gain insight on where and how their software is being used, which may help

them in developing software that better suits their users’ needs Finally, IT

managers who are considering using Open Source products will be better able

to understand some of the advantages and disadvantages associated with

Open Source adoption

This research intends to address the following research question:

What are the factors leading Singaporean organizations to adopt Open Source Software?

It will specifically explore what kinds of business are adopting Open Source

Software, in what areas, and for what reasons It will also examine if there is

resistance to this new paradigm and what is its future The unit of analysis is

the organisation

Trang 21

1.4 Methodology

A variety of qualitative research techniques have been employed in IS

research (Markus & Lee 1999; Myers 1997), including ‘‘Revealed Causal

Mapping’’(Nelson et al 2000), ‘‘repertory grids’’ (Hunter & Bock 2000) and

content analysis (Altinkemer K et al 1994; Pliskin & Romm 1997)

This study follows a post-positivist paradigm through the use of a

Multi-Organisational Case Study, conducted using in-depth interviews McPhail

(1996) groups research into three fundamental types, exploratory, descriptive

and causal research This study is exploratory in nature and is focused on

developing a good grasp of the situation, which is the facts, people and

concerns and to discover ideas and to generate ideas with which to clarify the

research issue and model (McPhail 2001; Zikmund 2001; Neuman 1997) In

this study, six organisations are studied, segmented into small, medium and

large companies, which have already adopted Open Source Software These

organisations were introduced to the researcher by Red Hat Software, a

leading provider of Open Source Software solutions to the Singaporean

marketplace

Table 1.1: Interview Cases

6 In-depth Interviews Companies

A prior study was conducted that informed the development of the research

model and the interview questions Several limitations were found in the

execution of this prior study, mainly due to a low response rate The response

rate (1.97 %) was considered poor and thus the results of that study cannot be

generalised to the population (Zikmund 2001) Measures were taken - such as

survey pre-testing and contacting non-respondents - to increase the response

rate The poor result may also be explained by limitations beyond the control

of the researcher, such as:

Trang 22

The use of the Green Book CD-Rom yielded undeliverable mail responses of

ten percent and over fifty two percent undeliverable email responses A more

accurate database would hopefully reduce this number Limited time and

budget also contributed to the low response rate Polling a larger percentage

of the sample population might yield a larger number of usable responses

The Prior Study was conducted in 2001, before embarking on this research

Masters It was used to refine the research model developed from the

Literature Review (Chapter 3) and to assist in the formulation of the

interview protocol used in the Data Analysis (Chapter 6) A description of the

Prior Study is included in Appendix E

The principal finding in driving the adoption of Open Source Software in the

study was the perception of a cost advantage The organisations interviewed

have acknowledged costs to be one of their biggest concerns and top

priorities The interesting thing about the outcome of the case study is that

while costs appear to be a major concern to the cases, objective measures of

costs such as Total Cost of Ownership (TCO) and Return on Investment

(ROI) are rarely used amongst the cases The perception of low cost is led by

industry sources such as publications, conferences and websites

Trang 23

The next principal finding is the greater need for open source software skills

in the industry One of the major drivers of Open Source Software Adoption

in the cases is that they posses pre-existing skills on Open Source Software

use This enables them to mitigate risk and to lower training costs

The final principal finding is that Open Source Software appears to be used

mainly in systems infrastructure that runs important, critical applications

Organisations interviewed reported a large degree of satisfaction with open

source including increased stability, scalability and cost effectiveness Issues

with Open Source Software include manageability, its quality of support and

ease of use

1.6 Contribution

This research examines the history of software from the beginnings of the

United States software industry to the current market for Open Source

Software licences and Closed Source Software Licences Informed by a prior

study, this research has developed a framework of Open Source Software

Adoption leading organisations to adopt or not to adopt open source software

The prior study suggests a relationship between External Organisational

Factors and External Information Technology Factors and the use of Open

Source Software Organisational comparisons in the analysis of the case study

uncovered the major issue of cost; it was repeatedly mentioned in the

interviews conducted Other considerations such as staffing and competitive

pressures are also revealed

1.7 Outline of the report

This introductory chapter provided a brief background to the study and

presents an overview of the research question and its objectives The scope of

the research as well as the methodology is outlined

Trang 24

The second chapter in this study represents the background review A brief

history of the overall software distribution industry is examined along with its

impact on the current software market

The third chapter presents the literature review A current analysis of Open

Source Software is given including its precise definition, history and specific

examples of its successes Factors leading to its adoption and rejection are

then examined Following this, a framework of Open Source Software

Adoption is proposed and several major factors are selected for further

discussion

The fourth chapter contains the discussion of the methodological techniques

employed for this research It discusses the development and adoption of

testing instruments used in this study; the measures taken to ensure Validity

and Reliability; the Research Design; Sampling and Case selection; the

interview protocol; and, its analysis

The fifth chapter contains a report on the data collected from the applied

research A discussion on the various organisations is presented along with

any other research issues which arose

The sixth chapter contains an analysis on the data collected from the applied

research The results on the prior study are presented and its impact

discussed The organisations are examined individually followed by a cross

case analysis detailing the findings of this research

The seventh chapter provides a summary of this research where the research

question, its design and methodology used are stated, followed by the

framework used and the findings The contribution of this research is then

presented, followed by directions for further research

Trang 25

Trang 26

Background Page 12

2 Background

Before examining the Open Source Software phenomenon, it is first

necessary to have an understanding of the events and circumstances leading

to its development This is achieved through the use of the ‘historical

method’ The historical method comprises the techniques and guidelines by

which historians use primary and secondary sources and other evidence to

research and then to document history Historical reviews are relatively rare

in information system research

This study involves the investigation of factors leading to the adoption of

Open Source Software The purpose of this chapter is to provide a review of

the literature leading to the emergence of Open Source Software It explores

the United States software industry from the 1950s, and through the

development of bundled software, the formation of Software Intellectual

Property It then looks into the limitations of this model, and the need for the

Open Source model It then examines a number of open source successes

being used in the present day The United States was chosen for its past and

current leadership in the global software industry

2.1 History of Propriety Software Development

Computer software prior to the 1960s was seen as mere mathematical

algorithms and not processes or machines (The History of Software Patents

n.d.) During this period, computer manufacturers distributed software freely

as part of their customer service policy (Johnson 1998) In a way, all

computer software in that era could be considered ‘open source’ This

changed when IBM, the leading computer manufacturer at that time,

announced that in June 1969 pricing of software would be separated from its

hardware (Johnson 1998) This ‘unbundling of software’ is seen as the

starting point of the software industry The separation was further reinforced

by the US Supreme Court decision to grant a patent in the case of Diamond v

Trang 27

Diehr (The History of Software Patents n.d) Computer software was thus

recognised as a form of intellectual property

2.1.1 The Birth of the software Industry

The software industry began in the late 1950s, when the use of computers for

business applications expanded rapidly This created a huge demand for

people with programming experience A number of people who had learned

their programming skills while working for computer manufacturers, or for

the large companies and government agencies (who were the first computer

users) saw this as an opportunity to start their own companies and sell their

services under contract (An Overview of the History of the Software Industry

n.d.)

By 1956 transistors began to replace vacuum tubes and memory core

technology improved Smaller, faster, more reliable and increasingly

energy-efficient machines led to improvements in speed Second generation

machines paved the way for the development of innovations such as stored

programs and programming languages (Informatics Press 1993) In turn, this

allowed for the development of computers that were affordable and flexible

enough to be integrated into large companies’ operations

Elmer Kubie and John W Sheldon, two former IBM employees, formed the

first one of these companies producing software, the Computer Usage

Corporation (CUC), in 1955 Their firm was founded with $40,000 in start-up

capital, and supported a staff of five in addition to the two founders Its first

project was to write a program, under contract, for the California Research

Corporation to simulate the flow of oil (Kubie 1994) This became one of the

earliest examples of contract programming where an organisation would

procure its software development from another firm CUC became a public

company in 1960 and by 1967 had a staff of over 700 people in 12 offices

Trang 28

around the U.S and revenues over $13 million Unfortunately, it suffered

financial losses in the late 1970s and eventually went bankrupt in 1986

In 1959, seven Univac programmers founded Applied Data Research (ADR)

to market their programming skills in developing systems software to

computer manufacturers such as Sperry Rand and Honeywell (Johnson 1998)

ADR went public in 1965 and, in the late 1960s, became one of the first

companies to sell software products, such as AUTOFLOW, a program which

automatically produced program flowcharts by reading the program source

code (ADR 1965) In order to protect their intellectual property rights in the

program, it was suggested to them that it would be better to "lease" the

program for three years at a time and call the program "equipment" At the

time contract law was very clear on the limited rights of a party when they

leased equipment It continued to be one of the largest U.S software product

companies until Ameritech acquired it for $215 million in 1986

Fletcher Jones and Roy Nutt, who had gained their computer experience in

the aerospace industry, founded Computer Sciences Corporation (CSC) in

1959 with $100 and a contract from Honeywell to develop a

business-language compiler called FACT By 1963, CSC was the largest software

company with revenues close to $4 million CSC continues to thrive today as

one of the world’s largest information technology services firms with more

than $10.2 billion in revenues (CSC at a Glance n.d)

By 1965, there were an estimated 45 major software contractors in the

U.S.A., some employing more than a hundred programmers and with annual

revenues of as much as $100 million In addition, there were hundreds of

small firms, typically with just a few programmers In 1967, it was estimated

that there were 2,800 software services firms in the U.S.A (An Overview of

the History of the Software Industry n.d.)

Trang 29

2.1.2 Bundled Software

The number of computers in use and their size and speed expanded rapidly in

the 1960s, increasing the demand for software to support the numerous tasks

for which computers were now being used Computers were not used for

making sophisticated calculations but for simplifying extremely large

numbers of simple calculations such as those needed by the Census

Department and by the human resource departments of large corporations

(SIAA 2001) This provided enormous opportunities for entrepreneurs to

create new companies to serve an expanding market

Figure 2.1: Vertically Integrated Mainframe and Minicomputer Industry Structure

Adapted: West, J & Dedrick J 2001 Open source standardization: The rise of Linux in the

network era Knowledge, Technology, & Policy 14(2): 88-112

Vertically integrated proprietary standards architectures were the norm for the

early computer industry By the early 1960s, a customer of any of the major

hardware manufacturers could expect to have access to a library of software,

which was included “bundled” into the cost of a computer (Johnson 1998)

This software included the computer’s operating system and also utility

programs such as sort programs, compilers for languages such as COBOL

and FORTRAN, and a growing library of programs written to handle specific

applications (Informatics Press 1993) IBM, for example, maintained a library

of application programs written by its programmers to meet the needs of

specific customers but these were then made available at no charge to other

IBM customers (An Overview of the History of the Software Industry n.d)

Trang 30

Each computer maker developed most if not all of its technology internally,

and sold that technology only as part of an integrated computer system This

systems era was ascendant from IBM’s 1964 introduction of its System 360

until the 1981 release of IBM’s first ‘Personal Computer’ (Moschella 1997)

During this era of vertically integrated computer systems, the most successful

architecture was IBM’s System 360 (later 370 and 390) line of mainframe

computers IBM produced most of the hardware, including electronic

components, CPU boards and peripherals, as well as the operating system,

tools and much of the application software The 360 marked IBM’s switch

from being a major buyer of electronic components to one of the largest

manufacturers, all for internal use (Sobel 1981) From 1950 to 1980, IBM’s

competitors in its mainframe and minicomputer markets, also sought vertical

integration, although most lacked IBM’s ability to achieve the same level of

integration This unfortunately caused compatibility problems between

systems from different vendors and started becoming a sore point for users in

a heterogeneous computing environment

IBM and its System 360 were fantastically successful in this strategy, at one

point garnering nearing half of the computer industry’s profits (Moschella

1997) A key reason was that the System 360 marked the first widespread

implementation of what Gabel (1987) refers to as “product line

compatibility” Unlike previous generations, IBM shared the same standards

architecture across its product line and thus enabled use of the same operating

system, tools and user applications

Computer users during this period had the choice of getting the software they

needed from their hardware vendor, or of having it custom-built for their

needs by their own programmers or by a contract-programming firm Many

executives in the software industry didn’t believe that there would ever be a

viable market for software products, since it was too difficult to compete

Trang 31

against free software from hardware manufacturers (Johnson 1998; An

Overview of the History of the Software Industry n.d)

But there were some contract programming firms, early in the 1960s, which

began to see opportunities, when there was no comparable product available

from the hardware vendor, to sell programs they had written for more than

one customer For example, CACI began selling SIMSCRIPT, a simulation

language, in 1962 (About CACI: Profile n.d.), and ADPAC Corporation

made several sales of its ADPAC compiler in 1964 to customers who had

seen it used by other firms’ ADPAC programmers and wanted it made

available to their own programmers (ADPAC Corporation Company Heritage

n.d.)

In 1965, Applied Data Research (ADR) released AUTOFLOW which

ultimately sold to thousands of customers (Johnson 1998) In November

1967, Informatics released MARK IV, a generalized file management and

report generation program, which surpassed $1 million in revenues within 12

months of its formal launch (An Overview of the History of the Software

Industry n.d)

ADR and Informatics, both very successful contract-programming firms,

were the first companies to set up the kind of organizations required to

market and support software products and thereby became true software

product companies

In the late 1960s, the concept of software as a product began to take hold,

despite an environment where customers were used to getting their software

free of charge In January 1967, International Computer Programs Inc (ICP)

published the first issue of its ICP Quarterly, a catalogue of software

programs that were available for sale (Welke 1998) Forty-nine programs

were offered in the first issue, but by the end of 1969, each issue of the ICP

Trang 32

Quarterly listed hundreds of software products (also called ‘software

packages’)

But IBM was forced to limit its vertical integration in the face of a 1969

federal anti-trust lawsuit To respond to one of the charges, that of illegal

“tying” of its mainframe hardware and software, five months later IBM

decided to end its decades-long practice of bundling its offerings allowing

companies to buy the hardware, software and services separately On June 30,

1969, IBM announced that effective from January 1, 1970, it would begin to

unbundle (charge separately for) some of its software, effectively ending the

expectation of its customers that they would always be able to get all the

software they needed from IBM free of charge (Johnson 1998)

As hoped by the government, this unbundling encouraged the third-party

supply of peripherals, software and support Still, many IBM customers

bought a complete IBM solution This helped IBM keep customers because

those customers who had custom-developed software would find it more

expensive to adapt their software to a new processor, operating system or

compiler.As Greenstein’s (1997) study of government procurement showed,

such strategies helped discourage switching between companies At the same

time, IBM’s vertical integration strategy had its limitations as an exemplar for

the rest of the industry Without IBM’s economies of scale and scope, rivals

were eventually forced to choose between inferior performance on key

dimensions, obtaining key components from outside suppliers, or exiting the

computer business

The 1970s saw the contract-programming industry continue to grow at a rapid

pace Companies within this industry came to be known as ‘professional

services’ firms reflecting the fact that they often provided a broad range of

consulting, analysis and design services in addition to programming (SIAA

2001) The software products industry became firmly established as a viable

source of software for computer users If, at the beginning of the 1970s,

Trang 33

customers were sceptical that software purchased from a vendor could meet

their needs as well as software written in-house, by the end of the decade

almost all computer users were buying some portion of their software from

software products companies As a result of unbundling, the hardware

vendors also became major players in the software products field (Johnson

1998)

Unbundling could be seen as the breakthrough of the proprietary software

industry The industry was further legitimised in 1981 in the case of Diamond

v Diehr when the U.S Supreme Court ordered the Patent and Trademark

Office to grant a patent on an invention even though computer software was

used (History of Software Patents n.d.) This gave computer software

manufacturers legal rights of a proprietary nature, protecting them from the

unlawful reproduction, modification and distribution of their products

Companies now protected their software through the use of intellectual

property laws They also distributed the executable code, which could only be

read by the specific machine it was designed for, versus the source code,

which could be read by human programmers (Murray 2000) This concept is

similar to a bakery manufacturing bread for its customers without disclosing

the recipe or its ingredients

2.1.4 Software Crisis

Resulting from the failure to deliver software products in time, computer

users were by the end of the 1970s facing a “software crisis” Software

development projects were consistently over schedule and over cost A large

percentage of programs could not be used or were never delivered This is

despite the fact that experts from academia, industry and research

laboratories, gathered for a NATO Conference in 1968, had attempted to

provide future directions for software development (Gibb 1994) They had

recommended the use of better system design methodologies The problems

Trang 34

of late delivery and cost overruns persisted throughout the 1970s and 1980s,

notwithstanding such new methodologies as Information Engineering,

Computer-Aided Software Engineering (CASE) and Rapid Applications

Design (RAD) Perhaps the most profound consequence of the conference

was the beginning of the use of open architectures in which different software

programs could interact with one another This led to the widespread

adoption of more flexible approaches to software development and

distribution, particularly to outsourcing development, custom software

developed by an external body for the specific needs of the organisation, and

the use of commercial off-the-shelf-software (COTS), a standard type of

software package that is marketed to a number of organisations (Hoffer,

George & Valacich 1998)

2.1.5 Standards and the need for interoperability

The vertically integrated segmented strategy, prevalent during the 60s and

70s, used in the computer industry assumed proprietary (single-vendor)

control of one or more standards in a system’s architecture The differences

between these proprietary approaches (and with subsequent “open”

standards) highlight the conflict between a firm’s competing objectives in

standards competition: adoption and appropriability

Figure 2.2: PC Industry Structure

Adapted: West, J & Dedrick J 2001 Open source standardization: The rise of Linux in the

network era Knowledge, Technology, & Policy 14(2): 88-112

Trang 35

Widespread adoption of the standard is important because the CPU, operating

system and key application tools such as compilers and databases are more

widely valued by users if they have a wide range of compatible software

Thus, to get the widest variety of software built upon their respective layer of

the architecture, sponsors of a standard used pricing and other incentives to

gain the largest number of early users and encourage the development of

co-specialized software (Teece 1986; Morris & Ferguson 1993; Shapiro, Carl &

Varian 1999)

At the same time, success of the standard provides no guarantee of the

sponsor’s ability to profit from that standard That ability will depend on

intellectual property protection for the standard, the relative importance of the

standard and complementary assets (Teece 1986) Control over the

architecture enables the technological evolution of the standard and provides

the incentive for the necessary investments (Morris & Ferguson 1993)

However, in the increasingly common case of divided control of a standard

architecture, control of the Application Programmer Interfaces (APIs)

determines access to the software and thus the right to profit from a standard

(West & Dedrick 2000)

So the sponsors of successful standards face competing incentives, as

illustrated in the 1980s PC industry They can gain wide adoption of their

standard, at the risk of not profiting from its success (as IBM did) Or they

can control the profit from the standard, at the risk of limited adoption and

potential abandonment

2.2 Emergence of Open Source Software

Software development through these approaches was not a feasible or

meaningful option for some organizations This created a need for a new

development paradigm Software typically accounts for 17% of the amount

Trang 36

spent to implement and maintain a computer system (Christiansen &

Cosgrove 1998) The reliability of software is still in question: major vendors

continue to create and release software packages with large numbers of

package’s source code, limiting the amount of customisation (Evers 2000)

The cost of commercial software, along with its development time, also

becomes an issue As a result of this, the Open Source model for software

development has gained increasing recognition and acceptance during the

mid 90s

2.2.1 History of Open Source Software

It was argued that Open Source development started with the Unix operating

system, the first operating system to be used on more than a single hardware

platform (Evers 2000; Murray 2000; The Open Source Initiative n.d)

However, software developed before Unix could also be considered open

source (Economist 2001) This was confirmed by Johnson (1998) in “An

Overview of the History of the Software Industry” (n.d), which mentions the

use of shared software libraries, where a variety of software applications were

made available to customers free of charge, established by IBM as part of its

customer support function until its decision to ‘unbundle’ software from its

hardware and before the recognition of computer software as intellectual

property As software was informally shared then, we could see Unix as the

formal start of the establishment of the open source movement and the

development of legal licences to protect open source users, developers and

distributors

Prior to Unix, computer systems did not communicate well with one another

(SIAA 2001; Severance 1995) Various computer lines manufactured by the

same company often needed interpreters to share data and there was

practically no interaction between machines developed by different vendors

(Rosenberg 2000) Operating systems often performed only limited tasks, and

Trang 37

only on the machines for which they were written If a business upgraded to a

bigger, more powerful computer, the old operating system might not operate

on the new computer, and often the company's software needed to be

re-written and the data had to be re-entered This created problems especially

during the mid 60s on the development of the Advanced Research Projects

Agency (ARPA) network, predecessor to the Internet, due to the need of

interoperability (Evers 2000; Murray 2000; The Open Source Initiative n.d)

In order to develop a convenient, interactive, useable computer system that

could support many users, a group of computer scientists from Bell

Laboratories Computer Science Research Center and GE in 1965 joined an

effort called Multiplexed Information and Computing Service or Multics,

then underway at MIT (Bell Labs 2000)

Multics was able to achieve its objective of providing for multiple users on a

single computer, but, over time, the effort failed to produce an economically

useful system (Ritchie 1996) Bell Labs withdrew from the effort in 1969 but

a small group of users at Bell Labs Computing Science Research Center in

Murray Hill Ken Thompson, Dennis Ritchie, Doug McIlroy, and J F

Ossanna, refused to drop the project (Bell Labs 2000)

In early spring of 1969, Thompson felt that it was obvious that Multics was

becoming redundant Consequently, he wrote an operating system of his own,

thereby creating an environment for doing future work He needed a system

that could separate different files from different users and thus developed the

idea of nodes and block addresses (Bell Labs 2000) He discussed the idea

with his fellow researchers and developed the working document for the Unix

file system

During the summer of 1969, Thompson first worked out the requirements for

an operating system, in particular the notion of processes He then developed

a small set of user-level utilities: the means to copy, print, delete and edit

Trang 38

files He also developed a command interpreter, or shell (Bell Labs 2000;

Ritchie 1996)

It soon became obvious that the development machine, a PDP 7 which the

UNIX group didn't own, was becoming obsolete In 1970, they proposed

buying a PDP-11 for about $65,000 (Bell Labs 2000; Ritchie 1996) Two

research department heads, Doug McIlroy and Lee McMahon, realized the

benefits of the new operating system and supported the proposal (Bell Labs

2000; Ritchie 1996) The PDP-11 arrived at the end of the summer

In 1971, Thompson first tried, to use Fortran to develop Unix on the PDP-7

but gave up after the first day (Bell Labs 2000; Ritchie 1996) He then wrote

a very simple language he called ‘B’, which he used on the PDP-7 It worked,

though there were efficiency problems Firstly, the implementation was

interpreted, making the execution slow Secondly, the basic notions of B were

based on the word-oriented BCPL, and therefore required translation to a

byte-oriented machine such as the PDP-11 (Severance 1995)

The ‘C’ programming language had recently been developed by Thompson

and a decision was made to completely re-write the operating system in C,

which is a compiled language As most operating systems were written in

Assembly, they had to be re-written from scratch for each and every hardware

platform (Evers 2000) By writing the entire operating system in C, only the

compiler portion of the operating system needed to be re-written for every

hardware platform This made Unix the first source-portable operating

system, a crucial advantage over competing operating systems (Severance

1995)

The first customer was the Bell Labs Patent Department, which was

evaluating a commercial system to prepare patent applications (Salus 1994)

In developing UNIX to support text processing, the Computing Science

Trang 39

Research Center supported three Patent Department typists who could type,

edit, and format patent applications on a single computer

In 1976 - 77, Ken Thompson took six-months leave from Bell Labs to teach

UNIX as a visiting professor at the Computer Science Department at the

University of California-Berkeley (UCB) (Salus 1994) As a result, the Unix

system became very popular throughout the academic community

After Thompson returned to Bell Labs, students and professors at Berkeley

continued to enhance UNIX (Kesteloot 1995) Eventually, many of these

enhancements were incorporated into what became known as Berkeley

Software Distribution (BSD) Version 4.2, which many other universities also

acquired (Kesteloot 1995)

Although UNIX is an ad-hoc development, it provides a powerful and

flexible environment that is useful for businesses, sciences and academia

Many telecommunications switches and transmission systems are controlled

by administration and maintenance systems based on UNIX (Salus 1994)

Initially designed for medium-sized minicomputers, the operating system was

soon moved to larger, more powerful mainframe computers (Severance

1995) As personal computers grew in popularity, versions of UNIX found

their way into these machines, and a number of companies began to produce

UNIX-based machines for the scientific and programming communities

(Salus 1994)

The writing of programs, which run on UNIX for commercial use, soon

developed into new market opportunities (Salus 1994) Unix became popular

for business applications due to its timesharing, multitasking capability,

permitting many people to use a mini- or mainframe computer; additional

benefits were its portability across different vendor's machines, and its e-mail

capability (Salus 1994)

Trang 40

UNIX was distributed through academic licenses, which were relatively

inexpensive, and government and commercial licenses from about 1975 The

University of California, Berkeley became important in popularising UNIX

when it established a Computer Systems Research Group (CSRG), originally

under the direction of Robert Fabry (Kesteloot 1995) The CSRG did much of

the formation of the TCP/IP protocols, which are the foundations of the

Internet, freely accessible with their BSD distributions under the BSD

Licences (Working Group on Libre Software 2000)

In addition, UNIX played a key role in the early days of the Internet, since

most of the VAX computers supporting the Internet ran on UNIX (Newman

1999) As UNIX and Open Source Software spread through the academic

world, businesses eventually became aware of UNIX from newly hired

programmers who had used it in college

As the use of Unix increased, several foundations were established to protect

the Open Source movement (Evers, 2000; The Open Source Initiative n.d)

The concerned groups in 1988 formed a special interest group, the Open

Systems Foundation (OSF), to lobby for an "open" UNIX within the UNIX

community (Bell Labs 2000) Soon several large companies, who at the time

were promoting their own proprietary operating systems in competition with

UNIX, also joined the OSF

Today UNIX is the operating system for most large Internet servers,

businesses and universities, and a major segment of academic and industrial

research in operating systems is based on UNIX Most commercial software

is written in C or C++, a direct descendant of C that was also developed at

Bell Labs, or more recently Java, a C++ descendant developed at Sun

Microsystems (Lévénez 2001)

Định dạng
Số trang	390
Dung lượng	4,22 MB