A Delphi Study of the Influences on Innovation Adoption and Process
Total Page:16
File Type:pdf, Size:1020Kb
Department of Computer Science and Information Systems College of Informatics and Electronics A Delphi study of the influences on innovation adoption and process evolution in a large open-source project The case of Debian Martin F. Krafft, B.A.hons Lero — the Irish Software Engineering Research Institute <[email protected]> — http://phd.martin-krafft.net A dissertation submitted for the degree of Doctor of Philosophy at the University of Limerick, Ireland Under the supervision of Prof. Dr. Brian Fitzgerald and Dr. Kieran Conboy 6 Apr 2010 ii To infinity, and beyond! iv Contents List of tables x List of figures xii Abstract xiii Originality note xiv Version history xv Copyright & Licence xvi Acknowledgements xvii 1 Introduction 1 1.1 Motivation . 1 1.2 Research question and objectives . 4 1.3 Outline of the thesis . 7 I Background and literature review 2 Debian 10 2.1 Introducing Debian . 10 2.2 The Debian community . 11 2.2.1 Project members and contributors . 12 2.2.2 Debian governance . 13 2.2.3 The role of volunteers . 14 2.2.4 Stereotypical traits of Debian developers . 15 2.2.5 Freedom and independence . 20 2.2.6 Debian derivatives and Ubuntu . 22 v Contents Krafft: Ph.D. dissertation 2.2.7 Communication basics . 23 2.2.8 Relation to organisational theory . 25 2.3 Debian packaging . 32 2.3.1 Tools, techniques, processes, approaches, solutions, and methods . 32 2.3.2 The Debian packaging workflow . 33 2.4 Process improvement and volunteers . 35 2.5 Previous research on Debian . 37 3 Assessing diffusions 39 3.1 History and terminology . 39 3.1.1 Innovation and invention . 40 3.1.2 Diffusion, adoption, and critical mass . 42 3.1.3 Adoption process and innovation stages . 43 3.1.4 Network externalities and excess inertia . 46 3.2 Diffusion frameworks . 46 3.2.1 In search of a framework . 48 3.2.2 An overview of existing frameworks . 50 3.2.3 A framework for influences to adoption behaviour in the Debian Project 58 3.3 Related studies . 59 II Research approach 4 Introduction 61 4.1 Philosophical considerations . 63 4.1.1 Positivism vs. interpretivism . 64 4.1.2 Research dichotomies . 68 4.2 Research objectives . 70 4.3 Secondary objectives . 70 4.3.1 Sub-objective 1: Data from the Delphi study . 70 4.3.2 Sub-objective 2: The Delphi method in the FLOSS context . 71 5 The Delphi method 73 5.1 The Delphi process . 74 5.2 Suitability for this research . 75 5.2.1 Delphi and asynchronous communication in FLOSS . 76 5.2.2 Benefits of controlled feedback in a FLOSS context . 78 5.2.3 Reasons for a Delphi approach . 80 5.3 Variants and modifications . 81 5.4 Design considerations . 83 5.4.1 Choice of the specific Delphi method . 84 5.4.2 “Experts” . 85 5.4.3 Participant selection . 85 vi Krafft: Ph.D. dissertation Contents 5.4.4 Panel size . 89 5.4.5 Degree of anonymity . 90 5.4.6 Communication medium . 91 5.4.7 Consensus vs. disagreement . 94 5.4.8 Question design . 95 5.4.9 Participant dropout and non-response . 96 5.4.10 Bias . 99 5.5 Sending e-mails to Debian contributors . 100 5.5.1 Avoiding the bulk-mail character . 100 5.5.2 Dealing with spam filters . 101 5.6 Deciding against a mock study . 102 6 Research approach details 105 6.1 Details of the participant selection process . 106 6.1.1 Picking nominators . 106 6.1.2 Calling for nominations . 107 6.1.3 Soliciting applications to the panel . 109 6.1.4 Questioning potential participants . 111 6.1.5 Selecting the panellists . 112 6.1.6 Declining applications . 115 6.2 Details of the Delphi study . 116 6.2.1 The first round: brainstorming influences . 116 6.2.2 Wrapping up the first round . 119 6.2.3 Preparing the second round . 120 6.2.4 Round two: discussing influences . 124 6.2.5 Wrapping up round two . 125 6.2.6 Preparing the third and final round . 126 6.2.7 Round three: identifying salient influences . 129 6.2.8 Design of the third-round email . 130 6.2.9 Wrapping up round three . 131 III Analysis 7 Results 134 7.1 Framing the results . 136 7.1.1 Individual and organisational adoption timelines . 136 7.1.2 A stage model for adoptions in the Debian Project . 138 7.2 Influences to Debian Contributors’ adoption decisions . 142 7.2.0 On the role of pioneers . 146 7.2.1 Stage one: knowledge . 147 7.2.2 Stage two: individual persuasion . 168 7.2.3 Stage three: individual decision . 184 7.2.4 Stage four: individual implementation . 195 vii Contents Krafft: Ph.D. dissertation 7.2.5 Stage five: organisational adaptation . 206 7.2.6 Stage six: organisational acceptance . 216 7.2.7 Stage seven: use-performance-satisfaction, individual confirmation . 221 7.2.8 Stage eight: incorporation . 223 7.2.9 Stage nine: organisational initiation . 231 8 Conclusion 236 8.1 Research contribution . 236 8.1.1 Influence salience . 238 8.1.2 A stage model of adoption in the Debian Project . 239 8.1.3 Specificity of results and stage model . 240 8.1.4 Secondary research objectives . 246 8.2 Implications for practice . 247 8.3 Review of the research approach . 250 8.3.1 Bias . 251 8.3.2 Design issues . 258 8.4 Review of the design . 266 8.4.1 Intuition and self-involvement . 266 8.4.2 Compensation for participation . 268 8.4.3 Feedback from the panellists . 269 8.5 Recommendations for future research . 276 8.5.1 Comprehensiveness and salience of influences . 277 8.5.2 Longitudinal diffusion studies . 278 8.5.3 Executive decision-making . 279 8.5.4 Consequences of innovation . 279 8.5.5 Investigation of single stages . 279 8.5.6 Needs and technology . 280 8.5.7 Information flow and channel effectiveness . 281 8.5.8 Revolutions and evolutions . 282 8.5.9 Corporate affiliation and credibility . 282 8.5.10 Elegance . 283 8.5.11 Level of abstraction . 283 8.5.12 On-line diffusion theory . 284 Bibliography 287 A Background information on Debian 306 A.1 Terminology . 306 A.1.1 System . ..