A STUDY OF LANGUAGE-ACTION PERSPECTIVE AS A THEORETICAL FRAMEWORK FOR WEB SERVICES

Abstract

A science of services is contingent on descriptive and normative theoretical contributions. This dissertation investigates the appropriateness of the language-action framework as a theoretical source for web services, the technological component of a science of services.

Web service supports interoperability among applications independent of their platform, through a set of standards. Current web service efforts are hindered by complexities created by the proliferation of standards which are sometimes competing and/or overlapping. These complexities undermine the usability of web services and therefore hinder widespread adoption. These complexities can be managed by developing boundaries for each web service aspect and new design methodologies and principles that are tailored towards web services. The Language Action Perspective (LAP) is a candidate theoretical perspective for addressing these concerns because it follows the basic premise of web services—supporting communication among disparate applications to achieve business goals.

The central hypothesis of this dissertation is that the Language-Action Perspective (LAP) would be an appropriate theoretical framework for web services. This hypothesis is tested through three inter-dependent essays. The first essay investigates whether LAP constructs can describe and explain the web services architecture. Findings from this essay reveal a need for conversation mechanisms to guide interactions among web services. The second essay responds to this finding. The domain selected for this investigation is enterprise integration. This essay develops a mechanism for accessing design knowledge with LAP constructs. The outcome is translated into web service conversation specifications. As proof of concept, a research prototype that uses a LAP based mechanism to generate web service conversation specifications as integration solutions is also developed. The third essay assesses the contribution of LAP constructs for designing effective integration solutions. The contribution of LAP constructs were assessed by designing and carrying out a controlled experiment. The assessment shows that LAP based mechanism produces higher quality integration solutions and reduces designer effort required for developing solutions.

Significant contributions from this dissertation include a LAP-inspired reference framework for web services (essay 1), a methodology and software artifact for designing enterprise integration solutions as conversations among web services (essay 2), and an evaluation of the relative effectiveness of LAP constructs for designing web service solutions (essay 3). Together they test the central hypothesis—the appropriateness of LAP as a theoretical framework for web services.

 

Table of Contents

LIST OF FIGURES…………………………………………………………………………………………………….. vii

LIST OF TABLES……………………………………………………………………………………………………… viii

ACKNOWLEDGEMENTS……………………………………………………………………………………………. x

CHAPTER 1  INTRODUCTION…………………………………………………………………………………… 1

1.1 INTRODUCTION……………………………………………………………………………………………………….. 2

1.2 PROBLEM DEFINITION……………………………………………………………………………………………… 2

1.2.1  Theories used in the web service domain……………………………………………………………. 3

1.2.2  Need for theoretical framework………………………………………………………………………… 5

1.3 A KEY CHALLENGE…………………………………………………………………………………………………. 7

1.4 MOTIVATION………………………………………………………………………………………………………….. 7

1.5 SCOPE……………………………………………………………………………………………………………………. 8

CHAPTER 2  LITERATURE REVIEW………………………………………………………………………… 9

2.1 WEB SERVICES……………………………………………………………………………………………………… 10

2.1.1  Service-Oriented Architecture………………………………………………………………………… 11

2.1.2  Web Service Technologies………………………………………………………………………………. 12

2.1.3 Service interactions………………………………………………………………………………………… 15

2.1.4 Web service complexity…………………………………………………………………………………… 21

2.2 LANGUAGE-ACTION PERSPECTIVE (LAP)…………………………………………………………………. 22

2.2.1  Speech Act Theory…………………………………………………………………………………………. 24

2.2.2  LAP-based applications…………………………………………………………………………………. 27

2.2.3  Criticisms on LAP approach…………………………………………………………………………… 36

2.3 THE SHIFTING PARADIGM AND ITS ISSUES…………………………………………………………………. 39

2.3.1 The LAP approach for web services…………………………………………………………………. 40

CHAPTER 3  RESEARCH QUESTIONS AND APPROACH………………………………………. 42

3.1 RESEARCH QUESTIONS…………………………………………………………………………………………… 43

3.2 RESEARCH METHODOLOGY…………………………………………………………………………………….. 44

3.2.1  First Essay: Conceptual Development and Logical Arguments…………………………… 44

3.2.2  Second Essay: Design Science Research Framework………………………………………… 45

3.2.3  Third Essay: Controlled Experiment……………………………………………………………….. 47

CHAPTER 4  A THEORETICAL INVESTIGATION OF THE EMERGING STANDARDS FOR WEB SERVICES………………………………………………………………………………………………………… 48

4.1 INTRODUCTION……………………………………………………………………………………………………… 49

4.2 WEB SERVICES AS THE REALIZATION OF SERVICE-ORIENTED COMPUTING………………………. 51

4.2.1 Alternative instantiations………………………………………………………………………………… 51

4.2.2 Commonalities across instantiations………………………………………………………………… 53

4.3 A THEORETICAL FRAMEWORK FOR THE WEB SERVICES STANDARDS SPACE…………………….. 55

4.3.1  A brief review of research on the language-action perspective……………………………. 57

4.3.2  An LAP-inspired framework for the web services standards space………………………. 58

4.4 ASSESSMENT OF EXISTING WEB SERVICES STANDARDS STACKS…………………………………….. 63

4.4.1  Assessment of the W3C initiative…………………………………………………………………….. 63

4.4.2 Assessment of the semantic web services initiative……………………………………………… 65

4.4.3 Assessment of the ebXML initiative…………………………………………………………………… 67

4.5 DISCUSSION………………………………………………………………………………………………………….. 70

CHAPTER 5  ACCESSING DESIGN KNOWLEDGE WITH SPEECH ACTS TO GENERATE WEB SERVICE CONVERSATION SPECIFICATIONS………………………………………………………. 74

5.1 INTRODUCTION……………………………………………………………………………………………………… 75

5.2 DESIGN KNOWLEDGE……………………………………………………………………………………………… 76

5.2.1 Patterns as design knowledge………………………………………………………………………….. 77

5.2.2 Enterprise integration patterns………………………………………………………………………… 78

5.3 ACCESSING ENTERPRISE INTEGRATION PATTERNS………………………………………………………. 79

5.3.1  Speech acts as mediator…………………………………………………………………………………. 80

5.3.2  Components facilitating mapping……………………………………………………………………. 82

5.3.3  Inferring mapping between task and integration patterns with speech acts…………… 86

5.3.4 Surfacing conflicts among integration patterns………………………………………………….. 89

5.3.5 Demonstration through an example………………………………………………………………….. 90

5.4 ASSEMBLING INTEGRATION PATTERNS INTO WEB SERVICE SOLUTIONS…………………………… 93

5.4.1 Generating conversation specifications…………………………………………………………….. 95

5.5 RESEARCH PROTOTYPE…………………………………………………………………………………………… 97

5.6 VALIDATION AGAINST DESIGN SCIENCE RESEARCH GUIDELINES………………………………….. 100

CHAPTER 6  EMPIRICAL EVALUATION………………………………………………………………. 103

6.1 NEED FOR EMPIRICAL EVALUATION………………………………………………………………………… 104

6.2 RESEARCH HYPOTHESIS………………………………………………………………………………………… 105

6.2.1 Design support for accessing integration patterns……………………………………………. 106

6.2.2 Dealing with complexity………………………………………………………………………………… 108

6.3 EXPERIMENTAL DESIGN……………………………………………………………………………………….. 112

6.3.1  Independent variables………………………………………………………………………………….. 112

6.3.2  Dependent variables…………………………………………………………………………………….. 113

6.4 SUBJECT SELECTION…………………………………………………………………………………………….. 115

6.5 EXPERIMENTAL PROCEDURES………………………………………………………………………………… 117

6.6 PILOT STUDY………………………………………………………………………………………………………. 118

6.7 FINAL EXPERIMENT……………………………………………………………………………………………… 119

6.8 DATA COLLECTION………………………………………………………………………………………………. 120

6.9 ANALYSIS…………………………………………………………………………………………………………… 121

6.9.1 Assumptions of MANOVA……………………………………………………………………………… 121

6.9.2 Canonical correlation test……………………………………………………………………………… 125

6.9.3  Bivariate correlation test……………………………………………………………………………… 127

6.9.4 MANOVA test for hypothesis 1 and 2………………………………………………………………. 128

6.9.5 MANOVA for hypothesis 3 and 4……………………………………………………………………. 133

6.10 DISCUSSION………………………………………………………………………………………………………. 137

CHAPTER 7  CONCLUSION……………………………………………………………………………………. 141

7.1 SYNOPSIS……………………………………………………………………………………………………………. 142

7.2 INFORMATION-TECHNOLOGY-PEOPLE (ITP) PERSPECTIVE…………………………………………. 144

7.3 CONTRIBUTIONS AND OUTCOMES…………………………………………………………………………… 146

7.3.1  Contributions to research……………………………………………………………………………… 147

7.3.2 Contributions to practice………………………………………………………………………………. 148

7.3.3 Contributions to the I-discipline…………………………………………………………………….. 148

7.4 LIMITATIONS AND FUTURE RESEARCH OPPORTUNITIES………………………………………………. 149

7.5 CONCLUSION………………………………………………………………………………………………………. 151

REFERENCES………………………………………………………………………………………………………….. 152

APPENDIX A PROPOSED WEB SERVICE STANDARDS……………………………………….. 173

APPENDIX B EXPERIMENTAL INSTRUCTIONS………………………………………………….. 176

APPENDIX C QUESTIONNAIRE…………………………………………………………………………….. 186

APPENDIX D BUILDING ENTERPRISE INTEGRATION SOLUTIONS WITH ……………

PATTERNS……………………………………………………………………………………………………………….. 189

APPENDIX E ENTERPRISE INTEGRATION SCENARIOS…………………………………….. 191

 

List of Figures

Figure 1.1.	Theories used……………………………………………………………………………………………. 4
Figure 1.2.	Challenge in using LAP as theoretical framework for web services…………………. 7
Figure 2.1.	Service-Oriented Architecture (SOA)…………………………………………………………. 12
Figure 2.2.	Simple conversation for action (taken from (Winograd & Flores, 1986), pg 65) 28
Figure 2.3.	Business Action Theory Framework (modified based on (Goldkuhl, 1998) ) ….. 30
Figure 3.1.	An overview of the research approach………………………………………………………… 43
Figure 4.1.	Comparing alternative instantiations with the travel agent scenario……………….. 54
Figure 5.1.	Request-Reply pattern………………………………………………………………………………. 78
Figure 5.2.	Mapping sets of tasks against integration patterns ……………………………………….. 82
Figure 5.3.	Enterprise integration patterns as sequence of speech acts…………………………….. 85
Figure 5.4.	Conceptual model of the OWL knowledge base ………………………………………….. 87
Figure 5.5.	Business process diagram for customer order processing ……………………………… 91
Figure 5.6.	Simplified skeleton of ceipXML specification…………………………………………….. 95
Figure 5.7.	Mapping between ceipXML and cpXML specifications……………………………….. 96
Figure 5.8.	Partial ceipXML and cpXML mapping for SCM scenario…………………………….. 96
Figure 5.9.	Process for developing integration solutions……………………………………………….. 97
Figure 5.10.	Architecture of the Integration Designer Assistant……………………………………….. 98
Figure 5.11.	Business process diagram editor interface…………………………………………………… 98
Figure 5.12.	Integration Designer Assistant…………………………………………………………………. 100
Figure 5.13.	Summary of research effort in accordance to design science guidelines………… 102
Figure 6.1.	Conceptual model of hypotheses to be tested…………………………………………….. 111
Figure 6.2.	Plot of ordered squared Mahalanobis distances………………………………………….. 125
Figure 7.1.	Contributions of this dissertation from ITP perspective………………………………. 145

 

List of Tables

Table 4.1	An LAP-inspired framework for the web services standards space ……………………. 62
Table 4.2	Assessing the W3C initiative against the LAP-inspired framework …………………… 63
Table 4.3	Assessing the Semantic Web Services-initiative against the LAP-inspired framework
	………………………………………………………………………………………………………………….. 66
Table 4.4	Assessing the ebXML-based initiative against the LAP-inspired framework………. 68
Table 4.5	Assessment of the three standards initiatives against the LAP-inspired framework 72
Table 5.1	Parsimonious set of speech acts…………………………………………………………………….. 83
Table 5.2	Mapping between speech acts and action types……………………………………………….. 86
Table 5.3	Partial set of enterprise integration pattern inference rules ……………………………….. 89
Table 5.4	Speech act sequence for identified interactions……………………………………………….. 92
Table 5.5	Integration patterns for identified interactions…………………………………………………. 92
Table 5.6	Design-Science Research Guidelines (Taken from (Hevner et al., 2004))…………. 101
Table 6.1	Summary of factors under study………………………………………………………………….. 111
Table 6.2	Design of Experiment for evaluating utility of the LAP constructs ………………….. 112
Table 6.3	Design error assessment scheme………………………………………………………………….. 115
Table 6.4	Subject problem and design support assignment groups…………………………………. 120
Table 6.5	Descriptive statistics ………………………………………………………………………………….. 122
Table 6.6	Descriptive statistics for Box-Cox transformed dependent variables………………… 123
Table 6.7	Descriptive statistics for LOG transformed dependent variables ……………………… 124
Table 6.8	Canonical correlation test results…………………………………………………………………. 126
Table 6.9	Correlation matrix among dependent variables……………………………………………… 127
Table 6.10	Correlation matrix among dependent variables with composite variable……….. 128
Table 6.11	Descriptive statistics for design support levels as independent variable………… 129
Table 6.12	Levene’s test of equality of error variances with design support levels as

independent variable…………………………………………………………………………………………….. 129Table 6.13Effects of design support on design error and design effort…………………………. 130

Table 6.14            Univariate test for dependent variables with design support levels as independent

variable …………………………………………………………………………………………………………….. 131 Table 6.15                 Comparison of means for dependent variables with design support levels as

independent variable…………………………………………………………………………………………….. 132

Table 6.16                            Descriptive statistics for design support and problem complexity levels as

independent variable…………………………………………………………………………………………….. 133

Table 6.17                       Levene’s test of equality of error variances with design support and problem

complexity levels as independent variable ………………………………………………………………. 135

Table 6.18                    Effects of design support and problem complexity on design error and design

effort       …………………………………………………………………………………………………………….. 135

Table 6.19                               Univariate test for dependent variables with design support and problem

complexity levels as independent variable ………………………………………………………………. 136Table 6.20Research hypothesis results summary ………………………………………………………. 138

Table 7.1       Contributions…………………………………………………………………………………………….. 146

Table 7.2       Opportunities for further research………………………………………………………………… 149

Chapter 1

 

1.1       Introduction

A distinct shift in the economy has begun since the last industrial revolution that brought us large-scale manufacturing industries. Two-thirds of global economic activity and eighty percent of the United States economy is engaged in the service sector (ABCDoha, 2006). Information technology is the catalyst for this shift (Applegate, 2004; Chesbrough & Spohrer, 2006). This economic shift is pushing organizations to fundamentally transform their businesses, requiring changes to both business models and underlying technology (IBMSummit, 2004; Umapathy, 2007).

 

A critical technology platform required for coping with this shift is Service-Oriented Computing (SOC) (Papazoglou, Traverso, Dustdar, Leymann, & Krämer, 2005). The fundamental element to develop software applications in the SOC paradigm is “service” (Curbera, Khalaf, Mukhi, Tai, & Weerawarana, 2003). Services are autonomous, platform-independent computational entities which can be combined in numerous ways to achieve business goals. The current realization of the SOC concept is web services (Weerawarana, Curbera, Leymann, Storey, & Ferguson, 2005). They provide interoperability among applications from different platforms and loose coupling of disparate systems through standardized mechanisms (Curbera et al., 2003; Papazoglou & Georgakopoulos, 2003). Web services support interoperability among applications through a set of standards.

 

1.2       Problem Definition

The efforts of the web services community to develop interoperable standards are coordinated by industrial consortiums (Peer, 2002). These consortiums use processes that were specifically designed to promote industry consensus and unite disparate web service standardization efforts (Graham, Pollock, Smart, & Williams, 2003). Through these processes, most major software companies heavily influence current web service efforts (Heywood, Jander, Roberts, & Saunders, 1997; Konuru & Mukhi, 2003; Virili, 2003). Therefore, current web service efforts have mainly been ad hoc (Lastra & Delamer, 2006). These efforts focus on a particular set of features without having a complete view of web services (Mukhi, Plebani, SilvaLepe, & Mikalsen, 2004). As a result, these efforts lack holistic view of web services and are not supported by any theoretical framework.

 

1.2.1     Theories used in the web service domain

I examined 640 research articles published in the proceedings of the International Conference on

Web Services (ICWS) and the Services Computing Conference (SCC) in the past three years. This examination was conducted with the objective of identifying theories that are being used to address web service problems. The examination revealed that 57 articles clearly identified theories that the authors used to address specific web service problems. A total of twenty-one theories were applied to address sixteen concerns related to web services, which are listed below. Automata theory, Petri net theory, and Set theory are the most highly used theories in the web service domain. Figure 1.1 shows the percentage of articles for each theory.

 

• Automata Theory
• Petri Net
• Set Theory
• Game Theory
• Queuing Theory ƒ Process Algebra
• Markov Decision Processes
• Control-flow complexity
• Sketch Theory
• Graph Transformation
• Situation Calculus

 

• Linear Logic
• Graph Theory
• IS Innovation Theory
• Controlled Language
• Case-based Reasoning
• Co-algebra
• Computational Learning Theory
• Regular Expression Language
• Progressive Auction
• Resource-based view

 

 

Figure 1.1.Theories used

 

This examination reveals that the most web service researchers use formal theories to address specific web service aspects. These formal theories provide systematic approaches to address specific web service problems. This examination strengthens the argument of Mukhi et al. (Mukhi et al., 2004) that current web service efforts focus on specific web service aspects and lack a holistic view. Due to the narrow scope of formal theories they are not appropriate to provide holistic view for web services.

 

1.2.2    Need for theoretical framework

Without appropriate guidance from a holistic theoretical framework, the advancement of web services has suffered (Balasooriya & Prasad, 2005). A manifestation of this problem can be seen in the significant confusion web service developers’ face when attempting to choose appropriate standards for the different aspects of web service solutions. For instance, WS-Notification (WSBaseNotification, 2004) and WS-Eventing (WS-Eventing, 2006) are competing standards for notifying the occurrence of an event (Krill, 2006). Without a strong theoretical framework to guide their efforts, developers cannot make informed choices (Hoskin & Harbour, 2005).

 

Another manifestation of this problem is evident from potential strategic missteps in web service standardization efforts (Festa, 2003). For instance, overlapping standards are being proposed, such as WS-BPEL (WS-BPEL, 2005) and WS-CDL (WS-CDL, 2005), and there is often no separation between standards that address different aspects of web services (Sedighi & Johnson, 2004).

 

Another manifestation of this problem is the slow adoption of web services (M. Chen, Chen, & Shao, 2003; Ran, 2003; Senf, 2005). Organizations have been slow in adopting web services, since existing development models and tools are inadequate to surmount web service problems (Bosworth, 2001; Papazoglou & Yang, 2002). In order to overcome this drawback, theoretical frameworks for web services need to be developed and would allow development of theoretically informed solutions to web service problems.

 

Any theoretical framework selected for web services should match with the core principle of supporting communication among disparate applications to achieve business goals (Gottschalk, Graham, Kreger, & Snell, 2002; Papazoglou, 2003). Communication underpins core web service operations such as “publish-find-bind” as well as additional functionalities such as exchanging proposals, negotiating, establishing contract, fulfilling commitments and renewing relationships. A theoretical framework for web services, therefore, should have communication at its core, account for a layered view of the many web service functionalities, and must be directed towards constructing web service solutions that achieve business goals.

 

Prima facie, these requirements are met by the Language-Action Perspective (LAP). The central premise of LAP is that much work in organizations is performed through language that organizational actors use to achieve their goals (Schoop, 2001; Weigand, 2006). LAP recognizes that language is not only used for exchanging information, (as in reports or statements) but also to perform actions (as in promises, orders, requests, and declarations) (Weigand, 2003). LAP emphasizes that such actions should be the foundation for creating effective information systems. This insight has profound implications for the theory and practice of designing systems using web service technologies to support organizational communication and interaction. LAP, with its essential qualities of communication and coordination, can provide an appropriate framework for web services to address its shortcomings.

 Therefore, the hypothesis I investigate in this dissertation is that The Language-Action

Perspective (LAP) is an appropriate theoretical framework for web services.

 

1.3       A Key Challenge

According to LAP, organizations are social systems with actors that communicate to achieve mutually agreed goals (Dietz, 2003). LAP considers information systems to be social systems that are technically implemented to support work in organizations (Goldkuhl & Lyytinen, 1982). As originally formulated, the purpose of LAP is to support human work in organizations (see figure 1.2), in other words, the foundations of LAP are based on human communicative theories. Web services, in contrast, primarily support communication among machines. A major challenge in this research, therefore, is adapting LAP constructs to explain machine to machine communication.

 

Figure 1.2.Challenge in using LAP as theoretical framework for web services

 

1.4       Motivation

The motivation of this study is to contribute to the discourse and search for theoretical foundations for web services. As the organizational computing paradigm shifts away from an object-orientation (Booch, 1993; Welke, 1994) to a service-orientation (Papazoglou & Georgakopoulos, 2003), theoretical foundations are necessary to inform design and development of software applications. The development of theoretical foundations for web services has the potential to greatly accelerate the standardization process, and can provide web service researchers ability to develop more effective web service solutions. In practical terms, stakeholders armed with such a theoretical understanding could explore and evaluate different choices before making heavy resource investments for developing standards or for implementing and testing web service solutions.

 

1.5       Scope

The scope of this dissertation is to investigate whether LAP is an appropriate theoretical framework for web services. This investigation is conducted as three part research. The first part investigates the ability of LAP to describe and explain web service architecture. The second part demonstrates the ability of LAP to provide solutions to web service problems. The third part evaluates the utility of the LAP based solution developed in the second part. This dissertation does not claim that LAP is the best theoretical framework for web services. The scope of this dissertation does not include comparative tests against other theories.

 

Following this introduction, chapter 2 provides background and review of prior work on web services and LAP. Chapter 3 describes the research approach and methodology adapted in this dissertation. Chapter 4, 5, and 6 describe the first, second, and third parts of this three-part research, respectively. Chapter 7 provides summary of outcomes and contributions of this

dissertation.

A STUDY OF LANGUAGE-ACTION PERSPECTIVE AS A THEORETICAL FRAMEWORK FOR WEB SERVICES

Sharing is caring!