Presentation of 3 slides

JOURNAL2ExploringtherelationshipbetweensystemdevelopmentlifecycleandknowledgeaccumulationinTaiwansITindustry..pdf

Home >Business & Finance homework help >Management homework help >Presentation of 3 slides

Article DOI: 10.1111/j.1468-0394.2012.00630.x

Exploring the relationship between system development life cycle and knowledge accumulation in Taiwan’s IT industry Wen-Hsiang Lai and Hsin-Cheng Tsen Feng Chia University 100, Wenhwa Rd., Seatwen, Taichung 40724, Taiwan Email:[email protected]

Abstract: Many projects fail because the knowledge learned from them is obtained too late or insufficient (Koenig & Srikantaiah, 2004); ‘knowledge in projects’, ‘knowledge about projects’, and ‘knowledge from projects’ are three types of knowledge that result from project-based work (Love et al., 2005). This study explores the relationships between the system development life cycle (SDLC) of project management, firm-level explicit knowledge of organizational knowledge accumulation (OKA), and implicit knowledge of employee knowledge accumulation (EKA) with respect to knowledge accumulation (KA) and knowledge integration (KI). First, it analyzes the competence of SDLC in Taiwan’s IT enterprises by adapting expert interviews, analytic hierarchy process (AHP), and fuzzy rule-based theory. This reveals that system planning (SP) and system analysis (SA) are the most important SDLC phases. Second, based on the above result, this study investigates how the effectiveness of SDLC (ESDLC) correlates with KI, OKA, and EKA. Results indicate that EKA and OKA have obvious mutual influences, and that both show significant impact on ESDLC. Furthermore, KI has positive influence on EKA, but negative influence on OKA.

Keywords: knowledge integration (KI), knowledge accumulation (KA), system development life cycle (SDLC), information technology (IT)

1. Introduction

Since more and more enterprises are becoming aware of an optimal balance of product standardization and customiza- tion (Yan et al., 2005), enterprises have started using infor- mation technology (IT) to run their businesses (Zhang & Wang, 2006), and thus the development of IT is becoming the driving force of economic growth and the basis of pro- ductivity enhancement in enterprises. One viewpoint is that the new era of IT requires faster, smaller, and smarter ma- terials, devices, and software for human beings. The main areas of Taiwan’s IT services industry are system integra- tion (SI) and IT outsourcing. In 2008, Taiwan’s Market In- telligence & Consulting (MIC) Institute indicated that the production value of SI in Taiwan’s IT industry had reached NTD 80.82 billion, and the production value of IT outsourc- ing had reached NTD 14.98 billion. Successful SI and IT outsourcing projects in Taiwan produce a virtuous cycle of Taiwanese IT growth. However, Zarrella et al. (2005) indi- cate that IT project failure is a widespread problem. Based on Keil and Mann’s (1997) investigation, at least 30% of IT projects reveal some project escalation; project escala- tion results in large payoffs, long-term payoffs, high clos- ing costs, and low salvage values for IT companies (Pan et al., 2009). IT teams that adopt the system development life cycle (SDLC) method often increase their project success rates. SDLC provides careful examination of functional goals and detailed guidelines for system development and system implementation.

Bengston and Lesser (1998) indicate that enterprises usu- ally obtain skills and knowledge during the informatization of operational procedures, and that the systematization of operational procedures involves an opportunity for knowl- edge accumulation (KA). Therefore, the key to enterprise

competitiveness is the refinement of valuable organizational memories (knowledge) from the current structure of oper- ational procedures; firms can systematize their most conve- nient and cost- and manpower-saving strategies. Enterprises should conduct KA during the informatization of their op- erational procedures in order to achieve effective project ex- ecution and KA.

An enterprise that informatizes its internal operational procedures must build an information system (IS). This study is an early investigation exploring the relationship between SDLC and KA in Taiwan’s IT Industry. It also reveals the relationships between an enterprise’s KA and the informati- zation of its operational procedures; these relationships can be used to analyze the implicit knowledge activities that hap- pen when an enterprise develops IS. In this study, Analytic Hierarchy Process (AHP) is used to calculate weights for the five phases of SDLC. Since AHP can evaluate weights only by making pairwise comparisons within the five phases of SDLC, this research employs a fuzzy logic inference system (FLIS) to generate 3D fuzzy surfaces to analyze the rela- tionships within those five phases. FLIS researchers often make use of the Delphi method to collect opinions from professionals and to generate a fuzzy logic gate for fuzzy rule-based calculations (Perng et al., 2005; Hsu & Chen, 2007). To simplify the interview processes, the AHP weights can be utilized in the formation of the fuzzy logic gate. As Section 3.1 explains, 3D fuzzy surfaces enable clear anal- ysis of growth-and-decline relations within the five phases of SDLC. Furthermore, Section 3.2 shows how, after the 3D fuzzy surfaces have been obtained, a statistical method is adapted to analyze the effectiveness of SDLC (ESDLC) correlating with knowledge integration (KI), organizational knowledge accumulation (OKA), and employee knowledge accumulation (EKA).

2. Literature review

Knowledge is an important asset for a firm; it provides the capacity to formalize and leverage its intellectual assets as to formal and informal structure, functions, and processes. Tsai et al. (2010, p. 840) point out that a knowledge reposi- tory system serves to ‘provide the exchanging intermediaries of explicit knowledge between knowledge contributors and knowledge seekers to assist knowledge sharing of employ- ees in the enterprise’. Codification and personalization are two kinds of strategies applied in the knowledge manage- ment within firms; knowledge-intensive organizations should pursue either codification or personalization as a dominant strategy (80%) and use the other as a supporting strategy (20%) (Hansen et al., 1999). Codification knowledge repre- sents explicit knowledge, which involves not only converting undocumented information into documented information, but also making institutional knowledge visible, accessible, and usable for the decision making of firms. Codification strategy connects people by using the information stored in database systems, and thus IT supports the storage of this knowledge and its retrieval across the organization (Dun- ford, 2000). Personalization knowledge characterizes the im- plicit knowledge where the knowledge of a firm is stored mainly in people, and the sharing channel relies heavily on human interaction (Palmer & Platt, 2005). Personalization strategy involves using interpersonal relationships to mobi- lize and use personal knowledge (Polanyi, 1967). Fong and Kwok (2009, p. 1348) indicate, ‘knowledge management is critical and beneficial as indicated by 64% at the project and 74% at the organization level’. This study analyzes firm-level knowledge in two aspects, namely, EKA and OKA. EKA has to do with implicit knowledge, and OKA refers to explicit knowledge.

Love et al. (2005) observe that ‘knowledge in projects’, ‘knowledge about projects’, and ‘knowledge from projects’ are three types of knowledge that result from project-based execution and management. ‘Knowledge in projects’ refers to the knowledge that resides in a project in the form of doc- umentation, meeting repositories, discussions, and project management systems. ‘Knowledge about projects’ includes project designing, analyzing, planning, implementing, and controlling. ‘Knowledge from projects’ is the knowledge achieved from executing a project in the form of best prac- tices, lessons learned, post-project reviews, or after-action reviews. However, according to Papows’s study, one-third of knowledge management projects fail (Papows, 1999); in some cases, the lessons learned are obtained too late or are forgotten once the review is obtained at the end of a project (Koenig & Srikantaiah, 2004).

2.1. SDLC

Harrington et al. (1995) suggest that since the computa- tional support for concurrent engineering design consists of the design issues of knowledge-based systems, it is essential to provide the life-cycle perspective in recommending de- sign alternatives. SDLC is one of many well-known phase models. It includes an overall process for developing, im- plementing, and retrieving ISs through phases of initiation, analysis, design, implementation, and maintenance. There are many different SDLC models and methodologies; the

Figure 1: Five phases of system development life cycle (SDLC) vs. knowledge accumulation.

typical method consists of a series of defined steps or phases. SDLC is a conceptual model used in project management to develop large-scale, functional business IT systems. SDLC in IS activities revolves around heavy data processing and number-crunching routines (Elliott, 2004); it leads to repeat- able, high-quality results, and decreases project development costs. In the context of IT, SDLC refers to a creative IS devel- opment process; SDLC in IT is an entire process of formal and logical steps taken to develop software products. The op- erational procedures within an enterprise can be viewed as the IS inherent to the enterprise. The IS development lifecy- cle can be generally divided into five phases; each phase has a different set of priorities, and each of the phases consists of structural knowledge1, which should usually be documented for maximum efficiency. Within the SDLC model, the mile- stones are considered as ‘baselines’, and the ‘baselines’ are critical to the iterative nature of the model (Blanchard & Fab- rycky, 2006). In the software industry, the ‘waterfall model’ is a traditional structured design technique based on SDLC; however, Yamamichi et al. note, ‘this conventional design technique has limitations with respect to the transference of technology and quality evaluation’ (Yamamichi et al., 1998, p. 55). Also, because knowledge is widely used in a multitude of applications within various domains (Srihari & Westby, 1992), and because a knowledge source operates on some in- put data and builds the capability to transfer information as new outputs (Wu et al., 2000), it is essential to consider KA within the SDLC phases of IT projects. Figure 1 shows the five phases of SDLC vs. KA. Table 1 contains explanations of the five SDLC phases.

2.2. KA between firms and employees

Knowledge is an invisible but valuable asset; many firms find it hard to make practical use of their knowledge as- sets. Knowledge ages fast and can be rapidly upgraded. The core issues of knowledge management are to obtain valuable knowledge and to accumulate enterprise-specific knowledge.

1Dienes and Scott (2005, p.339) state, ‘structural knowledge might con- sist of knowledge of particular items, knowledge of fragments of items, knowledge of other types of rules, or knowledge embedded in connection- ist weights’.

Table 1: Descriptions of five system development life cycle (SDLC) phases

Period Abbreviation Explanation Structural knowledge

System planning SP SP involves setting goals, defining targets, establishing schedules, and estimating budgets for an entire software project.

1. Project planning document 2. Meeting minutes

System analysis SA SA builds on the SP phase; SA analyzes interacting entities; it usually relates to the enterprise’s operational procedures.

1. System interface 2. Procedure pseudocode

System design SD SD is based on the SA phase; SD defines the architecture, components, modules, interfaces, and data for an IS to satisfy the specified system requirements (SR).

1. Procedure flow diagram 2. Dataflow diagram 3. System Specification 4. Graphic user interface (GUI) 5. Schema 6. Meeting minutes

System implementation SI SI builds on the SD phase; SI provides data processing consultation, installation, training, and other computer-related support.

1. Graphic user interface (GUI) 2. Program code

System maintenance SM SM usually involves data backup, malware management, file system maintenance, etc.; it also improves the IS based on the users’ new requirements.

1. Change requirements 2. Management report

In the face of an uncertain and competitive environment, it is common for enterprises to elevate risk by sharing re- sources with other companies. During long-term and inti- mate cooperation, an enterprise has the chance to acquire knowledge or inspiration from its partners that might spark innovation. Cusumano and Elenkov (1994) point out that if individuals transfer knowledge between organizations and want to produce value, those individuals must consider the problem of how to make the knowledge recipients accept, absorb, and adapt the knowledge in their actions. Bonora and Revang (1991) classify the storage of knowledge into ‘mechanic storage’ and ‘organic storage’. ‘Mechanic storage’ refers to the storage of propositional and objective knowl- edge, and ‘organic storage’ refers to the storage of individual and subjective knowledge. Organizational knowledge can be stored in the memories of employees, and the combination of knowledge and activities within enterprises allows them to store knowledge in their systematic routines (Argot, 1999). Roos and Roos (1997) indicate that enterprises are not the owners of the knowledge, and that enterprises should dis- cover valuable knowledge, transform employees’ knowledge into competitive advantage, and try to accumulate knowl- edge. Barry and Stephens (1998) argue that organizational knowledge can be created and consolidated only through accumulation and diffusion processes. KA depends on the knowledge-absorbing abilities of knowledge receivers and knowledge-receiving enterprises. Although specific organi- zational knowledge is hard to accumulate, KI, absorption, and sharing are important abilities for enterprises to possess if they are to accumulate knowledge.

2.3. KI between firms and employees

KI is the foundation of competitive advantage in a dynamic environment. According to Fuller, KI has been a perennial concern, and the earliest precedents of KI can be classified into two recurrent metaphysical strategies: instantiation and emergence (Fuller, 2001). The KI process synthesizes multiple knowledge representations into a common repre- sentation. KI also includes the task of identifying knowledge

while incorporating new information into a knowledge base. However, this task is usually considered difficult, since new knowledge may interact with prior knowledge in very subtle and surprising ways, and unanticipated interaction may require changes to the knowledge base. Kogut and Zander (1992) describe KI as the ability to apply existing knowledge so as to acquire and integrate new knowledge. This ability refers not only to application of IT tools, such as databases, but also to communication and coordination between individuals, as well as the common knowledge possessed by these individuals. Since knowledge in the main lies in and is transformed by individuals (Nonaka, 1995; Grant, 1996a), enterprises can accumulate knowledge, embodied as human capital, via recruitment and training. KI can also come from institutional procedures, informal routines, or methods of problem solving; such KI can be obtained through observation, discussion, and revision of existing knowledge. Because external knowledge bases are plentiful, enterprises should not only integrate internal knowledge resources, but also integrate any external knowledge that might be useful (Nonaka, 1994). In the context of knowledge-based theory and knowledge market mechanisms, organizational knowledge assets can be created, accumulated, and shared through an effective mechanism of knowledge management; construction and maintenance of such a mechanism can en- hance a firm’s core competitiveness. Enterprises should use KI to systematize and structure their knowledge by combin- ing various kinds of internal and external knowledge. Grant (1996b) states that in the course of KI, the more relevant knowledge the company has, the more easily that firm can express the knowledge in the form of a common language. Therefore, an enterprise’s ability to absorb knowledge has a positive influence on its ability to integrate knowledge.

3. Research methodology

In the real world, people make decisions every day, and most of the decisions are based on vague thinking logic. The at- titudes of decision makers are reflected in the data input

Table 2: Nine steps of analytic hierarchy process (AHP) im- plementation in this study

Step Explanation

1 A system can be dissolved into many components. 2 In the hierarchy structure, each factor has a single

level. 3 Factors of each level can be evaluated according to

the previous level. 4 Absolute numerical scale can be transferred into

proportional scale. 5 After pairwise comparison, use a positive reciprocal

matrix. 6 The relationship of factors’ priorities must have

transitivity (if A is better than B, and B is better than C, then A must dominate C), and strength (if A is two times better than B, and B is two times better than C, then A must be four times better than C).

7 Transitivity is hard. Thus, factors that do satisfy transitivity can be acceptable. However, each such factor must have a consistency ratio (CR < 0.1), where CR = CI/RI, and a CI = (λmax − n) / (n − 1); RI = random index, gained from the rank of RI (Satty, 1980, p. 21).

8 Through the weighting principle, factors gain more intense priority levels.

9 Regardless of how intense factors may be, all factors in the hierarchy are considered to relate to the overall assessment.

CI, consistency index; RI, random index.

stage, and the linguistic variables are major concepts used to assess preference ratings; these are ‘importance’ and ‘appro- priateness’ (Lai et al., 2010). The decision makers employ the assumed weighting set {Very Low, Low, Medium, High, Very High} to assess the relative importance of various cri- teria, and they use the linguistic rating set {Very Poor, Poor, Fair, Good, Very Good} to evaluate the appropriateness of alternatives against various criteria. This study analyzes the competence of SDLC implementations in Taiwan’s IT enter- prises through expert interviews, AHP, and fuzzy rule-based theory. The AHP assigns weights to the phases of SDLC, namely, system planning (SP), system analysis (SA), system design (SD), SI, and system maintenance (SM). The fuzzy rule set consists of many fuzzy IF-THEN rules.

3.1. Fuzzy AHP method

AHP is a process that induces cognitive awareness, evoking various decision-making problems with a multitude of quan- titative and qualitative variables and sub-variables. After ob- taining the values of weightings from AHP, Fuzzy Logic Decision Systems (FLDS) can be adapted to generate 3D fuzzy surfaces and analyze the growth-and-decline relations within variables and sub-variables (Lai & Tsai, 2008). Fuzzy inference is the core of the fuzzy system in a wide variety of engineering systems. Through inferences from a fuzzy rule base, the inference engine processes input sets and produces output sets. Table 2 shows the nine steps of AHP implemen- tation in this study.

3.1.1. Expert interviews and first-stage AHP questionnaire The first-stage questionnaire uses AHP methods to diagnose five phases of SDLC. The authors selected participants who

had worked in the relevant IT field (i.e. administration, MIS department, project management, office automation depart- ment, and factory automation department), and most of the chosen experts had more than 10 years of experience; open- structured interviews were conducted with them. This study delivered 32 AHP questionnaires, and 25 questionnaires were retrieved (78% retrieval rate). After using a consistency in- dex (CI) examination from Expert Choice 2000, this study deleted unqualified questionnaires, and finally 18 valid inputs remained. Two of the experts are corporate general managers, and the other 16 experts are managers in IT-related compa- nies. Table 3 shows the experts’ backgrounds from the 18 valid AHP questionnaires. The validity rate is approximately 56%. This study classifies factors from Table 1 into an AHP research framework to evaluate SDLC, shown in Figure 2. There are two levels in the AHP hierarchy framework; level 1 is the target level, and it is obtained after level 2 has been calculated. Level 2 consists of the five major factors: SP, SA, SD, SI, and SM. The weights (w) in level 2 are calculated using the AHP method.

Table 3: Experts’ backgrounds from 18 valid analytic hierar- chy process (AHP) questionnaires

No. Expert’s Professional Number. of Years of background title handled experiences

projects

1 Administration management

CEO 14 23

2 Administration management

Vice CEO 12 18

3 Administration management

Director 10 16

4 Administration management

Deputy director

8 10

5 Manager in MIS department

Manager 10 14

6 Manager in MIS department

Deputy manager

3 11

7 Manager in MIS department

Project manager

8 12

8 Manager in MIS department

Project manager

12 13

9 Project manager in IT department

Deputy manager

19 25

10 Project manager in IT department

Deputy manager

7 11

11 Project manager in IT department

Project manager

16 19

12 Project manager in IT department

Project manager

6 12

13 Project manager in IT department

Project manager

21 22

14 Office automation department

Senior engineer

4 10

15 Office automation department

Senior engineer

13 16

16 Office automation department

Engineer 3 17

17 Factory automation department

Senior engineer

9 10

18 Factory automation department

Engineer 3 12

All the occupations of experts’ backgrounds are IT-related industries.

Figure 2: The system development life cycle (SDLC) effectiveness hierarchy framework.

Figure 3: A fuzzy logic decision system (FLDS).

3.1.2. Construction of FLIS In Figure 3, X represents the set of key attributes influencing the ESDLC, and Y represents the output of SDLC effectiveness. If we consider Figure 3 based on the viewpoint of fuzzy set theory, it is obvious that X represents fuzzy memberships of the linguistic terms describing input attributes after ‘fuzzification’, and thus the linguistic terms are matched with the preconditions of fuzzy IF-THEN rules (Lai & Tsai, 2008). Any fuzzy IF-THEN rule with preconditions satisfied can be used.

3.1.3. Inference system Fuzzy logic theory generally deals with a set of fuzziness, fuzzy relation, fuzzy logic, fuzzy con- trol, and fuzzy measure theory. Through the adoption of a fuzzy rule base, the inference engine converts input sets to produce output sets, and the consequence can be con- verted into definite value. This process is called ‘defuzzifi- cation’; three commonly used defuzzifiers are centre of area defuzzifier, centre average defuzzifier, and maximum defuzzi- fier. This study adopts a centre of area defuzzifier, owing to its continuity and lack of ambiguity of fuzzy calculation. Figure 4 shows the inference regarding major influential fac- tors of SDLC, as implemented using Matlab fuzzy tools.

3.1.4. Membership function The membership function of a fuzzy set is considered as the indicator function in classical sets, and it also represents the degree of truth as an exten- sion of valuation. The membership function, derived from

Figure 5: The bell-shaped membership functions.

the characteristic function, is to express objects’ member- ship grade between 0 and 1. Based on Ragin’s (2000) study, this study sets three important qualitative anchors (Mod- erate High, Moderate Middle, and Moderate Low) within the ‘Moderate’ input and output criteria. Figure 5 shows the membership functions of the major influential factors.

3.1.5. Build IF-THEN rules This section builds IF-THEN rules out of the High (H), Moderate High (MH), Moderate Middle (MM), Moderate Low (ML), and Low (L) values of input and output criteria, shown in Table 4. Based on Lai and Tsai’s (2008) study, the fuzzy numbers assigned to the linguistic terms are (H, MH, MM, ML, and L) = (3, 2.5, 2, 1.5,and 1); therefore, the outcome (Y) of the SDLC effectiveness is equal to the multiplicative product of fuzzy numbers and weightings. Each of the five factors can be used as a rule input. All rules have a unique output defined for

Figure 4: The Mamdani inference concerning system development life cycle (SDLC) effectiveness.

every possible set of inputs. Each rule that relates factors takes five inputs (H, MH, MM, ML, and L). There is one input for each factor, that is, SP, SA, SD, SI, and SM, and each input can take one of five values. Thus, there are 5 × 5 × 5 × 5 × 5 = 3125 possible input sequences, because there are 3125 combinations of values. Table 4 indicates the definitions of input and output values. Table 5 shows the outcome ranges and linguistic terms for SDLC effectiveness. Table 6 shows an example of the fuzzy rule-based calculations and outcome values.

3.2. Statistical method

3.2.1. Second-stage statistical questionnaire The authors mailed 280 surveys to key decision makers in a randomly selected sample of IT organizations in Taiwan. The statisti- cal questionnaire adopted Likert 5 scales (‘strongly agree’, ‘agree’ ‘moderate’, ‘disagree’, ‘strongly disagree’) to measure the answers from questionnaire participants.

3.2.2. Hypotheses Systematic KA and various adaptations of knowledge are critical issues in knowledge management. The complexity of dynamic application domains and the un- certainty of future maintainability are the main problems of KA (Koh et al., 2005). Polanyi (1967) classifies knowl- edge into tacit (subjective) knowledge and explicit (objec- tive) knowledge. Knowledge flows exist between tacit and explicit states, creating a symbiotic relationship between the two states. From the theoretical viewpoint, Davenport and Prusak (1998) state that knowledge is a kind of professional intelligence that can be divided into four levels: cognitive knowledge, advanced skills, system understandings, and self- motivated creativities. Davenport and Prusak (1998) further indicate that knowledge is a kind of fluid synthesis, which includes structured experiences, values, and contextualized information. Knowledge also includes unique viewpoints from experts; such viewpoints offer a framework for the assessment and integration of new experiences. From the entrepreneurial viewpoint, organizational knowledge within enterprises consists of special knowledge (a company’s con- ventions, workflows, documents, and business secrets) and general knowledge (public knowledge) (Matusik & Hill, 1998). Purser and Pasmore (1992) state that knowledge is the aggregation of fact, opinion, and intuition used for de- cision making. Nonaka and Takeuchi (1995) observe that knowledge creation is developed in a collective manner, by continuous tacit and explicit knowledge sharing between in- dividuals and groups within enterprises. How to integrate knowledge inside and outside the organization, and how to

Table 5: Outcome ranges and linguistic terms for system de- velopment life cycle (SDLC) effectiveness

Ranges of Y Linguistic terms Abbreviation

2.6 < Y <= 3 High H 2.2 < Y <= 2.6 Moderate high MH 1.8 < Y <= 2.2 Moderate middle MM 1.4 < Y <= 1.8 Moderate low ML 1 <= Y <= 1.4 Low L

effectively accumulate and manage that knowledge, are two important subjects for enterprises today (Leonard-Barton, 1995).

Knowledge is dynamic, relevant, and unstable; it varies from person to person (Nonaka, 1994). In this innovative era, enterprises consider the sustained ability to innovate as critical to competitiveness. De Boer et al. (1999) state that enterprises should be able to integrate knowledge in order to improve their competitive advantages efficiently. To maintain competitive forces continuously, it is not sufficient for enterprises to merely rely on their internal management of research and development (R&D) to create knowledge; firms must acquire external knowledge to accumulate knowledge. Thus, acquisition of external knowledge plays an important role in generating new knowledge within enterprises. In the global knowledge economy, a growing number of enterprises are implementing methods for knowledge management and accumulation; these tasks are becoming important in their operations.

This study adapts theories of implicit knowledge and explicit knowledge from Polanyi, Nonaka, Takeuchi, and Leonard-Barton. Implicit knowledge generally means the knowledge inside a person’s brain, including that person’s way of doing things, experiences, judgments, decisions, intentions, unofficial skills, techniques, experiences, secrets of success, personal intuitions, inspirations, powers of obser- vation, values, intelligence, etc. Explicit knowledge refers to the kind of knowledge that has already been stored in media with characters, sound, or images. It can be filed and stored in documents, patents, trademarks, reports, manuals, books, files, databases, plans, summaries, forms, etc. Persons other than the original expert can access and organize the expert’s external knowledge much more easily than the expert’s inter- nal knowledge. Different bodies of knowledge, with different characteristics, exist in different knowledge carriers. People are the carriers of implicit knowledge, which in this context exists in employees of the enterprise. On the other hand, the carriers of explicit knowledge may be organizations. As mentioned in subsection 3.2.1, this study also divides

Table 4: Definitions of input and output values

Input criteria Output value

Effectiveness Factors Linguistic terms Values Name Linguistic terms Values

System planning (SP) High → 3 High → 3 System analysis (SA) High→ 2.5 High→ 2.5

System development life cycle (SDLC)

System design (SD) Moderate { Middle→ 2 Moderate { Middle→ 2System implementation (SI)

Low→ 1.5 Effectiveness of SDLC

Low→ 1.5

System maintenance (SM)

Low → 1 Low → 1

Table 6: An example of the fuzzy rule-based calculations and outcome values

Scenarios System planning System analysis System design System implementation System maintenance Outcome Linguistic (SP) (0.312) (SA) (0.273) (SD) (0.105) (SI) (0.083) (SM) (0.227) value terms

1 if H H H H H 3 H 2 if H H H H MH 2.773 H 3 if H H H H MM 2.546 H 4 if H H H H ML 2.917 H 5 if H H H H L 2.69 H 6 if H H H MH H 2.463 MH 7 if H H H MH MH 2.834 H

. . . . . . . . . . . . 3119 if L L L ML ML 1.166 L 3120 if L L L ML L 1.537 ML 3121 if L L L Low H 1.31 L 3122 if L L L Low MH 1.083 L 3123 if L L L Low MM 1.454 L 3124 if L L L Low ML 1.227 L 3125 if L L L Low L 1 L

Figure 6: The research framework of this study.

knowledge into EKA and OKA. Figure 6 shows the research framework of this study and the mediate variable, KI, between EKA and OKA. The hypotheses of the present study are listed below:

Hypothesis H1: EKA influences the KA embed- ded in OKA.

Hypothesis H2: OKA influences the KA embed- ded in EKA.

Hypothesis H3: KI can help transfer EKA knowledge to OKA knowledge.

Hypothesis H4: OKA has an influence on ES- DLC.

Hypothesis H5: EKA has an influence on ES- DLC.

4. Empirical results

4.1. Results of fuzzy AHP

Examination of the first-stage AHP questionnaire showed that CI was less than 0.1, which satisfied the requirements of AHP. Analysis of priority vectors (weightings) of five factors for SDLC effectiveness produced results for those factors, shown in Table 7. The result shows that SP (w = 0.312) is the most important factor in the SDLC process. That is, in

Table 7: Analytic hierarchy process (AHP) results of system development life cycle (SDLC) factors

Factors Priority Vector Ranking (Weighting, w)

System planning (SP) 0.312 1 System analysis (SA) 0.273 2 System design (SD) 0.105 4 System implementation (SI) 0.083 5 System maintenance (SM) 0.227 3

building an IS, setting goals, defining targets, establishing schedules, and estimating budgets are top priorities for the SDLC. Also, SA (w = 0.273) and SM (w = 0.227) are the second and third most important factors in the SDLC.

4.1.2. The fuzzy surfaces of SDLC After editing the fuzzy rule bases and considering the AHP weightings, the authors found four fuzzy surfaces of SDLC; they were based mainly on the factor SP. Figures 7 (a), 7 (b), 7 (c), and 7 (d) show the fuzzy surfaces of major variables: SA vs. SP, SD vs. SP, SI vs. SP, and SM vs. SP, respectively. Figure 7 (a) indicates that SA and SP have symmetric influence levels with similar strengths. Except for the strength range between 1 and 2, SA and SP have significant impact on the ESDLC. As long as the degree of strength of SA or SP increases, the SDLC effectiveness outcome also increases. Figure 7 (b) shows that, without considering SD, as long as the degree of strength of SP is close to 1, the SDLC effectiveness outcome is mostly in the high

Figure 7: Three-dimensional fuzzy surfaces of system development life cycle (SDLC).

range. Figure 7 (c) shows that SI has even less influence than SD on the ESDLC. Also, as long as the degree of strength of SP is close to 1, the SDLC effectiveness outcome is mostly located in the high range. Figure 7 (d) shows that SM and SP have similar influence levels, and also indicates that the SDLC effectiveness outcome is located in the high range. It is clear that when the influence level of SM or SP is close to 1, the SDLC effectiveness becomes significant.

4.2. Statistical analysis

Sixty-three responses (out of 280) to the questionnaire were received; these constituted a 22.5% response rate. Of the 63 questionnaires returned, 58 were valid. To ensure the valid- ity of the analysis, indicators were not retained unless their eigenvalues were greater than 1.0 and their Cronbach’s α val- ues were greater than 0.6. Table 8 shows the Cronbach’s α values.

The results of statistical factor analysis indicate that four factors in this study are to be retained: KI, OKA, EKA, and ESDLC. The statistical route analysis, shown in Table 9, suggests that KI is the factor with the greatest influ- ence on the ESDLC of a system. Its route coefficient is 0.598 (significance level 0.01). The second most influential factor is OKA (β = 0.189), and the least influential factor is EKA (β = 0.103). Table 9 also shows that EKA and OKA have mutual influences (β = 0.593 and β = 0.695), and that both of them reach the 0.01 significance level. According to the route analysis of KI, this factor has different influences on EKA and OKA. KI has higher significance to the EKA (β = 0.246) and it reaches the 0.1 significance level; however, KI is

Table 8: Cronbach’s α values

Variable Abbreviation Cronbach’s α value

Knowledge integration KI .805 Organizational

knowledge accumulation

OKA .831

Employee knowledge accumulation

EKA .654

Effectiveness of SDLC ESDLC .773

not significant to the OKA (β = −0.036). This study assumes that the hypotheses are supported at least at the significance level of 0.1 (p < 0.1). Figure 8 presents the results of statis- tical route analysis. Of five research hypotheses, three items, H1, H2, and H3, are supported by the route analysis; the other two items, H4 and H5, are not supported. This study demonstrates that for a company executing SDLC, KI has more influence on its system R&D than the EKA and OKA knowledge carriers. Thus, knowledge has a certain influence on SDLC effectiveness. Table 10 compares the hypotheses in this study.

5. Conclusion

Swaffield and Knight (1990) indicate that the techniques of data flow analysis, entity-relationship analysis, and entity- life cycle analysis are used to structure the data associated with knowledge engineering. The theories of this study are derived from Nonaka’s and Takeuchi’s (1995) and Leonard- Barton’s (1995) concepts of multiple hierarchical knowledge structures, and of implicit and explicit knowledge. This study has proved that during the SDLC process for various IT sys- tems, KA can be improved through the transformation of implicit knowledge (EKA) and explicit knowledge (OKA), and that KI plays an important role in the interaction be- tween EKA and OKA. Grant (1996) has opined that during the process of KI, obtaining more relevant knowledge can improve various applications of KA, and can have a positive influence on knowledge absorption and knowledge sharing. The results of this study agree with Grant’s (1996b) find- ing. Strengthening an organization’s education and employee training can improve employee knowledge absorption. Fur- thermore, employees’ KA activities can indirectly provide inherent organizational knowledge through a spiralling re- ciprocation of knowledge. KA can be influenced by different knowledge carriers; SDLC is too complex to be dominated by the direct influence of a single group. However, knowl- edge can be accumulated through knowledge reciprocation among employees and organizations, such as the company’s education and training program (which transfers knowledge from the organization to individuals) and training given by senior staff to junior staff (which transfers knowledge from

Table 9: Results of route analysis

Variables R2 value Route β Value T value p value

(Y1) System development life cycle (SDLC)

0.522 Knowledge integration (KI)→Effectiveness of SDLC (ESDLC)

0.598 4.006 0.000***

OKA→ESDLC 0.189 1.043 0.307 EKA→ESDLC 0.103 .537 0.596

(Y2) Organizational knowledge accumulation (OKA)

0.418 KI→OKA − 0.036 − 0.220 0.828 EKA→OKA 0.659 4.086 0.000***

(Y3) Employee knowledge accumulation (EKA)

0.476 KI→EKA 0.246 1.714 0.098* OKA→EKA 0.593 4.086 0.000***

*means at the level of 0.1, the result of two-tailed checking is significant. ***means at the level of 0.01, the result of twotailed checking is significant.

Figure 8: Route analysis in this study.

individuals to individuals). This study finds that even though KA cannot directly improve SDLC ability, KA can help im- prove SDLC ability by integrating EKA and OKA. Thus, when the employee and the organization start to integrate knowledge and to interact with mutual knowledge, KA will have a clearly beneficial influence on SDLC.

KI is an important ability for any enterprise’s information management. In order to fully employ the value of knowl- edge, an enterprise must not only integrate knowledge in- side the firm but also absorb useful knowledge from outside the firm. All enterprises must assiduously practice KA. In order to adjust nimbly to a rapidly changing environment,

Table 10: A comparative synthesis of the hypotheses

Hypothesis Description Result

H1 Hypothesis H1: employee knowledge accumulation (EKA) influences the knowledge accumulation embedded in organizational knowledge accumulation (OKA).

Supported

H2 Hypothesis H2: OKA influences the knowledge accumulation embedded in EKA.

Supported

H3 Hypothesis H3: Knowledge integration (KI) can help transfer EKA knowledge to OKA knowledge.

Supported

H4 Hypothesis H4: OKA has an influence on ESDLC.

Not Supported

H5 Hypothesis H5: EKA has an influence on ESDLC.

Not Supported

companies should foster efficient, pluralistic KI and flexi- ble knowledge reorganization. EKA plays the main role in the dynamic knowledge exchange process, bringing essen- tial improvements to IT organizations. Enterprises should strengthen and coordinate knowledge management abilities, such as with respect to intellectual property rights, the culture of trust, and KA resources, in order to improve communi- cation between enterprises and their partners. KA promotes the efficiency of industrial linkages and prevents the prob- lems caused by opportunism in business.

References

ARGOT, L. (1999) Organizational Learning: Creating, Retaining and Transferring Knowledge, Boston: Kluwer Academic Publishers.

BARRY, B. and C.U. STEPHENS (1998) Objections to an objectivist approach to integrity, Academy of Management Review, 23, 162– 169.

BENGSTON, D. and E. LESSER (1998) Turbocharging business pro- cesses with knowledge, Journal of Innovative Management, 4, 367– 521.

BLANCHARD, B.S. and FABRYCKY, W.J. (2006) Systems Engineering and Analysis, 4th edn, New Jersey: Prentice Hall.

BONORA, E.A. and O. REVANG (1991) A Strategic Framework for Analyzing Professional Service Firms Development Strategies for Sustained Performance. Toronto: Strategic Management Society Inter-organizational Conference.

CUSUMANO, M.A. and D. ELENKOV (1994) Linking international tech- nology with strategy and management:a literature commentary, Research Policy, 23, 195–215.

DAVENPORT, T.H. and L. PRUSAK (1998) Working Knowledge: How Organizations Management What They Know, Boston: Harvard Business School Press.

DE BOER, M., F.A.J. BOSCH, and H.W. VOLBERDA (1999) Managing organizational knowledge integration in the emerging multimedia complex, Journal of Management Studies, 36, 379–398.

DUNFORD, R. (2000) Key challenges in the search for the effective management of knowledge in management consulting firms, Jour- nal of Knowledge Management, 4, 295–302.

ELLIOTT, G. (2004) Global Business Information Technology: An Inte- grated Systems Approach, London: Addison-Wesley.

FONG, S.W. and W.C. KWOK (2009) Organizational culture and knowl- edge management success at project and organizational levels in contracting firms, Journal of Construction Engineering & Manage- ment, 135, 1348–1356.

FULLER, S. (2001) Strategies of knowledge integration, in Our Fragile World: Challenges, Opportunities for Sustainable De- velopment, M.K. Tolba (ed.), Oxford: EOLS Publishers, 1215–1228.

GRANT, R.M. (1996a) Prospering in dynamically-competitive en- vironments: organizational capability as knowledge integration, Organization Science, 7, 375–388.

GRANT, R.M. (1996b) Toward a knowledge-based theory of the firm, Strategic Management Journal, 17, 109–122.

HANSEN, M.T., N. NOHRIA, and T. TIERNEY (1999) What’s your strat- egy for managing knowledge? Harvard Business Review, 77, 106– 116.

HARRINGTON, J.V., H. SOLTAN, and M. FORSKITT, (1995) Negotiation in a knowledge-based concurrent engineering design environment, Expert Systems, 12, 139–148.

HSU, P.F. and B.Y. CHEN (2007) Developing and implementing a selection model for bedding chain retail store franchisee using Delphi and fuzzy AHP, Quality and Quantity, 41, 275–290.

KEIL, M. and J. MANN, (1997) The nature and extent of IT project escalation: results from a survey of IS audit and control profes- sionals, IS Audit Control Journal, 1, 40–48.

KOENIG, M.E.D., and T.K. SRIKANTAIAH (2004) Knowledge Manage- ment Lessons Learned: What Works and What Doesn’t. Medford, NJ: Information Today.

KOGUT, B. and U. ZANDER (1992) Knowledge of the firm, integra- tion capabilities, and the replication of technology, Organization Science, 3, 383–397.

KOH, H., S. HA, T. KIM, and S.H. LEE (2005) A method of accumulation and adaptation of design knowledge, The Inter- national Journal of Advanced Manufacturing Technology, 26, 943–949.

LAI, W.H., P.L. CHANG, and Y.C. CHOU (2010) Fuzzy MCDM ap- proach to R&D project evaluation in Taiwan’s public sectors, Jour- nal of Technology Management in China, 5, 84–101.

LAI, W.H. and C.T. TSAI (2008) Analyzing influence factors of tech- nology transfer using fuzzy set theory. Cape Town, South Africa: PICMET Conference, 2287–2295.

LEONARD-BARTON, D. (1995) Wellsprings of knowledge: building and sustaining the sources of innovation. Massachusetts: Harvard Busi- ness school.

LOVE, P., P. FONG, and Z. IRANI (2005) Management of Knowledge in Project Environments. Oxford: Elsevier.

MATUSIK, S.F. and C.W. HILL (1998) The utilization of contingent work, knowledge creation, and competitive advantage, Academy of Management Review, 23, 680–697.

NONAKA, I. (1994) A dynamic theory of organizational knowledge creation, Organization Science, 5, 14–37.

NONAKA, I. and H. TAKEUCHI (1995) The Knowledge-Creating Com- pany, New York: Oxford University Press.

PALMER, J. and S. PLATT (2005) Business Case for Knowledge Man- agement in Construction, London: CIRIA.

PAN, G., S.L. PAN, and M. NEWMAN (2009) Managing informa- tion technology project escalation and de-escalation: an approach- avoidance perspective, IEEE Transactions on Engineering Manage- ment, 56, 76–94.

PAPOWS, J. (1999) Enterprise.com: Market Leadership in the Informa- tion Age, Cambridge MA: Perseus Books.

PERNG, Y., S. HSUEH, and M. YAN (2005) Evaluation of housing con- struction strategies in China using fuzzy-logic system, International Journal of Strategic Property Management, 9, 215–232.

POLANYI, M. (1967) The Tacit Dimension, New York: M.E. Sharp Inc. PURSER, R.E. and W.A. PASMORE (1992) Organizing for learning,

in Research in Organizational Change and Development, Vol. 6 William A. Pasmore and Richard W. Woodman (eds.), London: JAI Press, 37–114.

RAGIN, C.C. (2000) Fuzzy-Set Social Science, Chicago: The University of Chicago Press.

ROOS, G. and J. ROOS (1997) Measuring your company’s intellectual performance, Long Range Planning, 30, 413–426.

SAATY, T. L. (1980) The Analytic Hierarchy Process. New York, McGraw-Hill.

SRIHARI, K. and G. WESTBY (1992) Decision support for PCB as- sembly using knowledge based expert systems, in Proceedings— NEPCON West 1992, Anaheim, CA: Reeds Exhibition, 545– 554.

SWAFFIELD, G. and B. KNIGHT (1990) Applying systems analysis tech- niques to knowledge engineering, Expert Systems, 7, 82–92.

TSAI, C.H., D.S. ZHU, B.C. HO, and D.D. WU (2010) The effect of reducing risk and improving personal motivation on the adoption of knowledge repository system, Technology Forecasting and Social Change, 77, 840–856.

WU, C.H., S.C. KAO, and K. SRIHARI (2000) GDKAT: a goal-driven knowledge acquisition tool for knowledge base development, Ex- pert Systems, 17, 90–105.

YAMAMICHI, N., K. YOKOCHI, and T. TANAKA (1998) Quality evalu- ation of software development process based on spiral modeling, Electronics and Communications in Japan, 81, 55–64.

YAN, W., C.H. CHEN, and P.K. LI (2005) A web-enabled product def- inition and customization system for product conceptualization, Expert Systems, 22, 241–253.

ZARRELLA, E., M. TIMS, B. CARR, and W. PALK (2005) Global IT Project Management Survey: How Committed are You? London, UK: KPMG International.

ZHANG, L. and H. WANG (2006) Intelligent information processing in human resource management: an implementation case in China, Expert Systems, 23, 356–369.

The authors

Wen-Hsiang Lai

Dr. Wen-Hsiang Lai is an Associate Professor in the Grad- uate Institute of Management of Technology at Feng Chia University, Taiwan. He received his PhD in Mechanical En- gineering from The University of Kansas, USA in 1999. His research focuses on the R&D management, such as the fields of technology transfer, R&D alliance, optimization of knowledge accumulation and engineering outsourcing, university–industry collaboration, industrial service flexibil- ity, and R&D performance evaluation.

Hsin-Cheng Tsen

Dr. Hsin-Cheng Tsen is an Associate Professor in the De- partment of Business Administration at Feng Chia Univer- sity, Taiwan. He received his PhD in Computer Science from The University of New Mexico, USA in 1990. His academic research mainly focuses on the areas of Information Manage- ment, such as the fields of Management Information System, Business Intelligent, Data Mining, Knowledge Management, and Electronic Commerce. His research often adapts quan- titative methods, such as factor analysis, regression analysis, ANOVA, SEM, AHP, and some other statistical methods to analyze research data.

to multiple sites or posted to a listserv without the copyright holder's express written permission. However,

users may print, download, or email articles for individual use.