ict, society and human beings 2008

2 downloads 1009 Views 304KB Size Report
IT Consulting company. Figure 4. Demographics of 10 experts. 4.2 Perceived Nature of Conflict. This section presents our findings regarding the experts' ...
IADIS International Conference ICT, Society and Human Beings 2008

USING AN EXPERT PANEL TO EMPIRICALLY VALIDATE A REQUIREMENTS ENGINEERING MEDIATION MODEL Nan Maa, Tracy Hallb, Trevor Barker a a

School of Computer Science, University of Hertfordshire, College Lane, Hatfield A10 9AB, UK b Department of Information Systems & Computing, Brunel University, Uxbridge, Middlesex UB8 3P

ABSTRACT Objective: In this paper we present findings from an empirical validation study of a newly developed narrative-based requirements engineering mediation (NREMM) model. The main objective of this study is to validate whether the motivation for developing the NREMM model is justified and how well the NREMM model meets its objectives. Methods: We used semi-structured interviews with a combination of closed and open-ended questions to elicit 10 requirements experts’ opinions of NREMM model. Both quantitative and qualitative data were collected. Basic frequency analysis of the quantitative data was performed, as well as content analysis of the qualitative data. Findings: Overall, we found that NREMM is potentially capable of helping RE practitioners to identify, define, and resolve conflicts in the requirements engineering (RE) process. However, the NREMM needs further improvement in terms of integrating with existing RE modelling and goal analysis techniques. In addition a set of guidelines to ensure practitioners can correctly implement NREMM in practice is also required. Conclusions: Our expert panel findings indicate that our newly developed NREMM model meets its original objective to provide an effective framework guiding practitioners to mediate conflicts in the RE process. However, NREMM needs further improvement to make it easier to adopt by practitioners. KEYWORDS Requirements Negotiation, Conflict, Conflict resolution, Narrative Mediation

1. INTRODUCTION In this paper we describe how we use a panel of experts to initially validate NREMM. Conflict is a common phenomenon in everyday life. It also has been recognized as an inevitable part of RE process (Sommerville and Sawyer, 1997). However, in the current RE literature, conflict is typically recognized as a technical issue that may lead to inconsistency in the requirements specification (Easterbook, 1996). Most previous work in this area focuses on presenting technical methods or techniques for modelling, analyzing, and managing conflict or inconsistency e.g. KAOS (Van Lamsweerde, 2000), Problem Frames (Jackson, 2001) and I* (Yu, 1998) or tools for automating conflict identification and resolution e.g. Oz (Robison, 1990), Synoptic (Easterbrook, 1996), or prompting groupware systems for remote negotiation e.g. Win-Win (Boehm et al, 2001). Little attention is given to the socio-psychological aspect of the conflict. Moreover, previous RE work consistently views the resolution of conflicts as a purely negotiation-based process (Boehm et al, 2001; Damian et al. 2003) in which a requirements engineer acts as a representative of a developer site and negotiates with users’ site to develop the trade-offs among the conflicting requirements. This paper adopts a complementary viewpoint and differentiates itself from this previous work by recognizing conflict as a social, human, and organizational issue. We adopt Barki and Hartwick (2001:157)’s definition of conflict as “a phenomenon that occurs between interdependent parties as they experience negative emotional reactions to perceived disagreements and interference with the attainment of their goals.” Furthermore, we also view the process of resolving conflict in RE as a mediated process, in which a requirements engineer acts as a mediator among different stakeholders (Ma et al. 2008b). To address this

99

ISBN: 978-972-8924-61-4 © 2008 IADIS

socio-psychological aspect of conflict in RE, we have developed NREMM to help RE practitioners effectively identify, define, and resolve conflicts (Ma et al., 2008a). NREMM borrows the original narrative mediation theory from Winslade and Monk (2000) and translates it into the context of RE (Ma et al. 2008b). This paper aims to validate whether our motivation for developing NREMM is justified and how well NREMM meets its objectives. At this initial stage we simply validate whether NREMM meets its design objectives. Our evaluation of the model’s usability forms the next stage of the model’s development. This paper is organised as follows: Section 2 gives an overview of NREMM model. Section 3 presents our methodological design of an expert panel validation. Section 4 provides our expert panel results. Finally, Section 5 concludes the paper and outlines our plans for future research.

2. OVERVIEW OF NREMM In this section, we present an overview of NREMM, which is translated from Winslade and Monk (2000)’s original narrative mediation approach (see Ma et al., 2008b for detailed descriptions of Winslade and Monk (2000)’s model). We follow a rigorous and systematic model translation approach (see Ma et al., (2008a). NREMM model contains three sub-models: conflict identification (See figure-1), conflict definition (see figure-2), and conflict resolution (see figure-3). In every sub-model, relationship practice and dialogical practice play an important role. Conflict resolution in RE is a cooperative practice in which the parties to the conflict are viewed as partners in mediation. Relationship practice thus provides a set of techniques to guide a mediator to create a relational climate among conflicting stakeholders. Conflict resolution in RE also lies in the dialogues between conflicting stakeholders. Dialogical practice thus offers useful listening and questioning techniques to help a mediator to uncover the real conflicts and separate people from the conflicts.

2.1 Sub-Model-1: Conflict Identification The aim of this sub-model is to establish a workable relationship with the conflicting stakeholders and initially identify conflict between them. The major activities in this phase include selecting meeting settings, relationship practice, dialogical practice, and stakeholder modelling. Each activity recommends a list of good practices or techniques. Figure 1 shows that the new model retains most elements of relational practice, and dialogical practice from the original model, and is complemented by the RE specialised techniques of stakeholder modelling and preparing an RE meeting setting. Relationship practice Showing respect Value personhood Build trust Rituals of engagement (If necessary)

Dialogical practice Watching tacit communication Watch facial expression Discursive listening Genuine curiosity questioning Inviting production of meaning Inviting collaborative conversions

Conflict Identification

Selecting Meeting setting Seat and Meeting layout Light, hearting, general Artifacts (Flip chart, whiteboard, PC)

Stakeholder modelling Identify stakeholder' role Create a persona Create an extreme character

Figure 1. Conflict identification sub-mod

100

IADIS International Conference ICT, Society and Human Beings 2008

2.2 Sub-Model-2: Conflict Definition The aim of this sub-model is to gain an accurate understanding of conflict. The original model refers to this phase as “deconstructive” in that it gently seeks to undermine the certainties on which the conflict feeds (Winslade and Monk, 2000). The sub-model thus retains the two elements from the original narrative mediation model: dialogical practice and relationship practice (shown in Figure 2). In addition, the sub-model is complemented by adding a new activity: writing a good story. The original model strongly emphasises the importance of verbal communication, but overlooks the importance of writing a good story document relevant for RE. The elements added on this activity are thus adopted from the fields of social science in which the concept of narrative first emerged. In this activity, some good practices for writing a rich and detailed story are provided. Relationship practices Mediator guarding against • Enlistment in problem narrative • Assuming an expert knowing position Calling parties into a co-authoring relationship Encouraging Affirming trust and respect

Dialogical practices Curious, persistent, and resilient in questioning Discursive listening Using externalizing language Identifying and deconstructing dominant problem discourses

Conflict definition

Documenting and writing a conflict story Using structured story template • Defining fleshed-out characters • Defending the detailed settings • Defining parties’ goals • Dentifying causality • Defining dramatic element

Figure 2. Conflict definition sub-model

2.3 Sub-Model-3: Conflict Resolution Once the conflict itself is clearly defined, a mediator then can begin to find a solution. In the context of RE, this can be understood as the requirements engineer inviting the conflicting stakeholders to propose their preferred solutions as the alternatives for the conflicting situation. Figure-3 shows that sub-model-3 retains two activities from the original model: dialogical practice and relationship practice. In addition, to help parties reach a fairly objective decision, a semi-quantitative RE prioritization technique by (Wiegers, 1999) is integrated with the original model. Relationship practices Celebrating and honouring moves towards redecoration of relationship Being open to new possibilities Mediator invites parties to identify with the alternative account of relationship Mediator seeks to learn from parties’ experience Describing impact of parties’ alternative story on mediator

Dialogical practices Documenting change Drawing attention to unstudied experience Engaging with parties in crafting an alternative story Recruiting audience to alternative story Speaking with agency and authority Receptive to alternative stories about the relationship

Conflict Resolution

RE prioritization List all possible solutions. Estimate its relative “value” Estimate its relative “cost” Estimate its relative “risks” Calculate a priority number

Figure 3. Conflict resolution sub-model

101

ISBN: 978-972-8924-61-4 © 2008 IADIS

3. METHODOLOGY The aim of an expert panel is to provide a flexible and robust approach to eliciting unbiased evaluations from domain experts. In the SE community, the value of expert opinions is widely recognized (Dybå T., 2000). For example, Beecham et al. (2004) use a panel of experts to validate their Requirements Capability Maturity Model (R-CMM).In this section, we briefly present our methodological approach to conducting an expert panel validation of NREMM.

3.1 Sampling: Selecting an Expert In this study, a purposive sampling strategy is adopted to target a panel of specific RE experts. The experts who participated in this study are targeted from a population of the two different categories: experienced RE practitioners and senior RE researchers. These two categories of experts either publish extensively in the RE discipline, or have extensive practical experiences. These two types of experts are complementary to each other. Our narrative RE mediation model aims to be not only theoretically robust but also, most importantly, practically useful. Consequently, the experts invited from academia provide useful insights on the theoretical aspects of our model, whereas the experts invited from industry provide insights on our model’s practicality.

3.2 Data Collection Methods and Procedures Semi-structured interviews are used as our primary data collection technique. Our interview script includes a combination of closed questions and open-ended questions. Our closed questions are measured on a 3-point measurement scale - agree, disagree and don’t know. We notice that answers generated from a 3 point scale are coarse, and much information can be lost (Dyba, 2000). However, due to the small number of experts involved, viewing responses in this form eases interpretation and analysis of the data (Beecham et al. 2004). We also used several open-ended questions to elicit experts’ overall comments on the model. We conducted a pilot study involving interviewing one expert to evaluate our overall expert panel design. Our main study was conducted from February to March in 2008. In total, 10 RE experts were interviewed. Before each interview, participants were provided, via email, with details of the purpose and procedures for the interview. To safeguard the confidentiality of the data and privacy of the participants we also sent to each interviewee a statement of the ethical principles the research team would follow. To avoid any pre-judgment bias we did not provide participants with interview questions in advance. Each interview lasted 45 minutes and contained 4 sub-sessions as follows: 1. Interviewees complete a demographics questionnaire. 2. Interviewees are asked questions regarding the nature of conflict they perceive in RE. 3. The researcher presents a 25-minute introduction to the NREMM. 4. Interviewees are invited to comment on the model’s structure, presentation, theoretical underpinning, practicability, easy of use and the level of required training needs We ask for experts’ perceptions of the nature of conflict prior to making a presentation of our model. As our presentation contains some literature background regarding the nature of conflict, which may lead the expert’s response.

3.3 Data Analysis Both quantitative and qualitative data were generated during data collection. Interviews were transcribed and analysed. Due to a small number of experts involved, the analysis of quantitative data is very straightforward and based on frequency analysis. Qualitative data was analyzed using content analysis and frequency analysis. Content analysis involves establishing classifications, and systematic links between them, and then counting the number of instances based on the pre-defined classifications. A formal inter-rater reliability test was conducted. We also calculated the Cohen’s kappa (k) statistic - 0.78, which indicates a “Substantial” level of confidence in classification.

102

IADIS International Conference ICT, Society and Human Beings 2008

4. FINDINGS 4.1 Demographics Figure-4 shows the profile of the experts who participated in the study. Most of the experts (9 out of 10) have more than 10-years of RE experience. The only expert who has less than 10-years of experience in RE is a well-known RE scholar with a PhD in human computer interaction (HCI). In this study, we deliberately targeted experienced RE practitioners, as we aim NREMM to be practically useful. We are also included experts from different disciplines with slightly different theoretical viewpoints. Consequently our experts have software engineering, requirements engineering and information systems backgrounds.

1 2

Roles A mix of academic and practitioner Academic

Years of experience 11

Highest Qualification PhD in IS

Company or institute University

7

PhD in HCI

University

3

A mix of academic and practitioner

12

PhD in CS & PhD in organizational behaviors

University

4

14

PhD in IS

University

5

A mix of practitioners Practitioner

11

MSc in IS/IT

IT Consulting company

6

Practitioner

14

MSc in management

IT Consulting company

7

Practitioner

11

Phd in management

IT Consulting company

8

Practitioner

12

MSc in managEment science

Independent consultant,

9 10

Practitioner Practitioner

12 13

BSc in CS MSc in information system

IT Consulting company IT Consulting company

academic

&

Figure 4. Demographics of 10 experts

4.2 Perceived Nature of Conflict This section presents our findings regarding the experts’ perception of the nature of conflict in terms of its occurrence, causes and consequences. This data helps us to justify the motivation for building a conflict resolution model in RE. It also helps us gain new insights from practitioners to improve our theoreticalbased model.

4.2.1 Occurrence of conflict in RE Table-1 indicates that 9 out of 10 experts agree with our proposition. Although an expert disagrees, he acknowledges that conflict “does often occur in RE, but not inevitably” (expert-1). This result strongly supports literature reports on conflict in RE (Easterbrook, 1996; Sommerville and Sawyer, 1997; Nuseibeh, 1998). Table 1. Occurrence of conflict in RE Questions/propositions Conflicts among different stakeholders are an inevitable part of RE process.

Agree

Disagree

Don’t know

9

1

0

4.2.2 Key causes of conflict in RE We also are interested in knowing what experts’ perceive to be the causes of conflicts. We use an openended question to elicit the causes of conflicts from the experts. Experts suggest many different factors

103

ISBN: 978-972-8924-61-4 © 2008 IADIS

causing conflict. To identify the key causes of conflict, we use the following criterion approach (Niazi et al, 2005): If a cause of conflict is cited in the interviews with a frequency of >=50%, then we treat it as a key cause of conflict. Table-2 suggests that differences between goals, cognitive understanding, education background, and organizational roles are four most frequently cited causes of conflict. For example, expert-2 said:” The causes of conflict can be different understanding of the same problems and different individual or organizational goals. Most importantly, conflict occurs often due to people’s different holistic perceptions on the whole project rather than different perceptions on certain aspects of software requirements.” Expert-5 pointed out two other causes that “people always approach the same problem with sorts of different understanding. They try to be different. This is particularly inherent between people from different departments with different organizational roles and education background.” Table 2. Four key causes of conflict in RE Causes Goals Cognitive understanding Education background Organizational roles

Definition Stakeholders have different organizational and individual goals on the project Stakeholders have different understanding on the same problem. Stakeholders have different education background. Stakeholders have different organizational roles and responsibility within an organizational

Frequency 10 out of 10/ 100% 9 out of 10/ 90% 6 out of 10/ 60% 5 out of 10/50%

4.2.3 Consequences of Conflict We found that conflict leads to many negative consequences in RE (see table-3). For example, 9 out of 10 experts think conflict leads to disagreement between stakeholders. 8 out of 10 experts view conflict leads to poor user satisfaction on final system. 6 out of 10 experts view conflict as leading to negative emotions, resistance and failure to collaborate. These negative consequences do not surprise us. Recent empirical findings indicate that conflict consistently and negatively affects software project success and team performance (Yeh and Tsai, 2001). However, far fewer of them (2 out of 10 experts) think the overall success of the project is impacted. Table 3. Consequences of conflict in RE Questions/propositions 1. Conflict leads to disagreement 2. Conflict leads to poor user satisfaction on final system 3. Conflict leads to resistance and failure to collaborate 4. Conflict leads to negative emotions 5. Conflict leads to ambiguity in RE specification 6. Conflict leads to inconsistency in RE specification 7. Conflict leads to overall project failure 8. Conflict promotes creative thinking 9. Conflict encourages stakeholder involvement

Agree 9 8 6 6 3 2 2 2 0

Disagree 0 0 1 0 1 2 6 3 7

Do not know 1 2 3 4 6 6 2 5 3

Interestingly, we also found that conflict does not necessarily lead to inconsistency and ambiguity in the specification. This appears opposite to what is reported in the RE literature. For example, Easterbrook (1996) indicates that conflict is mainly characterized by disagreements, and those disagreements may lead to inconsistency in the specification. Sawyer (2002) says that conflict may be caused by many different and conflicting viewpoints, and therefore may lead to ambiguity in the specification (Sawyer, 2002). There is also lack of strong evidence to claim that conflict promotes creative thinking or encourages stakeholder involvement. Nonetheless, expert-7 points out that Conflict certainly can be a good thing, but it all depends on how it is brought out, when it is brought out and most importantly how well it is being managed. If it has been surfaced early and managed well, it will definitely lead to a better understanding of the problem situation and a better requirements specification. We finally asked all experts whether conflicts need to be explicitly managed. All 10 experts agree the necessity of managing conflicts in RE. In particular, a very experienced expert (expert-10) says:

104

IADIS International Conference ICT, Society and Human Beings 2008

“Based on my practice, users often had conflicting attitudes towards each other and developers. They were enmeshed in politics and power games. In short, they behaved just like human beings. In many cases, I was required to be a mediator between them. I was soon aware that I need to learn more from other soft disciplines, and your social-psychology based mediation approach certainly seems very interesting to me and timely to address this problem.”

4.3 Three Sub-Models Validation Results This section discusses whether NREMM meets its objectives. As mentioned in Section 2, NREMM includes 3 sub-models. We thus validate whether each individual sub-model meets its individual objective.

4.3.1 Results for conflict identification sub-model Table-4 presents our findings regarding the validation of the conflict identification sub-model. As mentioned in Section 2, this sub-model aims to establish a workable relationship with the conflicting parties and initially identify conflict between them. Our results show that 9 out of 10 experts agree with the importance of making conflict “surface early” (expert-3). Expert-3 also says: “conflict often remains as hidden rather than surfaced. It more likely occurs when different perceptions, goals, and motivations have not been brought out and resolved early on.” Results also show experts agree that the techniques described in the sub-model meet its aim. In particular, experts recognise the value of techniques described in dialogical practice. For example, expert-5 notices that “in my practice, two key success factors are using the right questioning and listening skills to uncover the real requirements.” However, some experts (4 out of 10) think that the techniques described in relationship practice are “just common sense, and don’t need to be highlighted in this sub-model (expert-8).” Table 4. Findings for conflict identification sub-model Questions/propositions 1. The sub-model, which focuses on clearly identifying conflict, is important for RE conflict resolution. 2. The discursive listening and curious questioning techniques help practitioners clearly identify conflicts. 3. Stakeholder modelling techniques help practitioners get sufficient background information of conflicting stakeholders. 4. Relationship practice techniques help practitioners develop a strong starting relationship with conflicting stakeholders. 5. The techniques described in the sub-model are easy to understand.

Agree 9

Disagree 0

Do not know 1

9

0

1

10

0

0

6

0

4

9

0

1

4.3.2 Results for Conflict Definition Sub-Model Table-5 presents the results regarding the validation of conflict definition sub-model. As mentioned in Section 2, this sub-model aims to gain an accurate understanding of conflict. Our results show that all 10 experts agree with the aim of conflict definition sub-model. Experts also mainly agree the techniques described in this sub-model are easy to understand, and fit for its purpose to gain an accurate understanding of conflict. However, the use of a rich and detailed story is questioned by the experts. Our results show only 4 out of 10 experts agree with our proposition-4 in table-5. Indeed, most of the experts appreciate the value of a rich and detailed story, which aims to provide new insights and deep understandings of a conflicting situation. However, “it is simply not realistic in practice (Expert-8)”. Expert-5 also says: “we are certainly interested with the story told by users. But, it’s not possible to document it in a full, detailed, and structured story form. We more prefer documenting and representing it in a form of model e.g UML.”

105

ISBN: 978-972-8924-61-4 © 2008 IADIS

Table 5. Findings for conflict definition sub-model Questions/propositions 1. The sub-model, which focuses on gaining an accurate understanding of conflict, is important for RE conflict resolution. 2. The curious, persistent, and resilient questioning techniques help practitioners gain an accurate understanding of conflict. 3. Using externalizing language helps separate the problem from the problem. 4. A rich and detailed story helps practitioners gain new insights and better understandings of conflict. 5. The techniques described in the sub-model are easy to understand.

Agree 10

Disagree 0

Do not know 0

9

0

1

10

0

0

4

3

3

9

0

1

4.3.3 Results for Conflict Resolution Sub-Model Table-6 presents the results regarding the validation of conflict resolution sub-model. As mention in Section 2, this sub-model aims to lead to a solution. Although our results show that all experts agree there is a need to find a solution, experts show a lack of confidence with the use of requirements prioritization techniques to achieve this aim. For example, expert-7 says “different groups of stakeholders tend to prioritize requirements differently…e.g., they tend to prioritize a certain requirement higher when they are the direct benefitor from the execution of the requirement; even different stakeholders within the same group prioritize them differently because of their different individual needs, experiences, and hidden agendas.” Expert-8 also notices that “Requirements prioritization is always dominated by project politics. The result is thus likely subjective and biased. It only plays a supporting role. I personally do not think there will be a magic solution for this.” Table 6. Findings for conflict resolution sub-model Questions/propositions 1. The sub-model, which focuses on finding out a solution, is important for RE conflict resolution. 2. Conflict can be more easily resolved when stakeholders have established a firm and collaborative relationship. 3. Requirements prioritization can offer practitioners a rational analysis to evaluate possible solutions. 5. The techniques described in the sub-model are easy to understand.

Agree 10

Disagree 0

Do not know 0

9

0

1

4

2

4

9

0

1

4.3.4 Overall impression on NREMM At the end of each interview, we use the open-ended questions to ask for their overall impressions of the NREMM in terms of its structure, presentation, theoretical underpinning, practicability, ease of use, and the possibility of being adopted by practitioners. The key findings are summarised as follows: • NREMM is capable of helping RE practitioners to identify, define, and resolve conflicts in the RE process. • NREMM is easy to understand by these 10 experts. • NREMM has a clear structure and presentation. • The use of story-telling concept as a theoretical underpinning matched well with the current state of the RE practice. • NREMM can be applied as a general framework to most project types. It particularly suits those businessoriented Information Systems, which are enmeshed in a complex human, social and organizational environment. • NREMM needs further improvements in terms of integrating with some RE modelling and goal analysis techniques. • NREMM needs to be supplied with a set of guidelines to enable practitioners to correctly implement it in practice.

106

IADIS International Conference ICT, Society and Human Beings 2008

5. CONCLUSIONS In this paper a validation of the NREMM through interviewing with a panel of experts is presented. The overall validation results show that NREMM has the potential to help practitioners to deal with conflicts in the RE process. However, the results also indicate that NREMM needs further improvement to make it more practical-oriented and more likely being adopted by practitioners. We also found out that an expert panel is particularly suitable in the early stages of the validation work to use small samples to gain insights on newly developed technologies. In the future, NREMM model is also about to be tested in a real-life context.

REFERENCES Barki H, Hartwick J, (2001) Interpersonal conflict and its management in information system development, MIS Quarterly Vol. 25 No.2, pp195-228, June 2001 Beecham, S., Hall, T., Britton, C., Cottee, M., Rainer, A., (2004). Using an expert panel to validate a requirements process improvement model. Journal of Systems and Software, Volume 76, Issue 3, June 2005, Pages 251-275 Boehm, B. Grünbacher P. and Briggs R. (2001) Developing Groupware for Requirements Negotiation: Lessons Learned, IEEE Software, 18(3), 2001. Damian, D.E. (2003) A research methodology in the study of requirements negotiations in geographically distributed software system, Proceedings of 11 the IEEE International Requirements Engineering Conference, 2003 Dybå T., (2000) An instrument for measuring the key factors of success in software process improvement, Empirical Software Engineering 5 (2000) (4), pp. 357–390. Easterbrook S. M. (1996) “Resolving Requirements Conflicts with Computer-Supported Negotiation”. In M. Jirotka & J. Goguen (eds) Requirements Engineering: Social and Technical Issues, Pp41-65. London: Academic Press. Jackson M. (2001) “Problem Frames: Analysing and Structuring Software Development Problems” Addison-Wesley Longman Publishing Co., Inc. Ma N., Hall T., Barker T. (2008a) Building a narrative based Requirements Engineering Mediation Model, a paper accepted on EuroSPI 2008 conference, 2-5, Sep, 2008, Dublin, (to be appeared in Lecture Note in Computer Science) Ma N., Hall T., Barker T. (2008b) Using Mediation Theory to Build a Requirements Conflict Resolution Model, a paper submitted on WRE 2008, 12-13, Sep, Barcelona, Catalonia, Spain Nuseibeh B., (1996) To Be And Not To Be: On Managing Inconsistency in Software Development , Proceedings of 8th International Workshop on Software Specification and Design (IWSSD-8), pp164-169, Scloss Velen, Germany, 2223 March 1996, IEEE CS Press. Robinson, W.N. (1990) Negotiation Behaviour During Multiple Agent Specification: A Need for Automated Conflict Resolution, Proceedings of 12th International Conference on Software Engineering (ICSE-12), 268-276, Nice, France, IEEE Computer Society Press, March 1990. Sawyer S. (2001) Effects of intra-group conflict on packaged software development team performance, Journal of Information System. 11(2), p. 155, 2001 Sommerville I. and Sawyer P. (1997) Requirements Engineering: A Good Practice Guide, Wiley, 1997, Van Lamsweerde, A. Requirements Engineering in the Year 00: A Research perspective, Invited Paper for ICSE'2000 22nd International Conference on Software Engineering, Limerick, ACM Press, June 2000. Winslade J. and Monk G. (2000) “Narrative Mediation: A New Approach to Conflict Resolution” Jossey-Bass Wiegers Karl E., (1999) First Things First: Prioritizing Requirements, Software Development, Vol. 7, no. 9 Yeh Quey-Jen and Tsai Chih-Ling, (2001) Two conflict potentials during IS development, Information and Management, v.39 n.2, p.135-149, December 2001 Yu, E. and Mylopoulos, J. Why Goal-Oriented Requirements Engineering, Fourth International Workshop on Requirements Engineering: Foundation for Software Quality (REFSQ'98), E. Dubois, A. Opdahl and K. Pohl (ed.), Pisa, Italy, 1998.

107