Investigating the Role of 'Soft issues' in the RE Process

2 downloads 123682 Views 238KB Size Report
Mar 11, 2009 - A method aimed at improving elicitation and analysis of such 'soft' issues is described. The method consists of a taxonomy of users' values, ...
16th IEEE International Requirements Engineering Conference

Investigating the Role of ‘Soft issues’ in the RE Process Sarah Thew & Alistair Sutcliffe Manchester Business School, University of Manchester Booth Street East, Manchester M15 6PB, UK [email protected] [email protected]

(OED). In the context of requirements engineering we are interested in the principles and norms users associate with their work role. For a research scientist this might mean valuing rigour and accuracy, resulting in a commitment to methodical and precise working practices, or valuing innovation resulting in an enthusiasm for creative working and data exploration. A useful conception of ‘value’ is to consider it as a complex memory schema which is used to interpret events, communications or situations we encounter. Values are complex concepts or knowledge schema, related to beliefs and attitudes, which shape our response to events; they are also related to motivations and ultimately our personality. Our responses and reactions result in actions, which may also give rise to feelings and emotional responses. Understanding values and emotions helps requirements engineers interpret the concerns held by individuals and to predict their actions and responses. Our method aims to facilitate this process. Based on the literature and a series of interviews with expert analysts, we developed an approach which encourages reflection by the analyst on their users’ values and emotions. The taxonomy of values and their consequences for process guidance are illustrated in Table 1. Nine upperlevel value categories are proposed based on Rescher’s theory [5] and our own investigations from card sort experiments and expert interviews. Six categories accord with generally recognised concepts: trust, morals, aesthetics, security, sociability and creativity/innovation. These terms hide many subcategories, some of these are given in the related terms column. Personal characteristics is also diverse, so we propose using personality theory dimensions (introvert/extrovert, sensing/intuition, thinking/feeling, judging/perceiving), with some additions. Unlike other values, personal characteristics are simpler attributes which describe people. Personality characteristics are closely related to motivations and both have

Abstract Soft issues, such as politics and people’s feelings, are often cited as problems in the RE process. A method aimed at improving elicitation and analysis of such ‘soft’ issues is described. The method consists of a taxonomy of users’ values, motivations and emotions, with process guidance for eliciting and analysing these issues for the RE process.

1. Introduction Soft issues, such as politics and people’s feelings are often cited as problems in the RE process [1], although there is little advice about how to deal with these issues. Few studies have directly considered stakeholders’ emotions during the analysis phase, though there have been numerous studies which report the impact of negative user emotions after implementation e.g. [2]. Value in RE has been analysed in the sense of worth or monetary value, notably the e-value method which had been applied to business process analysis and socio-technical systems requirements [3]. However, values may also be interpreted as a set of issues which are frequently referred to as problematic in the RE process, such as politics, culture and sensitivities about the consequences of automation [1, 4]. This paper discusses the role of values and emotions in the RE process and attempts to taxonomise the concepts of values, motivations and emotion as they relate to RE.

2. Method Development Definitions of value vary from worth and desirability, to judgement of what is important in life

1090-705x/08 $25.00 © 2008 IEEE DOI 10.1109/RE.2008.35

63

Authorized licensed use limited to: IEEE Xplore. Downloaded on March 11, 2009 at 10:07 from IEEE Xplore. Restrictions apply.

implications for team management in the RE process and customisation for personal RE. Motivations are a placeholder for a more detailed taxonomy, while beliefs and attitudes are a diverse category including socio-political, cultural and religious beliefs. Column three suggests some potential conversation topics which might expose particular values.

The process implications in column 4 vary from organising the team composition in response to aesthetic needs (e.g. include aesthetically aware designers), to specialisation of the RE process to include safety and risk analysis. In many cases, especially with motivations, beliefs and attitudes, value analysis may both alert the analyst to potential stakeholder conflicts, and help the analyst better understand the causes of those conflicts. When a value is absent or not strongly held, the only consequence is usually not to invest in the subprocesses for aesthetics, security, and creativity, etc; but the absence of, or negative values for, trust, sociability or morals will need corrective action to ensure a productive relationship among the project team. Table 2. Motivations and their consequences.

Table 1. Values: elicitation hints and implications for RE process management. Value Related Potential Process concept terms sources implications Trust Openness, Relationships Less control integrity, with other checks, loyalty, individuals improved team respons/departments confidence ibility, Privacy reliability policies SociaCollab Relationships Improved team bility oration, with others, cooperation, cooperawareness of shared ation, office politics awareness friendship, altruism Morals/ Justice, Behaviour Open process, Ethics fairness, toward others, use of workequality, opinions of shops to tolerance others’ promote behaviours inclusivity CreatOriginalWork Include in ivity, ity, processes, problem Innovaadventure, problem solving & tion novelty solving creative challenges Aesthet- Beauty, Appearance, Designers as ics nature, art reaction to team members, images and storyboards design Security Safety, Data managehazard/ threat privacy, ment policies, analysis risk Attitudes to change Personal serious/ Self image, Customisation charactplayful, personae, analysis for eristics introvert/ scenarios, personal RE, extrovert, psychological team conflict systematic questionnaires management /opportunistic Motivat- Ambition, Goals, Stakeholder ion achieveambitions analysis, ment rewards for see also members Table II Beliefs Cultural, Leisure Stakeholder & political, interests, user analysis, team attitudes religious background, composition topics reaction to news events

Motivation Power

Possession Achievement Selfesteem Peeresteem Selfefficacy Curiosity, learning Sociability Altruism

Description Need to control others, authority, command Desire for material goods, wealth Need to design, construct, organise Sense of selfworth Need to feel valued by others Confidence in own capabilities Desire to discover, understand world Desire to be part of a group Desire to help others

Implications Work organisation, responsibility, control hierarchy Resource control, incentives, marketing Goal oriented, coordinate personal and project aims Link personal & project goals, praise personal achievement Team composition, social feedback & rewards, Confidence building, training, skill matching Extensible systems, self tutoring Collaboration in work organisation Cooperation in work organisation

Motivations are important for understanding stakeholder groups and for individual-level requirements when systems can be customised or configured. Table 2 summarises the more important motivations for requirements analysis, synthesised from Maslow’s motivational theories [6]. Motivations are not easy to detect so elicitation guidance from the description in column 2 can only provide hints to guide identification, some of which are suggested by the motivation type itself, i.e. expressions of interest in learning, or willingness to help others. Column 3 suggests implications for personal goals and needs for each motivation type; for example, self-efficacy,

64

Authorized licensed use limited to: IEEE Xplore. Downloaded on March 11, 2009 at 10:07 from IEEE Xplore. Restrictions apply.

of a new system. Frustration may lead to the stronger emotion of anger expressed overtly; more often frustration is not so obvious. The causes may be lack of involvement in the requirements analysis process, being marginalised at the expense of other stakeholders, or not having one’s contributions or values discussed. Frustration in the long term will be manifest in stress, leading to illness and absenteeism; in the short term users may be uncooperative and uncommunicative. Finally, disgust and revulsion are strong emotional responses which are usually obvious from the users’ reference to a design feature.

curiosity and learning point towards the need for extendable systems where the user can take control and customise or program additional functionality. Altruism and sociability suggest people who will collaborate and cooperate with others in group working, both in the delivered system and in the system development process. Motivations of self- and peer-esteem can indicate designing systems to suit individual needs; for instance, in e-commerce, marketing tools can be customised to praise customers and thereby improve their self-esteem. Table 3. Emotional responses and their possible causes. Emotion Related Possible causes Remedial feelings action Fear Fright, Project /design Review & worry, is threatening remove threat threats Pleasure Joy, Project /design None happiness is rewarding Anxiety Uncertain- Project /system Explain ty, worry consequences specification, not clear, little reassure involvement users Frustrat- AnnoyConflict, valueRevisit ion ance, interest clashes, stakeholder anger values ignored analysis Disgust Revulsion, Design has Radical horror complete clash design with values/ review culture Depress WithLack of involve- Re-engage -ion drawn, ment in process, users, isolated, values ignored improve communicati on & motivation

3. Using the Taxonomy Questionnaires are effective for surveying values and emotions within a population; however, they restrict investigations to a pre-defined set of values and provide no scope for respondents to expand their opinions. Therefore our method provides a systematic approach to the analysis of rich data gathered during the course of requirements elicitation. Interviews combined with ethnographic techniques help gather insight into values and emotions from implicit expression of users’ likes and dislikes in verbal and non-verbal communication. Scenarios, storyboards and evaluation of prototypes help to elicit user feedback when their comments reveal how the system reflects (or does not reflect) their values or motivations. The key steps of the method are summarised in Figure 1. Novice analyst Preliminary analysis

Emotions can give useful feedback about project plans and designs, and may indicate potential problems leading to user dissatisfaction or system rejection. Emotions can be detected via body language, voice tonality and facial expressions. Strong emotions are hard to miss (e.g. anger) but others are less obvious, e.g. a frown could indicate frustration, anxiety or just being puzzled. The more important emotions and their consequences are given in Table 3. These emotions are based on the classification of emotions as responses (positive or negative) to events, people or artefacts by Ortony et al. [7], who list 22 separate emotions. Pleasure, joy and happiness are all positive responses, so no remedial action is needed (see column 4 in table 3). Fear is usually overt but may be tacit, for instance when a new system threatens someone’s job security. Anxiety is a milder form of this emotion, potentially arising from uncertainty, and fear of the consequences

Hunches

Taxonomy of values, motivations & emotions Interviews, observations, meetings etc.

Transcribe interviews

Internalise taxonomy and method Preliminary analysis Hunches

Review interview notes

Annotate transcripts

Revised hunches

Expert Analyst

Refined understanding of values, emotions, motivations

Revised hunches

Non functional & functional requirements Process Implications

Figure 1. Process stages, and expert and novice pathways.

65

Authorized licensed use limited to: IEEE Xplore. Downloaded on March 11, 2009 at 10:07 from IEEE Xplore. Restrictions apply.

but also recommendations for project organisation, functional allocation and work design. In expert mode, the taxonomy and process will be learned in training courses and by preliminary experience. The method knowledge is internalised so it can be used to formulate appropriate questions framed by the analyst’s understanding of the application domain. The method becomes part of the expert’s battery of techniques in scenario analysis, and questioning using storyboard and prototype probes. An elaboration of the expert mode is to use the internalised knowledge of the taxonomy to review interview notes, using Tables 1-3 as aide-memoires.

The method can be used in two modes to suit novice or expert analysts and the time resources available. Preliminary analysis leads to identification of key issues or ‘hunches’, i.e. a sub-set of the users’ values, motivations or emotions from the taxonomy which are potentially relevant. For example, we may suspect that our users prioritise certain aspects of their work: creativity, aesthetics, collaboration. Making these intuitions explicit encourages evidence gathering to support or challenge initial hunches. In novice mode interview audio-recordings are transcribed. The transcript is marked up, summarising each snippet of evidence with a short name, as shown in table 4. Table 4. Sample of a marked-up transcript. Italics and underlining are used to show the links between the transcript and the annotation, with the taxonomy categories in brackets. Transcript Annotations Interesting as in a good Looking for opportunities position to do this kind of work – and would like to Positioning of teamdevelop links with external recognition Experimental Medicine given the upcoming (personal development, rearrangement of schools. self and external esteem) I’m still in the literature. I Thoroughness want a good understanding of the problem if it’s (Systematic) tricky.

4. Conclusions The contribution of this paper is to propose a taxonomy and analysis method to deal with soft issues in RE that complements existing analysis of nonfunctional requirements. The method introduces new considerations into the RE process by drawing attention to stakeholders’ values, motivations and emotions. This analysis has implications for management of the RE process, as well as providing input to the definition of requirements for personalisation and customisation. Future work will focus on validating the method by testing via case studies with industrial practitioners and giving tutorials to evaluate the method’s comprehensibility and utility.

If time resources are constrained, an alternative is to use expert review of audio recordings to make notes. During the review a simple checklist prompting the analyst on likely sources of values, motivations and emotions is used to support the annotation process. In addition the value, motivation and emotion tables may be used in conjunction with existing taxonomies of non-functional requirements which already provide lists of quality attributes, e.g. accuracy or reliability [8]. Most values and motivations held by the user (self) will be positive, but those related to colleagues, design features and the environment may be negative, indicating potential issues for the project or design. Emotional responses of frustration, anxiety and fear during interviews provide corroborating evidence of such problems. Once the annotation is complete the transcript is inspected for frequent value and motivation categories, possible causations thereof, and differences between individuals and groups. The implications of the analysis for both the project process and the design are reviewed by referring to the tables and the annotated notes. The output is lists of functional and nonfunctional requirements, familiar in most RE methods,

References [1] S. [2]

[3] [4] [5] [6] [7] [8]

Robertson and J. Robertson, Mastering the Requirements Process: Addison Wesley, 1999. A. C. W. Finkelstein and J. Dowell, "A Comedy of Errors: The London Ambulance Service Case Study," in 8th International Workshop on Software Specification and Design, Schloss Velen, Germany, 1996. J. Gordjin, E. Yu, and B. v. d. Raadt, "E-Service Design Using i* and e3value Modelling," IEEE Software, vol. 23, pp. 26-33, 2006. E. Mumford, Values, Technology and Work: Martinus Nijhoff Publishers, 1981. N. Rescher, Introduction to Value Theory: Prentice-Hall Inc, 1969. A. H. Maslow, R. Frager, C. McReynolds, R. Cox, and J. Fadiman, Motivation and Personality. New York: Addison Wesley-Longman, 1987. A. Ortony, C. G, and C. A, The Cognitive Structure of Emotions: Cambridge University Press, 1990. G. C. Roman, "A Taxonomy of Current Issues in Requirements Engineering," Computer, vol. 18, pp. 14-23, 1985.

66

Authorized licensed use limited to: IEEE Xplore. Downloaded on March 11, 2009 at 10:07 from IEEE Xplore. Restrictions apply.