Emotional Attachment Framework for People-Oriented Software

Emotional Attachment Framework for People-Oriented Software

Mohammadhossein Sherkat School of Computing and Information Systems
The University of Melbourne
VIC 3010, Australia
sherkatm@unimelb.edu.au
Antonette Mendoza mendozaa@unimelb.edu.au Tim Miller tmiller@unimelb.edu.au Rachel Burrows rachel.burrows@unimelb.edu.au
Abstract

In organizational and commercial settings, people often have clear roles and workflows against which functional and non-functional requirements can be extracted. However, in more social settings, such as platforms for enhancing social interaction, successful applications are driven more by user emotional engagement than functionality, the drivers of user engagement are difficult to identify. A key challenge is to understand people’s emotional goals so that they can be incorporated into the design. This paper proposes a novel framework called the Emotional Attachment Framework that is based on existing models and theories of emotional attachment. Its aim is to facilitate the process of capturing emotional goals in software engineering. To demonstrate the framework in use, emotional goals are elicited for a software application that aims to provide help for homeless people. The outcomes are evaluated by domain experts and compared with an alternative approach. The results indicate that the Emotional Attachment Framework has the potential to help system analysts uncover additional emotional goals as well as provide valuable insights into these emotional goals.

Abstract

A new relationship between software and people is emerging. In organizational and commercial settings, people often have clear roles and workflows against which functional and non-functional requirements can be extracted. However, in more social settings, such as platforms for enhancing social interaction, successful applications are driven more by user emotional engagement than functionality, the drivers of user engagement are difficult to identify. A key challenge is to capture and understand users emotional requirements so that they can be incorporated into the interaction design.

This paper proposes a novel framework called the User Emotional Attachment Framework that is based on existing models and theories of emotional attachment to facilitate the process of capturing and understanding emotional requirements in software design. To demonstrate the framework in use, emotional requirements are elicited for a software application that aims to provide help for homeless people in Australia. The results are evaluated by domain experts and compared with an alternative approach. The results indicate that the User Emotional Attachment Framework has the potential to help designers uncover additional emotional requirements as well as provide valuable insights into these emotional requirements. The analysis shows that the root of majority of homeless people’s emotional requirements goes back to their needs for expressing personal identity as well as seeking pleasure.

keywords:
People-Oriented Software, Requirements Engineering, Emotional Goals, Emotional Attachment Framework
journal: Journal of Systems and Software

1 Introduction

Computing is not about computers any more. It is about living.” Nicholas Negroponte (calvo2014positive, , p. 197).

The main goal of Requirements Engineering (RE) is to understand what people want and desire, so that we can design and implement a successful software system guinan1986development (); dieste2008understanding (); gonzales2011eliciting (); colomo2010study (). In this process, stakeholders must decide on the most effective combination of software features. To make this decision, two critical activities are requirements elicitation and requirements modelling.

This process has been described as ‘getting into someone’s head’ to capture his or her needs and desires guinan1986development ().

Despite the maturity of existing requirements elicitation processes, many software applications struggle engage people

clancy1995standish (); OASIG1995 (); whittaker1999went (); tichy2008business (); gonzales2011eliciting (), ultimately leading to people rejecting the software van2001interactive (); platt2007software (); Dix:2003:HI:1203012 (); 1663532 (). It is widely argued that a major concern is that software engineers seem to only focus on traditional functional and quality requirements and overlook key drivers of engagement such as people’s emotions and values bentley2002putting (); draper1999analysing (); gogueny1994requirements (); hassenzahl2001engineering (); krumbholz2000implementing (); miller2015emotion (); proynova2011investigating (). It is therefore important to take into account software requirements that will create the desire to engage with the system - in addition to functional and quality requirements. These requirements may be related to social values or emotions that people wish to feel, what we call emotional goals in this research sutcliffe2010analysing (); miller2012understanding ().

Addressing emotional goals in design is not straightforward and software engineers face several barriers due to the nature of emotions. First, identifying and understanding people’s emotional goals is difficult as they are the subjective attributes of people and not the property of software callele2006emotional (); mendoza2013role (). People rarely are able to express their emotions directly. Although people recognize their emotions when they encounter them, it is difficult for them to articulate them beforehand since emotions are a subjective part of their consciousness Goguen:1994:RER:177970.184582 (). Second, even if people are able to express emotional goals, there is limited knowledge about how these emotions can be captured and then translated into concrete requirements as part of design salzer1999atrs (). Third, the unstructured and ambiguous nature of emotional goals make converting them to a software specification difficult. Thus, the lack of a systematic methodology to map such intangible emotions to a tangible set software features remain a challenge. Consequences related to overlooked emotional goals are particularly profound in social applications such as platforms for enhancing social interaction, social networking and public health software systems. In these types of software systems users posses a variety of characteristics, including their needs, goals, motivations, and lifestyle that ultimately influence the ability of software to appeal and engage. In this paper we define these types of software systems as People-Oriented Software (POS) systems. In POS systems, requirements elicitation is challenging as people are often not obliged to use a software system, do not generally have well-defined roles and responsibilities, have different cultural and social backgrounds, and do not necessarily have the same requirements as each othermiller2015emotion ().

Reviewing the requirements engineering literature shows that the above mentioned barriers do not allow emotional goals to be easily understood, elicited and analyzed by using traditional requirements engineering methods miller2015emotion (); thewvalue (). Traditional requirements engineering methods capitalise on well-defined business work-flows and tasks that are less applicable in contexts that are not as structured or protocol-driven like emotional goals in POS.

In terms of using scientific theories of emotions in software engineering, there are numerous theories and frameworks such as Parrott’s Emotions by Groups parrott2001emotions (), Plutchik’s Wheel of Emotions plutchik2003emotions () and the Hourglass of Emotions cambria2012hourglass () that have categorized the human’s emotion. While some ideas from these theories can be used in considering emotional goals in software engineering, they often are not suitable for the design of software systems for several reasons: (i) they typically assume emotions operate independently of users’ cognitive processes; (ii) they do not model the fact that people may experience two or more emotions at the same; (iii) they cannot be used to describe the complex range of emotions that people may feel in daily use of a software system; and (iv) are not contextualized in everyday life (i.e. they are too abstract) cambria2012hourglass (); smith2009critiquing (). These limitations motivate the need for bridging the gap between the psychological frameworks and software engineering. Accordingly, a key question at the heart of successful requirements engineering activities is: How can we get a better understanding of people’s emotional goals in designing a software system?

In this paper we investigate the nature of emotional goals within the context of requirements engineering. Our research question is How can people’s emotional goals be elicited in the process of requirements engineering? We present a novel framework called the Emotional Attachment Framework, which is based on current theories and models of emotion and appraisal arnold1970perennial (). The framework facilitates the process of capturing and understanding emotional goals in requirements elicitation process. We also outline a process model for using the proposal framework in practice. Identifying emotional goals is an essential precondition for detailed design and development efforts, therefore the focus of this work spans this first step of capturing and understanding emotional goals.

We evaluate our proposed framework by eliciting the emotional goals in an industry case study; a mobile website to help homeless people find information about available support services. We used existing interview data to analyze and extract emotional goals. We then compared our findings with an alternative approach for eliciting emotional goals miller2015emotion (). Further, we tested the applicability and effectiveness of our proposed framework by interviewing two independent domain experts involved in this case study. The results show that our approach identified and categorized people’s emotional goals that the domain experts agreed with, and that they believe the elicited emotional requirements provide improved support for requirements engineering. Comparison analysis with the alternative requirements elicitation approach indicated that we discovered more subtle the emotional goals by using the proposed framework.

The following section discusses some main concepts, complexity of emotional goals and prior efforts in considering emotional goals in design. In Section 3, we present our framework for understanding the people’s emotional goals and this proposed framework will be evaluated in Sections 4. The last two sections of this paper are dedicated to discussion and conclusion.

2 Literature review and related work

Frog Design founder Esslinger states that: “even if a design is elegant and functional, it will not have a place in our lives unless it can appeal at a deeper level, to our emotions” (cited in (sweet1999frog, , p.9)). In this section we briefly review the nature and foundations of emotions and what emotional goals are. We also review the efforts in considering emotional requirements in design and development including software system domain.

2.1 Concepts and Definitions

The importance of considering emotions is recognized in software system design, however, this subject has been of interest in different fields of science and research such as philosophy sep-emotions-17th18th (), biology de1990rationality (), physiology de1990rationality (), sociology thoits1989sociology () and psychology Desmet2002 (). It means that the main roles and concepts of emotions have been borrowed by the field of software system design from other domains. In this section it is important to recognize the characteristics of emotional goals and the aspects which make them distinguish them from other types of requirements. In order to understand emotional goals, it is first essential to have an understanding of the concept of emotion.

2.1.1 Definition of Emotion

Although it seems the definition of emotion is clear, as Arnold arnold1970perennial () mentioned, understanding the nature of emotions, especially emotions’ characteristics, how do emotions work?, how are emotions felt and expressed?, etc. is a lasting problem. Emotions influence everything we do from the way we behave and think, to the way we communicate and make decisions. People are capable of a vast range of emotions, from satisfaction in daily tasks to the sadness of losing valuable things and death of a loved one ortony1990cognitive (); norman2005emotional ().

By reviewing the literature, we found several different terms for reflecting people’s different feelings. Affect, emotion, mood, passions, value, motivation and sentiment are just a few to name. Although the meaning of these words differ from one another and refer to different concepts, they are routinely used interchangeably. Among all of these words, the word “emotion” is often used in a wide variety of research domains. In this research we also use the term emotion, as this is a common term that people use to refer their subjective and hard-to-measure feelings and it is more related to people rather than software miller2015emotion (). In this research what we mean by the people’s emotions is the people’s different feelings and mental states which have different characteristics and effects on how people make decisions and behave.

2.1.2 Forming Emotion

Prior studies in the field of product design have tried to understand how consumers’ emotions form Desmet2002 (). As Arnold, the founder of ‘Appraisal Theory’ argued, people form emotions when encountering a product or service. This is the consequence of their assessment (or their ‘appraisal’) of the product or service capabilities and how it may harm or benefit themarnold1970perennial (). Based on the Arnold’s definition, an appraisal is the “direct, immediate sense of judgment of weal and woe” (cited in (cowie2001emotion, , p.9)). He also defined emotion as the “felt tendency toward anything intuitively appraised as good or away from anything intuitively appraised as bad” (cited in (rulla2003depth, , p.33)). Based on these definitions without appraisal there is no emotion and positive or negative emotions are the consequence of appraising a product or a service beneficial or harmful respectively. In other words, people can experience different emotions such as joy, frustration and desire in response to products that they appraise as matching or mismatching with different concerns that they have.

Desmet and Hekkert argued that what causes humans’ emotions is their interpretation of a product or a service rather than the product or service itself desmet2007framework (). This interpretation is the consequence of cognitive appraisal process which often happens automatically and unconsciously in the people’s minds. Desmet used Arnold’s appraisal theory to visualize and develop this concept as the ‘Basic Model of Emotions’. Desmet’s model contains three variables: Appraisal, Concerns and Stimulus (product or service) (Figure 1) Desmet2002 (). The definitions of these three variables are as below:

Figure 1: Desmet’s Basic Model of Emotions Desmet2002 ()
  • Appraisal: According to appraisal theory, appraisal refers to automatic and unconscious evaluation of a product or service for the people’s personal concerns scherer2001appraisal (). The output of this evaluation in Desmet’s basic model of emotions shapes the people’s positive or negative emotions.

  • Concerns: Based on what Frijda has discussed, behind each emotion there is a concern Frijda2010handbook (). If people’s concerns can be addressed by product or service capabilities, the product or service is appraised as beneficial and lead to positive emotions. In the opposite side, the product or service capabilities that cannot match with the people’s concerns are appraised as harmful and lead to negative emotions. Desmet Desmet2002 () categorized concerns in two main groups, universal (such as the concern for safety and love) and culture and context-dependent (personal concerns such as concern for being home late or missing good seat at the cinema).

  • Stimulus: In Desmet’s basic model of emotions, the stimulus refers to the element labelled as ‘Product‘ in Figure 1. It represents any product or service that can address one or more concerns. Based on the Desmet’s definition, physical products, services, remembered and imaginary items can be considered in the basic model of emotions as a stimulus.

2.1.3 Emotional Attachment

Emotional attachment is an attempt to create emotional links between people and products or services. It has received a lot of attention in recent years particularly in the field of marketing 11015023620150701 (); norman2013design (); guo2014emotional (). The main basis of emotional attachment is derived from ‘Attachment Theory’ was initially developed by Bowlby to explain the relationship between the parents and infants bretherton1992origins (). However in last two decades a wider range of relationships and contexts have been encompassed. Recent studies indicate that individuals not only emotional attach to other individuals, but also to brands whan2010brand (); thomson2005ties (), pets sable1995pets (), gifts mick1990self (), places giuliani2003theory (); scannell2010defining (), collectables slater2001collecting (), firms and organizations mende2011attachment () and, objects ainsworth1969object (); page2014product ().

People experience emotional attachment at different levels of abstraction. They may emotionally attach to a product or a service in general or specific capabilities. For instance, people may emotionally attach to Facebook™ (as a general social network application) or like one or more capability(ies) (sharing media as a specific capability).

Mugge Mugge () proposed four categories of sources for creating product emotional attachment:

  • Self-expression: People desire a distinct personal identity from others, what Mugge et al., call Self-expression mugge2008product (). Based on Collins on-line dictionary, self-expression refers to the expression of personal identity such as feelings, thoughts or ideas111http://www.collinslanguage.com. Past studies argue that people become more attached to products or services if they can show their personal identity ball1992role (); kleine1995possession ().

    Further, Mugge et al. mugge2009emotional (); mugge2009development () show that more self-expressive occurs if people can ‘embed their personality’ into a product. Embedding the personality is based on the ‘self-congruity theory’, which says humans need to express and create a positive and consistent view of one self govers2005product (). Accordingly people attach emotionally to a product or services that are reflect their own ‘self-concept’.

    Sirgy sirgy1982self () shows that this ‘self-concept’ has two categories: (1) (Ideal Self), which is the view people have of themselves; and (2) (Public Self), which is the image people would like other people to have of them.

  • Affiliation: The role of affiliation in forming emotional attachments stems from the people’s desires to have relationships, to be connected, associated, and involved with others meschtscherjakov2014mobile (). Murray’s Theory of Human Personality describes people wanting to feel a sense of affiliation with others and belonging to social groups champoux2016organizational (). Although it may seem that the affiliation opposes self-expression, they can exist simultaneously. Kleine et al. argued that although people like to establish a unique identity, they are also motivated to have interpersonal connections kleine1995possession (). A good example of coexisting affiliation and self-expression is Facebook™. A Facebook™ page expresses a unique identity by writing posts and sharing individual pictures (self-expression) and simultaneously he/she is a member of specific groups (affiliation).

  • Memories: Kleine notes that people like to maintain a sense of the past, happy moments in life, and be reminded of people, occasions, places etc. that are important for them klein2012memory (). Csikszentmihalyi and Halton argued that if a product has clues or symbols for their consumers regarding “where they have been, who they are now and what they aspire to be”, these symbols can attached them emotionally to the product csikszentmihalyi1981meaning (). Memories can be both positive and negative. Accordingly, people like to remember or re-enforce positive memories and forget negative ones. Mugge contends that if people experience positive memories in using a product, they emotionally attach to it regardless of its utility and appearance mugge2010product ().

  • Pleasure: Based on the Hedonism school of thought, pleasure and happiness are the primary aim of human life Stanford (). Accordingly, if a product provides sensory pleasure and enjoyable activity, it causes emotional attachment Mugge (). Jordan defines pleasure in designing a product as “emotional, hedonic and practical benefits associated with products” (jordan2002designing, , p.11). Maclachlan et al., argued people experience pleasure in two different ways: utilitarian and reflective benefits maclachlan2009exploring (); maclachlan2009let (). In this classification, utilitarian benefits are related to products primary function, while reflective benefits refers to additional benefits.

It seems reasonable to extend the boundary of emotional attachment and assume that people can attach emotionally to a software system as a type of product or service. Based on the emotional attachment definitions in other disciplines we define this concept in software engineering as ‘the strength of the emotional relationship people experience with a software system’. People’s emotional attachment will determine whether people use the software system as a beloved tool, they simply put up with, reject it, or never appropriate it at all.

2.1.4 Emotional Goals

Based on the definitions of emotion, emotional goals can be defined as the people’s psychological or mental goals which have intrapsychic origin glanze1990mosby (). Callele et al., define emotional goals as a way to help system analysts to deliver emotional experience to people in order to achieve optimal software system experience callele2008balancing (). Lopez-Lorca et al., have considered emotional goals in two different categories. 1) Personal Emotional Goals which represent how people want to feel as a general sense and independent of any particular system (feeling independent) and, 2) System-dependent Emotional Goals which represent people’s desires towards a specific software system lopez2014modelling ().

In this research we define emotional goals as what people would or would not like to feel by using a software system. As the root of functional or non functional capabilities goes back to people’s functional or non functional requirements, we can define people’s emotional goals as what people would or would not like to feel in the process of satisfying their functional or non functional requirements. Recent research has demonstrated how emotional goals are distinct from the application’s non-functional requirementsmiller2015emotion (). If we consider the social application of Facebook™  it is possible to see how a functional requirement of connecting friends is quite different from the emotional goal of feeling connected. Alternatively, we can see how the quality goal of being secure is also quite different to feeling secure from the perspective of the user.

Emotional goals are influenced by people’s values, aspirations, socio-cultural norms and mood that people would like to experience when using a particular software system. The interplay of these social variables further reinforces the idea that emotional goals cannot be confined to immediate sensory pleasures, but are in fact more reflective and long-term.

2.2 Emotional Goals in Design

Research into capturing emotional goals in design dates back to the 1970s. Most of the research in this area focuses more generally on product design jordan2002designing (); desmet2007framework (); norman2005emotional (); hassenzahl2001engineering (); chapman2015emotionally (); schifferstein2004designing ().

Although emotional goals is not emphasised in software engineering, there is research in the field of human-computer interaction ingram1984designing (); marcus2015emotion () and emotionally intelligent software agents bates1994role (). Considering emotional goals is imperative in game design norman2013design (); salen2004rules () where the main goal of such games is to have ‘fun’ callele2006emotional (). According to Draper’s study draper1999analysing (), fun is a dependency between the game application and players goals and as such is not a property of game application itself.

Ramos and Berry ramos2005emotion () investigated software features that caused fear (as a type of emotion) to show the influence of emotion in system acceptance. They analyzed an application that stores information about mistakes within a company, and who was responsible for them to show the relationship between negative emotions and system acceptance. Although in their value taxonomy they categorized people’s soft goals in four categories including structural, social, political and symbolic, however, they did not propose specific guideline for eliciting people’s emotion.

Colomo-Palacios et al. colomo2011using () developed a method called the ‘Affect Grid’ to understand the effect of emotions on software requirements. Their results indicated that emotions were a factor to take into account in establishing requirements stability. Colomo-Palacios et al. concluded that understanding the stakeholders’ emotions includes knowing “the reliability and stability of the definition of those requirements” colomo2011using ().

Hassenzahl et al., hassenzahl2001engineering () considered the people’s emotional goals as a ‘hedonic quality’. In their study, hedonic quality is an aspect of a software system (especially in graphical user interface) that creates positive subjective experience in the people based on their emotions. According to Hassenzahl et al., hedonic qualities are non-task-oriented qualities which prevent product or service consumers from boredom and discomfort, and create motivation, stimulation and challenge. Thew and Sutcliffe thewvalue (); thew2008investigating () in their studies reviewed different issues related to poor understanding of stakeholder values and proposed a taxonomy of stakeholders‘ values, motivations and emotions (VME). In their taxonomy they suggested eight, nine and six upper level categories for value, motivation and emotions respectively. Although they proposed the elicitation guides in the form of conversation topics for each category, the proposed taxonomy makes no recommendations for solving value clashes.

Probably the most famous studies in considering the people’s emotions in designing products can be seen in Norman’s studies norman2005emotional (). Norman believes that there are three levels of emotion: visceral, behavioural and reflective. The visceral level is pre-conscious formed automatically and related to product appearance such as colours and style. The behavioural level is sub-conscious, and related to the use and experience with a product based on function, performance, and usability. The reflective level is conscious, and as Norman stated “the highest level of feelings, emotions and cognition reside” in this level. Accordingly the visceral and behavioural levels are about the “now”, while the reflective level is longer over time and is more about the “satisfaction” produced by using a product and main causes of long-term attachments. Miller et al. miller2015emotion () argue that the reflective level has been neglected in software engineering research and practice.

2.3 Eliciting Emotional Goals

Although there are limited studies about how to deal with people’s emotional goals in software engineering, perhaps the initial efforts in highlighting the importance of eliciting and capturing people’s emotional requirements was escalated in Mumford’s ETHICS method mumford2013values (). Although this method encouraged system analysts to find ways to capture people’s values and emotions in system design process, it is a questionnaire-based technique with less flexibility and people may not be comfortable with this time-consuming method.

Another technique that has been applied widely for eliciting and capturing people’s soft goals is ethnography sommerville1993integrating (); hughes1995presenting (); viller1999social (). In this technique, a system analyst gather data including observing people in their workplace and may participate in their everyday work to realize people’s soft goals. Although this technique does not require asking questions directly for capturing people’s soft goals, the quality of output depends entirely on ethnographer expertise as it does not have any guideline supporting ethnographers in soft goals elicitation process. Yu eric2009social (); eric2010social () by proposing the i* modelling notation tried to model soft goals such as as “trustworthy”, “flexible”, “minimal intrusion” or “normal lifestyle”. However, the primary contribution of i* accredited to its capability as a modelling language as opposed to an elicitation technique.

In the field of human-computer interaction, Friedman suggested Value-sensitive design approach for eliciting human values friedman1996value (); friedman2002value (). In this technique user scenarios and storyboards are used for eliciting people’ feelings toward a developed software system or its prototype in the form of cue cards. In other words, this technique does not propose any guideline for eliciting people’s soft goals such as values and emotions in the requirements engineering process. Friedman’s idea was followed by Cockton et al. cockton2009evolving () in developing the Worth Maps technique for capturing informal descriptions of people’s feelings and emotions. In some other studies such as gordijn2003value (); komssi2011integrating (); fuentes2010understanding () instead of outlining a holistic and repeatable technique, only general advice have been given for eliciting specific types of people’s feelings and emotions.

Based on what we discussed in this section, we can argue that although many of the previous methods have been successfully applied for requirements elicitation, they did not propose a systematic and repeatable process for capturing and understanding people’s emotional goals for software development purposes. The current paper aims to address the above mentioned issues and provide something specific for POS.

3 Proposed framework

To answer the research question in this study we first need to answer this question ‘How do people’s emotions form during their encounter with a software system?’. Understanding the process of forming emotions provides insight about the aspects that we need to investigate to realize emotional goals.

In this section, we first present our conceptualisation of forming the people’s emotions, which we call Emotional Appraisal Model. We then introduce our proposed framework – the Emotional Attachment Framework (EAF) – and process model for understanding and capturing the people’s emotional goals.

3.1 Emotional Appraisal Model

According to our presentation of the Basic Model of Emotion (Section 2.1.1), we can understand that people’s emotions form during and after using a software system as a consequence of people’s appraisal regarding whether a software system’s capabilities satisfies their concerns. If the capabilities match their concerns, the software system is appraised as beneficial, otherwise as harmful. Therefore, forming emotions is the outcome of match and mismatch of the people’s concerns with software systems capabilities that result in a positive emotion, a negative emotion or an absence of emotion (neutral) respectively.

In this paper, we adapt and refine Desmet’s model of emotions for our purpose and present the ‘Emotional Appraisal Model’ (EAM) in the domain of software engineering (Figure 2). In EAM, instead of considering any product, we consider a specific type of product: software applications. In addition, Desmet’s definition of a concern referred to all types of concerns that customers may have, however, we adapt it to focus specifically on concerns that are related to emotional attachment. Emotional concerns therefore comprise of a sub-set of concern types that facilitate emotional attachment, which we discuss later.

Figure 2: Emotional Appraisal Model

Figure 2 represents the relationship between emotions, concerns and a software system. The various types of positive, negative or absence of emotion are the result of match and mismatch between people’s concerns with software system capabilities. It is clear that the type and importance of emotional concerns depends both on people and the software application domain. The interplay between these aspect also varies over time. For instance in computer games, pleasure is prioritized, whereas in social network applications group affiliation could be more important.

3.2 Emotional Attachment Framework

Based on the ‘Emotional Appraisal Model’ (Figure 2), forming an emotional attachment or detachment with a software system depends on people’s appraisal of their emotional concerns and software system’s capabilities. Accordingly, we can categorize the main drivers of forming emotional attachment under emotional concerns in ‘Emotional Appraisal Model’ (Figure 2). Adding the drivers of forming emotional attachment to Emotional Appraisal Model forms a new framework that we call it in this research Emotional Attachment Framework - EAF. Figure 3 represents the schematic view of EAF.

In proposing the EAF it is our assumption that although people may have different emotional concerns, the root of their emotional concerns can always be found by reinforcing what causes people to emotionally attach to (or detach from) a software system. Identifying these will support a better understanding of emotional requirements. This perspective stems from this fact that people have a need to support their emotional sense and define and maintain their self-concept ball1992role (); kleine2004integrative (). For this purpose, we have adopted the core ideas of customer emotional attachment from the field of marketing and used Mugge’s attachment drivers that we discussed in Section 2.1.3. The main reason for this selection is that the Mugge’s model offers the most complete mode of emotional attachment in the existing literature.

As Figure 3 represents, according to the four sources of creating emotional attachment in Mugge’s study, people’s emotional concerns are driven by their aspirations related to (i)self-expression, (ii)affiliation, (iii)pleasure and (iv)memories.

1

Self-expression: People will become emotionally attached to a software system if it can help them to display their individuality to themselves and to others (e.g. status among peers). For instance, a person may experience emotional attachment with a software system for educating society about climate change as it expresses their identity of being an environment friendly person winfree2017learning (). As an other example, in an emergency systems for older people miller2015emotion () the sense of ‘independent’ or ‘empowered’ originated from this fact that using the system’s capabilities can facilitate self-expression between people. According to the concept of self-expression and its drivers, if a software application can help people to 1) express their specific identity – Ideal Self, and 2) distinguishing themselves from others (e.g. status among peers) – Public Self, it gives their potential people sense of self-expression which influences people’s preference positively and can enhance their emotional attachment.

2

Affiliation: People may also attach emotionally to a software system as it provides them with a platform to form their social identity. Accordingly, if a software system can help people to belong to a group or enhance that part of the self that needs to feel connected, it will satisfy the people’s desires for affiliation meschtscherjakov2014mobile (). For instance in a system to help elderly people feel more cared for miller2015emotion (), the older person may experience a stronger positive emotional attachment with an application that makes them feel more connected to their family members, friends, carers, or other social groups.

3

Pleasure: According to what Jordan jordan2002designing () argued within the context of product design, in addition to function and usability, people seek pleasure from a software system that they use. Therefore we can infer that people in using a software system look for different forms of pleasure including 1) Physical Pleasure: relates to people physical body and the pleasure experienced by sensory perception - five senses, 2) Social Pleasure: relates to people relationships with others and with status and image, and 3) Ideological Pleasure: relates to peoples values and beliefs jordan2002designing ().

4

Memories: In terms of memories, if a software system can help remind people of the past, a specific occasion, an important person, create sense of personification, convey cultural-religious meanings or cause feelings of nostalgia, it may cause people emotional attachment. For example, reminders about family and friends’ birthdays or bringing up old photos posted to a social networking size.

It is clear that the four above-mentioned sources of emotional attachment in EAF are intertwined, rather than independent. For instance, in designing a mobile app for educating people to adopt low carbon products, gamification can stimulate the pleasure and self-expression of people to connect them to other environmentally friendly people or green communities (affiliation). A brief description of each emotional attachment driver is found in Table 1. Based on the definitions presented in Section 2.1.3 and Table 1 we realize that ‘affiliation’, ‘social pleasure’ and ‘public self’ are all closely connected. While their definitions and underlying theories are distinct as illustrated in the framework, it is expected that goals that are related to one have a high chance of also being related to the other two due to the fact that they are all related to social drivers of emotional attachment. The dashed oval in Figure 3 represents this proximity between these three drivers.

Figure 3: Emotional Attachment Framework
Themes \Drivers Definition Source(s)
Self-expression A person’s aspiration to feel a distinct personal identity from others through expressing their personal identity such as feelings, thoughts or ideas mugge2008product (); ball1992role (); kleine1995possession ()
   Ideal-self A person’s aspiration to feel like a person they would imagine to be sirgy1982self ()
   Public-self A person’s aspiration regarding the image they would like others to have of them sirgy1982self ()
Affiliation A person’s aspiration to feel a sense of relationship with others, and a belonging to social groups champoux2016organizational (); kleine1995possession ()
Pleasure A person’s aspiration to feel happiness satisfaction and enjoyment Mugge (); jordan2002designing (); maclachlan2009exploring (); maclachlan2009let ()
   Physical pleasure A person’s aspiration to feel pleasure by means of a sensory perception (five senses) jordan2002designing ()
   Social pleasure A person’s aspiration to feel pleasure resulting from social interaction or social status/image jordan2002designing ()
   Ideological pleasure A person’s aspiration to feel pleasure as a consequence of supporting their values and beliefs jordan2002designing ()
Memories A person’s aspiration to feel a sense of their past, happy moments, people, occasions and places that are important for them klein2012memory (); csikszentmihalyi1981meaning ()
Table 1: Themes and Drivers Definitions
Figure 4: Process Model

3.3 Process Model

As we discussed in Section 1, people’s emotional goals are usually represented at a high level of abstraction and even if people are able to express them, there is limited knowledge about how these emotions can be captured. One of the main purposes of proposing the EAF in this study is developing a technique for helping system analysts to understand the underlying drivers behind people’s emotional goals and breaking people’s emotional goals down to illustrate the key emotional goals of a software system. For this purpose in this section we outline a process model for understanding and capturing emotional goals by using the EAF. As shown in Figure 4, the required input to the process model is data from which the relevant information can be extracted. Data collation can be achieved using any proper qualitative, quantitative or mixed data collection approach such as interviews, surveys, observations, focus groups, ethnographies, documents and records. neuman2005social ().

3.3.1 First Step: Discover Emotional Concerns

The purpose of the first step is to analyse the data and extract the relevant information that is related to people’s emotional goals. As we discussed in Section 1, although system analysts may be able to analyse the data for exploring the emotional goals without using the EAF, we argue that this approach would be at risk of overlooking key emotional concerns. The EAF can provoke system analysts’ minds in analysing the data through its themes and drivers (Table 2). For this purpose, in the first step, the emotional attachment themes are used to analyse data. According to the EAF (Figure 3) the main emotional attachment themes are Self-Expression, Affiliation, Pleasure and Memories. In this step system analysts need to extract Emotional Codes (e.g. quotes) from the data that are related to any themes in the EAF. Table 2 shows some clues to provide prompts for this activity. Similar Emotional Codes are then grouped together to form a list of distinct Emotional Concerns.

3.3.2 Second Step: Derive Emotional Goals

As we discussed in Section 2, emotional concerns can come in both positive and negative forms. Although for systems to be adopted and engaging we use both positive and negative emotions, the negative emotions in this study need to be converted to positive emotions as a preparatory step for goal consolidation. It also is important because we believe it is difficult to design, implement, and track for a lack of concern, compared to a positive concern. For instance ‘having a cloudy mind’ as a negative emotional concern may be converted to ‘good feeling’ or ‘sense of clarity’ as a positive emotional goal. If multiple emotional concerns result in the multiple similar emotional goals then they will be combined and represented by a single encompassing emotional goal. The output of this stage is a list of unique emotional goals.

Themes \Drivers Clue
Self-expression Any evidence regarding a person’s aspiration to have a distinct personal identity from others
    Ideal-Self Any evidence regarding a person’s aspiration to have particular characteristics, skills, etc.,
   Public-Self Any evidence regarding a person’s aspiration to create a particular image from themselves in the eyes of others
Affiliation Any evidence regarding a person’s aspiration for connection, joining, associating, and involving with others
Pleasure Any evidence regarding a person’s aspiration of having pleasure and happiness as the primary or most aim of their life
   Social Pleasure Any evidence regarding a person’s aspiration to have or improve their social interaction and social status
   Ideological Pleasure Any evidence regarding a person’s aspiration to follow their personal values, beliefs and attitudes
   Physical Pleasure Any evidence regarding a person’s aspiration to feel sense of realism in using technology
Memories Any evidence regarding a person’s aspiration for constructing and maintaining a sense of the past, happy moments in life or remind people, occasions, places etc.
Table 2: Themes and Drivers Guideline

3.3.3 Third Step: Develop Emotional Goal Profiles

The final output of the process model is Emotional Goal Profiles (EGP). An EGP contains a summary of an emotional goal, key emotional drivers in EAF that are associated with this emotional goal and the emotional goal priority. Table 4 shows a sample of EGP. The output of the final step was presented in a format that can be used as an input to existing software engineering techniques in order to proceed the derivation process into a final requirements specification. As a user story is a frequently used artifact in software engineering domain for representing high-level requirements paetsch2003requirements (), we adapt this format for our purposes and generate User Emotional Stories (UES). It is our assumption in this study that the UES facilitates sensemaking the abstract emotional goals in a way that software engineers are familiar with, i.e. user story. The general format of UES is based on traditional user stories with a specific the focus and composition including emotional concerns and goals. In this study we use the following format for representing the UES:

As a type of user, I want emotional concern so that I feel emotional goal.

A sample of UES can be found in Table 4. As Table 4 shows the emotional user story explaining the need to have social interaction and relationships as a driver for social pleasure is more useful for a system analyst than simply “designing for Joy”.

Like functional and quality goals, there are many reasons that some emotional goals may be selected over others for analysis and system development, for example (1) no software system can consider all users’ goals in the first iteration, (2) budget and time limits do not let system developers consider all users’ requirements concurrently and, (3) in an incremental system development strategy, system analysts consider the most important requirements at first and then extend the system’s boundary based on the other requirements. There are different techniques for prioritizing users’ goals such as MoSCoW, Paired Comparison and 100-Point hatton2008choosing (). In this research it is our assumption that the frequency of emotional goals is a good representation of the importance of users’ goals. As a result, a simple method for prioritizing each emotional goal is checking how often the related emotional concerns has been reported by stakeholders. It is possible to elicit priorities from people as well, but we believe that the count is a reasonable indicator. For this purpose we calculate the Percentage of Frequency (POF) by using the following:

(1)

In this study we propose a prioritization scales as shown in Table 3. Based on the POF, a higher priority indicates that the related emotional goals have higher importance and priority and system analysts needs to set a higher priority to those emotional goals in designing and developing the software application.

POF (%) Priority
Low
Medium
High
Table 3: Prioritization Scales
Emotional Goal Profile (EGP)
Emotional Concern: Homeless people would like to have social interaction and relationship with others.
Emotional Goal: Connected
Key Emotional Drivers: Ideal Self, Public Self, Affiliation, Social Pleasure
Priority: High
Emotional User Story (EUS): As a homeless person, I want to have social interaction and relationship
with others so that I feel connected.
Table 4: A sample of Emotional Goal Profiles (EGP)

4 Evaluation

The aim of our evaluation is to determine whether the proposed emotional attachment framework and process model are useful and usable in eliciting complete, consistent and correct emotional goals. In the following sections, a summary about the evaluation design, case study and achieved results will be discussed.

4.1 Case Study: Ask Izzy

The pathways to homelessness in Australia are varied and this leads governments and industries to fund a range of services to support people who are homeless or at risk of homelessness. Based on the situation of homeless people, a mobile phone can play a significant role in their life since in crisis, as even without credit, they can call emergency services or use free wi-fi. Survey statistics show that in 2016, 95% of homeless people in Australia have a mobile phone and majority of them (77%) are equipped with smart phones. It is estimated that 42% of the population of homeless or at risk people in Australia are young people (24 years and under) who are more likely to own a smart phone infoxchange ().

In Australia, there are more than 1200 government-funded specialist homelessness services and 300,000 health and welfare community support services infoxchange (). This abundance and diversity of services makes it difficult for those affected by homelessness to find the service they may need. A system that can connect people to the help they need will be beneficial. For this reason, Infoxchange, a not-for-profit social enterprise in Australia, proposed a website application called Ask Izzy222See http://www.askizzy.org.au to connect homeless or people at risk with essential support services and important information. This project is in collaboration with founding partners Google, REA Group and News Corp Australia.

Ask Izzy provided us with the opportunity of evaluating our proposed framework using interview data from people who had a genuine investment in the software application under development. The case study brought a rich representation of a variety of complex emotional concerns drawn from lifetime experiences. Additionally, this case study provided the realistic setting of an industrial case study, along with the realistic pressures and constraints that come with an agile software development process. This rich socio-technical testbed would not have been possible to replicate in any artificial / toy study setup.

4.2 Evaluation Design

To address the evaluation aims, we applied the proposed framework and process model on a case study in which we are involved, which is the design of a website and mobile application called Ask Izzy. As part of this case study, some of the authors of the present paper applied a People-Oriented Software Engineering (POSE) model miller2015emotion () to understand homeless people’s emotional goals in designing the initial version of Ask Izzy. For data gathering they used semi-structured interviews with 22 participants including 15 homeless people and 7 social support workers. Each interview taking between 60-70 minutes, each participant was interviewed once. The interviews were focused on homeless people’s requirements for designing a homelessness social support web application entitled Ask Izzy.

In each interview some general questions were asked. These questions were based on the following themes: 1) what should the technology (web application) do for you?; 2) How should it be?; and 3) How do you want to feel? For those participants with experience in using technology, interviewers also asked questions regarding problems they experienced using technology. The interview data was analysed by using the POSE approach to extract and model the key emotional desires and concerns by different stakeholders. All interviews were recorded and transcribed.

In this study, we analyzed the set of interview data at hand and using following steps based on the proposed process model (Figure 4):

  • Based on the first stage of this process model, we used the EAF to analyze the interview transcripts. The interview data was analyzed to specifically extract emotional codes. The analysis involved careful reading of the interview transcripts by using the themes and drivers guideline (Table 2) for eliciting emotional codes.

  • Based on the second stage of the proposed process model as explained in Section 3.3.2, after grouping emotional codes to emotional concerns, we transformed negative emotional concerns to positive emotional goals.

  • Based on the third stage of the proposed process model as discussed in Section 3.3.3, for each emotional goal, POF was calculated and EGPs were prepared.

For evaluating the aims in this study the following analyses were conducted:

1

Comparison Analysis

In the first round of evaluation, usefulness and usability of the proposed method was measured by comparing the method presented in this paper with what was achieved in a prior emotional goal elicitation approach using the same Ask Izzy data set.

2

3C’s Analysis

In the second round of evaluation, three quality measures, Completeness, Consistency and Correctness (3C’s) zowghi2002three (), were adopted for measuring the usefulness and usability of the proposed method. These three quality measures are described below within the context of our proposed approach:

  • Completeness: The proposed framework and process model generate a complete emotional goals if all emotional goals for addressing people’s emotional concerns have been specified.

  • Consistency: According to the definition of consistency trochim2001research (), the proposed framework and process model generate consistent emotional goals if the same result can be repeatedly derived. In other words if different domain experts derive similar emotional goals, we can infer that the proposed framework and process model lead to consistent results.

  • Correctness: The proposed framework and process model lead to correct emotional goals if it represents the accurate and correct reflection of people’s emotional goals.

Domain Experts Analysis

Domain experts’ judgment is a popular indicator of completeness and correctnessbourque2014guide (); rosqvist2003software (). Accordingly, we conducted evaluations with two independent Ask Izzy domain experts. Both of the domain experts were Ask Izzy development collaborators, exceedingly well versed in the domain of homelessness in Australia and were heavily involved in the elicitation and design of the first version of Ask Izzy. The feedback from two experts that worked directly in design of this application is valuable given the wealth of experiences these individuals posessed. To mitigate the potential bias, the project manager of developing and designing the first version of the Ask Izzy and the expert goal modeling analyst were selected since they were not involved in either developing and using the proposed users emotional attachment framework and its process model. Domain experts worked independently, except for times when they required clarifications. For implementing the domain experts analysis we presented the proposed framework, process model and EGPs to the domain experts and asked them to reflect their thought regarding the completeness and correctness of achieved emotional goals as well as usability and usefulness of the proposed methods333The questionnaire is available at https://goo.gl/p0ioi9.

Inter-rater Reliability Analysis

In extracting the emotional codes in Ask Izzy case study, we used the textual materials (interview transcripts). We also used one author’s judgment for extracting the emotional codes, grouping them and forming the EGPs. Although our main findings in this research were adapted from well-defined emotional attachment theories, there is potential for disagreement on the outputs related to (i) what is considered to be an emotional concern (ii) what driver each concern is related to, and (iii) how these concerns are grouped and labelled as goals. For measuring consistency different techniques such as Inter-Rater Reliability, Test-Retest Reliability, Parallel-Forms Reliability and Internal Consistency Reliability are available trochim2001research (). Among the mentioned techniques inter-rater reliability is a popular technique in the system development domain boudreau2001validation (). As a result, we used Cohen’s Kappa values for measuring the inter-rater reliability to understand to what extent the proposed method in this research leads to consistent results. Cohen’s Kappa statistical measurements range from to ; larger numbers represent better reliability and smaller numbers near zero suggest agreement has happened by chance saeed2013computer ().

As part of the process, after analyzing the transcripts by one of the co-authors, we recruited and trained two other participants to repeat the same activities. To mitigate potential bias, we split inter-rater reliability test into two rounds. As asking the participants to do a complete analysis of the data was unreasonable, in the first round % of quotes were selected randomly and one of the participants (Coder No. 1) analyzed the selected quotes. Coder No. 1 extracted the emotional codes and selected the relevant emotional concerns for each emotional codes from Table 6. In the next step Coder No.1 was asked to group emotional concerns, convert the negative one to positive emotional goal and form the EGP for the captured emotional goals.

In the second round for mitigating the potential bias in selecting quotes, we selected a transcript randomly and asked Coder 1 and Coder 2 to review the first ten pages, 5 middle pages and the last 5 pages. After reviewing the transcripts, coders selected some quotes that they thought associated with emotional concerns. Then we asked coders to select 4 different quotes randomly from each parts and extract emotional codes associated with each quotes based on their understanding. For analyzing the emotional codes and selecting the relevant emotional concerns, the coders followed our process. Then we calculated the Cohen’s Kappa values saeed2013computer () for measuring the proposed method consistency.

4.3 Results

In the following, we first present the results of using the proposed method for understanding and capturing the emotional goals in Ask Izzy. Then an overview of results for each evaluation activity will be presented.

4.3.1 Case Study Analysis Results

Based on what we discussed in Section 3.3.1 and Section 4.2 regarding the data analysis and process model, one of the authors analyzed each interview transcripts by using the EAF themes’ definitions. emotional codes were elicited. Table 5 shows samples of quotes and related emotional codes444The complete list is available at https://goo.gl/vK86i1. As we discussed in Section 3.3.1, based on the similarity between the achieved emotional codes, the research group consolidated similar emotional codes to achieve a list of emotional concerns. In this study the initial emotional codes were grouped into emotional concerns. Table 6 presents a summary of results in the Ask Izzy case study. In this table for each emotional concern we present a short quote (emotional code) from the transcripts to support each emotional concern, however each emotional concern is backed by additional emotional codes. Since all of the emotional concerns in this case study were expressed in negative form, they were subsequently converted to positive emotional goals. As the research group did not find any significant overlap between the achieved positive emotional goals, then the list of unique emotional goals including the emotional goals was prepared. In the next step, for each emotional goal an EGP had been developed555The complete list of EGPs is available at https://goo.gl/j7DxRy.

Quotes Emotional Codes
“It makes you feel bad when they interrogate you and make out like you’re trying to evade the system” Lack of People’s Trust on me, Lack of People’s Respect on me, Lack of People’s Trust me, Lack of People’s Respect for me
“They’d tell you one thing and then you’d see them again a week later and they’d tell you something completely different” Distrust of others, Lack of clarity and stability
“What about the people they need to look after here?” Lack of priority
“You’re in need, you’re in crisis, you’re in everything and you haven’t… there’s nothing for you that you can help.” People’s reluctance to support and provide assistance
“The thing that used to annoy me is they say “Go online, register online and you can access all this stuff”. People’s reluctance for support and assistance
“You have to be lucky of getting the right person in the right call otherwise you could be like turning around.” Lack of clarity, transparency and accuracy
“I think there should be more education around that for the young. I stress now because my son, he’s ten” lack of awareness between the next generations
“There’s just so many things that you learn through other people” Clarity, Transparency, Accuracy
“When people are homeless, they don’t consider being ill is an occupational hazard, and they don’t go to a doctor until they’re really sick, and a lot of times it’s really too late.” Hopeless /loneliness
“You don’t even talk to the other people. They don’t even talk to you.” Emotional and social isolation, Fear of abandonment, Fear of loneliness
“That’s what I’m trying to do is just change the attitude towards it” Losing self esteem, Lack of priority
“There is nothing worse than support worker saying: Well look, I think it’s time for us… you know our three month period is up now, I’ve used all 66 hours with you, you know it’s time to stand on your own feet.” Lack of empathy
“The collapse actually happened?” But you don’t, because you’re shameful, you’re not worthy, you’re not a winner, all those sort of things resonating through you, so there’s that element …” Social shame / losing self confidence
“If you put all the people that live in this country together, we’re all talented in something” Losing self confidence
Table 5: Samples of Emotional Codes - Ask Izzy Case Study
No. Emotional Concerns Emotional Goals Frequency Associated Drivers in EAF

Ideal Self

Public Self

Affiliation

Social Pleasure

Ideological Pleasure

Physical Pleasure

Memories

Frequency of Emotional Codes
1 Feeling of unfairness and injustice behaviour in supporting different types of homeless people Sense of fairness and justice 1 1
2 Feeling like they can’t cope In control 1 1
3 Feeling of dependency to others after becoming homelessness Independent 2 2 2 2
4 Feeling lack of ability to follow my interests and dreams after becoming homelessness Empowered 3 3 2
5 Feeling depressed and confused after becoming homeless Positive 3 3
6 Feeling lack of importance as a homeless people Sense of priority 3 3 3
7 Feeling of distrust and suspiciousness of others Sense of trust in others 3 3 3
8 Feeling unsafe after becoming homelessness Safe and secure 4 4 3 4
9 Feeling useless after becoming homelessness Useful 4 4 4 4 4
10 Feeling lack of self-esteem after becoming homelessness Sense of self-worth 5 3 5 1 5
11 Feeling scary after becoming homelessness Calm 5 5 3
12 Feeling of disappointment and reluctance Motivated and hopeful 5 5 5
13 Feeling lack of privacy after becoming homelessness Sense of privacy 5 1 1 5 4
14 Feeling of lack of self confidence after becoming homelessness Self-confident 5 5 5
15 Feeling people don’t trust on me as a homelessness person Trusted 6 6 6 6
16 Feeling people don’t respect me as a homelessness person Respected 11 11 11 11
17 Feeling like I am not heard as a homelessness person Empathized with and listened to 11 11 11
18 Feeling people are reluctant to help homelessness person Supported and assisted 12 12 1
19 Feeling of emotional and social Isolation, Feeling of loneliness after becoming homelessness Connected 13 13 13 12 13
20 Feeling of misleading and confusion in receiving information Sense of transparency and trust in the information 15 6 15
Sum of of Emotional Codes 117 69 50 36 101 10 0 0
Table 6: A Summary of Results -Ask Izzy Case Study

The fourth column in Table 6 represents the frequency of each emotional codes in the interview data. For example, number in row means that six different emotional codes in the interview data were related to the emotional concern of ‘Feeling people don’t trust me as a homeless people’. A higher frequency in this column indicates that the related emotional codes of selected emotional concern was more common between interviewees, has higher importance and priority for homeless people. As the fourth column in this table shows, between different emotional concerns, ‘feeling misled and confused in receiving information’, ‘feeling of emotional and social isolation, feeling of loneliness after becoming homeless’ and ‘feeling people are reluctant to help homelessness person’ are more common people’s emotional codes with the frequency of , and respectively.

The numbers in the remaining columns represent how many times related emotional codes for each emotional attachment driver have been seen in the interview data. For instance, row in Table 6 shows that interviewees mentioned the concept of this emotional concerns times in different emotional codes. In all cases, the content of mentioned emotional concern was related to ‘Public Self’ and ‘Ideological Pleasure’, three times to ‘Ideal Self’ and just in one case to ‘Social Pleasure’. A higher value indicates that 1) related emotional concern can be addressed more effectively through this driver and, 2) for those prospective people who have expressed this emotional concern, this driver has potential to create emotional attachment between people and software system.

As we discussed in 3.3.1 we identified four main themes in our proposed framework including: 1) Self-expression, 2) Affiliation, 3) Pleasure and 4) Memories that represent a hierarchy of emotional attachment drivers. By consolidating the number of different drivers of each theme, we understand which theme could potentially create more emotional attachment between users and software system. Table 7 shows the number of different emotional attachment drivers of each theme.

Ideal Self

Public Self

Affiliation

Social Pleasure

Ideological Pleasure

Physical Pleasure

Memories

Sum
Self-expression 69 50 119
Affiliation 36 36
Memories 0 0
Pleasure 101 10 0 111
Table 7: Summary of emotional attachment drivers for each theme - EAF approach

From Table 6 we can understand some useful facts that can help us to contextualize the emotional goals in this case study. For instance, according to the distribution of emotional codes in Table 6 we can infer that homeless people feel loss of control in their life when they become homeless. In other words, one of the emotional drivers for the potential users of Ask Izzy will be that the software system capabilities that help users to regain this feeling that they have this ability to control their every day life. As another example, Table 6 shows that the Social Pleasure is the most important driver for forming the emotional attachment between people and software system. In other words, the majority of the people’s emotional concerns can be addressed by considering Social Pleasure in designing the Ask Izzy. Row in Table 6 also shows the complex range of associated drivers for feeling Connected. The complexity of this emotional goal may be an accurate reflection of the complexity of the problems that were expressed by participants. Almost all of the emotional concerns elicited for this goal were found to be associated with Public Self, Affiliation and Social Pleasure. This high degree of overlap found between all three of these drivers supports what have has been argued in Section 3.2 regarding the correlation and internal relationship of these three drivers in forming people emotional attachment. In Figure 5 we summarized the homeless people emotional goals in the AskIzzy case study and associated drivers that influence their emotional attachment to the AskIzzy.

Figure 5: Ask Izzy Emotional Requirements and Attachment Drivers

As Figure 5 and Table 7 indicate, the majority of the homeless people emotional concerns were related to self-expression and pleasure ( and respectively). In other words, the root of majority of the homeless people emotional goals goes back to people’s needs for expressing their personal identity as well as seeking pleasure. The reason of importance for these two categories is not so complicated to understand if we consider the nature of the homelessness phenomenon. As a consequence of becoming homeless, homeless people become so dependent on others and are busy for the activities that they have to do to survive. As a result, they do not have this opportunity to do some activities to feel some degree of pride, self-respect, and self-worth, follow their dreams and interests or have some degree of entertainment. Therefore it is not a surprise if their main emotional concerns are related to Self-expression and Pleasure.

4.3.2 Comparison Analysis Results

As we discussed at the start of Section 4.2, some of the authors of the present paper applied POSE to understand and model homeless people emotional goals. The resulting experiences were included into the design of the initial version of Ask Izzy, which is now a functioning application that has real users, and lots of them.

Reviewing the prior study results shows that emotional codes were found by using POSE666The complete list of emotional codes and results are available at https://goo.gl/oQBdn2. By analyzing the similarity between the emotional codes, we can categorize emotional codes in the previous study into emotional concerns. By reviewing the list of emotional concerns generated using POSE and the proposed method in this paper, we are able to match all emotional concerns in the previous study (POSE) to an equivalent in some of the emotional codes that we achieved by using the EAF.

Ideal Self

Public Self

Affiliation

Social Pleasure

Ideological Pleasure

Physical Pleasure

Memories

Sum
Self-expression 13 7 20
Affiliation 7 7
Memories 0 0
Pleasure 41 4 0 45
Table 8: Summary of emotional attachment drivers for each theme - POSE approach
Count POSE EAF
Emotional goals 12 20
Emotional codes 44 117
Ideal self 13 69
Public self 7 50
Affiliation 7 36
Social pleasure 41 101
Ideological pleasure 4 10
Physical pleasure 0 0
Memories 0 0
Self-expression 20 119
Affiliation 7 36
Memories 0 0
Pleasure 45 111
Table 9: The Comparison of the Results in Using POSE and EAF Approaches

Table 9 shows the comparison between the results generated using POSE versus our EAF. As Table 9 indicates, while all the emotional concerns in the previous study were achieved in this study, by using the proposed method in this study, number of elicited emotional codes were almost doubled (from to ). The additional codes cover something important that the previous did not and we were able to achieve further emotional goals in this study. ‘Sense of fairness and justice’, ‘Independent’, ‘Useful’ and ‘Trusted’ are some samples of emotional goals that were uncovered by only this proposed method, but not in the POSE method.

The only different trend seen is the number of associated emotional concerns with Pleasure. In the previous study the majority of elicited emotional codes are associated with Pleasure () which is two times more than Self-expression (). However the number of elicited emotional concern in this study that have association with Self-expression and Pleasure are almost the same ( and respectively).

4.3.3 3C’s Analysis Results

As we discussed at the start of Section 4, we also investigated three evaluation criteria (3C’s). In the following, a summary of results for each quality measure will be discussed.

  • Completeness:

    To assess completeness, two domain experts were asked to review the results and proposed method and answer a selection of open ended questions regarding the completeness of specified emotional goals. They were also asked to comment on whether the additional emotional goals were valid and useful or not777The questionnaire is available at https://goo.gl/p0ioi9.

    The domain experts answers to the open-ended questions provided some valuable points. One domain experts stated:

    “It is useful to elicit [emotional] concerns. … using the framework is much more comprehensive.”

    “Although you cannot guarantee the completeness, it is a very good starting point for capturing [emotional goals].”

    Although domain experts acknowledged that the captured emotional goals by using the EAF appeared to be complete, they commented that for users who identified as Australian Aboriginal and Torres Strait Islanders, a ‘Sense of Belonging’ was not adequately represented or prioritised in the final goal set. While this goal overlaps with existing goals, such as the goal of feeling ‘Connected’ it was found to be distinct in terms of the underlying drivers. In this case, the ‘Sense of Belonging’ was driven more by a need for cultural identification as opposed to a requirement for social support.

    The analysis of domain experts responses and number of emotional codes and goals that we have achieved by using the EAF show that the proposed framework has the potential to help system analysts to achieve a complete list of emotional goals that need to be considered. It is important to note that people’s emotions change over time and it is therefore difficult to argue that the list of emotional requirements are complete. However, these initial results appear promising given that the domain experts claim that the proposed framework has captured all expected emotional goals that need to be considered.

  • Consistency:

    As we discussed at the start of Section 4.2, for measuring the consistency in this study we used Cohen’s Kappa values. Table 10 shows the statistical values for the first and second rounds. As all the Cohen’s Kappa values in the both rounds are above 70%, which is considered the minimum value for inter-rater agreement to be considered consistent boudreau2001validation (), we can say that the process of extracting emotional codes and forming the EGPs have consistency.

    Round % of Agreement for Emotional Codes % of Agreement for Emotional Drivers
    Co-Author
    Coder 1
    Co-Author
    Coder 1
    1 Coder 1 71.0 % N/A 73.0 N/A
    2 Coder 1 71.9 % N/A 75.9 N/A
    Coder 2 72.1 % 74.0 % 71.3 % 74.1 %
    Table 10: Cohen’s Kappa Values
  • Correctness:

    For measuring to what extent the proposed framework and process model lead to accurate and correct reflection of people’s emotional goals, we again asked the two domain experts to reflect their thoughts regarding the correctness of achieved emotional goals. Results show that both domain experts believe the people’s emotional codes and emotional goals have been elicited correctly.

    For instance, one domain experts stated:

    “All this [the list of emotional goals] is true certainly. They [the captured emotional goals] are all true and valuable and it is very hard to find any counterexample to argue.”

    Although domain experts had some suggestions for better wording of the concerns and goals, by answering the related questions, domain experts approved the correctness of the elicited emotional goals. The domain experts had some comments to the open-ended questions which provided some valuable points. From these results, we see several expected results. First, EAF can help system analysts and requirements engineers, who may not have strong domain expertise, to achieve better insight about the people’s emotional goals.

    “It is useful to identify emotional concerns particularly if someone has no experience. Even if you have experience, when you are designing an interview, it is useful.”

    Another domain expert also stated:

    “The framework completely makes sense to me, …, I think it is really helpful for analysts and designers to put together the questions that they might want to ask.”

    Understanding the people’s emotional goals is so important for system analysts, and our results show that the EAF can help system analysts in this process to some extent. One domain expert mentioned:

    “Having the categories help you focus on relationship between the concerns to consolidate or remove … for consolidating [the] emotional requirements, it is good to look within the proposed categories in the framework.”

    and,

    “Having this [framework] helps me to think in different ways. It could give you a hint and is helpful to create a pool of solution for each [emotional] category. It is good to have them in the background when you designing.”

The results of the above mentioned analyses also support our hypothesis that the EAF can help system analysts to understand and captures people’s emotional goals. We attribute these results to proposed themes and drivers for analyzing the data, hierarchical layout of the emotional attachment drivers in EAF and, well-structured and easy to use process model.

4.4 Limitations

There are some limitations in this study. First, we consider only one case study and two domain experts. This is the trade-off that we made when deciding to ground this evaluation in a realistic industrial case study. Accordingly generalising the usability and usefulness of the proposed framework is limited and further case studies and evaluations with larger numbers would be a logical next step for this research. Additional studies would improve the validity of the proposed framework for eliciting emotional goals. Second, the case study is subject to bias, since our initial understanding from the case study and the homeless people emotional goals would be used for data analysis and comparison the results. Third, the analyzed data in the industry case study in this research focused only around the homeless people and social support workers goals. While data from others stakeholders such governmental sector organizations and business partners may be necessary for further validation.

5 Discussion

In this study, the research question, “How can we get a better understanding of the people’s emotional goals in order to design a software system?” is addressed. Although emotional goals and requirements have received some attention in software engineering domain, to the best of our knowledge, the proposed process model in this study is the first method that helps system analysts to systematically focus on people’s emotional goals. As we experienced it in practice, a priority set of emotional drivers is a good starting point to elicit and capture emotional goals. In doing the industry case study in this research, as we expected, the EAF and the related themes and drivers gave us a good attitude towards the phenomenon of study and the process of understanding the homeless people’s emotional goals. Due to having the EAF, we started the data analysis with a well-developed analysis framework that facilitated the emotional goals elicitation process.

This study suggests that although people may have different emotional goals, the root of their emotional needs can always be found behind what causes people emotionally attach to (or detach from) a software system. However, existing approaches overlook indirect drivers of emotional attachment. There are many social situations that would be overlooked if they are categorized in other frameworks. For instance, categorizing the concern of social isolation or the goal of feeling connected in other frameworks would end up with something quite abstract like “joy” which completely misses the “social” breakdown, and would potentially result in very different requirements. Accordingly, understanding the people’s emotional attachment drivers help system analysts to understand better the people’s emotional goals.

In our study, we captured emotional codes from 130,000 words transcripts that subsequently formed emotional goals (Table 6) under the four main emotional themes: 1) Self-expression; 2) Affiliation; 3) Pleasure; and 4) Memories.

In the AskIzzy case study participants may have been talking about the stigma associated with being homeless. The negative perceptions that society has on homeless people can obviously have a negative impact on the way they feel about themselves. According to the results, homeless people feel loss of control in their life when they become homeless. Therefore they want to regain control their everyday lives. This study also reveals that homeless people generally feel an inability to follow their interest that they had previously. The results indicate that empowering the potential users of AskIzzy in following their interests would be not only important for them as a characteristics that they would like to have, but also it is important for them as part of their social interaction and social status.

In this study, the distribution of emotional goals associated with Affiliation and Social Pleasure drivers indicates that the Clarity and Trust of information is an important goal for potential users and in designing a web application would be need to be considered. The results of this study also reveal that there is a high degree of overlap between Public Self, Affiliation and Social Pleasure drivers. This fact supports the correlation and internal relationship of these three drivers in the EAF in forming emotional attachment. As the results show, the framework guides system analysts to extract subtle references related to the goal of being accepted and Feeling Respected without the need to hear a direct verbal reference to how the situation is related to the way they feel.

Findings from this study shows that utilizing a framework to help capture emotional codes within semi-structured conversations appears to have great potential. However, without such support system analysts are at risk of overlooking important emotional goals. One potential reason for eliciting a greater number of emotional concerns by using the EAF is that it allows the developer to extract quotes that are indirect (as well as direct) references to emotional concerns. A better understanding of different aspects of people’s emotional goals provides valuable insights for system analysts to design successful software application.

Reviewing the results raise the question of whether goals that contain a high number of drivers should be divided into smaller goals, allowing them to become easier to understand and design for. The optimal level of granularity for goals is still unknown, however there are certainly multiple different aspects of these goals that need to be addressed within design. The proposed framework and process model in this study may not capture all emotional goals, and further, the quality of the proposed techniques is only as good as the data that backs them. However, we believe that the list of emotional attachment drivers for each emotional goal directs the system analysts’ attention to the aspects that they need to consider in designing a software system to address the people’s emotional goals. Furthermore, the proposed framework in this study (EAF) can be used before system analysts start the interviews to drive their dialogues and preparing the questionnaire.

Finally, evaluations of our proposed framework in Ask Izzy case study provides support that from domain experts point of view the EAF can lead to improved systems analysts understanding of the people’s emotional goals and ultimately, better software system design. Furthermore, the process model in this study gives this ability to system analysts to prioritize emotional goals based on their frequency and focus on those that have a higher priority. A higher priority indicates that 1) related emotional goals can be addressed more effectively through the specific emotional drivers, and 2) for those people who have this emotional goals, which drivers have more potential to create emotional attachment between people and software system.

6 Conclusion and Future Work

Over the past decade, there has been a paradigm shift from designing software systems for satisfying functional requirements towards the applications that are trying to enhance people’s quality of life. We argue that due to this, system analysts need to engage with emotional relationships between software systems and people. Considering people’s emotional requirements in designing a software system aids in satisfying the basic human physiological needs. As part of this process, understanding and capturing people’s emotional requirements must be integrated into the requirement engineering process as an important aspect of software systems design.

Based on the unstructured nature of emotional requirements, systematic frameworks such as the one presented in this paper, can increase rigor and improve analysis of the domain. It also can stimulate system analysts’ sensitivity to consider the emotional aspects that they might not have considered otherwise. In this paper, we introduced a novel framework and process model for understanding people’s emotional goals, based on the sound theory of emotional attachment. This framework is simple, but conveys the important part of the information of the domain.

The contributions of the EAF and its process model are first to propose an analysis method to deal with people’s emotional goals for complementing existing requirements engineering techniques. In the process of analysing the people’s emotional goals, the EAF draws attention to people’s emotional attachment drivers. The EAF provides a concrete guideline towards emotional requirements engineering however, its accuracy in terms of volatile and changeable nature of people’s emotions is a matter of negotiation. Although the output of process model (EUS) provides input to the requirements engineering process, converting them into functional requirements and design implications require further research.

A hierarchical classification nature of EAF can give strong indications of the emotional goals overlap and clashes. It also help system analysts to consolidate emotional goals with the same root and source for avoiding the confusion in requirement engineering process. The proposed framework (i) encourages system analysts to consider emotional drivers before starting the data gathering such as interviews or workshops and, (ii) it sensitises system analysts’ minds by providing a set of taxonomies and themes to discover emotional goals. Furthermore, the Emotional User Story by using a known structure of user story gives a clear guidance to system analysts to know how to contextualize people’s emotional goals. Although people’s emotions are context based, as we have used the basic concepts of forming the people’s emotion in developing the EAF, the proposed framework can be considered as a generic and context-free framework that can be applied in variety of domains and contexts.

We applied this framework to a software application that aids homeless people finding support services. Our transcribed interviews consisted of more that 130,000 words. We found that the users emotional attachment framework provided a useful and structured way to analyse and collate emotional requirements. The proposed framework in this study helped to find the right level of detail. We believe that our proposed framework is at an appropriate level for understanding the stakeholders’ emotional requirements. Although the quality of what we analysed in the case study is only as good as the data that backs it, we believe that these results support our assumptions that having a view of emotional attachment drivers can facilitate the process of understanding emotional requirements.

In this study we only focused on developing a framework for understanding and capturing people’s emotional goal. In future work, we aim to provide a systematic method for discovering possible design solutions for software systems. We will also use the proposed framework and techniques in other POS domains such as public health and education software systems for further investigation of proposed methods’ usability and usefulness. We encourage other researchers to use the proposed framework in software system domain for developing the proposed framework and as a method for validation and repeatability.

Acknowledgements

The authors would like to take this opportunity to express their gratitude to Alex Lopez-Lorca and Diana Brown for their participation in our domain-expert studies, and valuable feedback on the model. This research is funded by the Australian Research Council Discovery Grant DP160104083 Catering for individuals’ emotions in technology development. The first author is funded by a University of Melbourne MIRS scholarship and a top-up from the CRC for Low-Carbon Living grant Increasing knowledge and motivating collaborative action on Low Carbon Living through team-based and game-based mobile learning.

References

  • (1) Oasig, [cited on-line: accessed 19/12/2015] http://www.it-cortex.com/stat-failure-rate.html the%20oasig%20study%20). 1995.
  • (2) http://www.infoxchange.net.au/homeless-assist-app-qanda, [cited on-line: accessed 12/01/2016]. 2016.
  • (3) https://plato.stanford.edu/entries/hedonism/, [cited on-line: accessed 24/01/2017]. 2017.
  • (4) M. D. S. Ainsworth. Object relations, dependency, and attachment: A theoretical review of the infant-mother relationship. Child development, pages 969–1025, 1969.
  • (5) M. B. Arnold. Perennial problems in the field of emotion. In Feelings and emotions: The Loyola symposium, pages 169–186. Academic Press New York, 1970.
  • (6) A. D. Ball and L. H. Tasaki. The role and measurement of attachment in consumer behavior. Journal of Consumer Psychology, 1(2):155–172, 1992.
  • (7) J. Bates et al. The role of emotion in believable agents. Communications of the ACM, 37(7):122–125, 1994.
  • (8) T. Bentley, L. Johnston, and K. von Baggo. Putting some emotion into requirements engineering. In Proceedings of the 7th Australian workshop on requirements engineering, pages 227–244, 2002.
  • (9) M.-C. Boudreau, D. Gefen, and D. W. Straub. Validation in information systems research: a state-of-the-art assessment. MIS quarterly, pages 1–16, 2001.
  • (10) P. Bourque, R. E. Fairley, et al. Guide to the software engineering body of knowledge (SWEBOK (R)): Version 3.0. IEEE Computer Society Press, 2014.
  • (11) I. Bretherton. The origins of attachment theory: John Bowlby and Mary Ainsworth. Developmental psychology, 28(5):759, 1992.
  • (12) J. Brooks, F.P. No silver bullet essence and accidents of software engineering. Computer, 20(4):10–19, 1987.
  • (13) D. Callele, E. Neufeld, and K. Schneider. Emotional requirements in video games. In Requirements Engineering, 14th IEEE International Conference, pages 299–302. IEEE, 2006.
  • (14) D. Callele, E. Neufeld, and K. Schneider. Balancing security requirements and emotional requirements in video games. In International Requirements Engineering, 2008. RE’08. 16th IEEE, pages 319–320. IEEE, 2008.
  • (15) R. A. Calvo and D. Peters. Positive computing: Technology for wellbeing and human potential. MIT Press, 2014.
  • (16) E. Cambria, A. Livingstone, and A. Hussain. The hourglass of emotions. In Cognitive behavioural systems, pages 144–157. Springer, 2012.
  • (17) J. E. Champoux. Organizational behavior: Integrating individuals, groups, and organizations. Routledge, 2016.
  • (18) J. Chapman. Emotionally durable design: objects, experiences and empathy. Routledge, 2015.
  • (19) T. Clancy. The standish group report. 1995.
  • (20) G. Cockton, D. Kirk, A. Sellen, and R. Banks. Evolving and augmenting worth mapping for family archives. In Proceedings of the 23rd British HCI Group Annual Conference on People and Computers: Celebrating People and Technology, pages 329–338. British Computer Society, 2009.
  • (21) R. Colomo-Palacios, C. Casado-Lumbreras, P. Soto-Acosta, and Á. García-Crespo. Using the affect grid to measure emotions in software requirements engineering. Journal of Universal Computer Science, 17(9):1281–1298, 2011.
  • (22) R. Colomo-Palacios, A. Hernández-López, Á. García-Crespo, and P. Soto-Acosta. A study of emotions in requirements engineering. In Organizational, Business, and Technological Aspects of the Knowledge Society, pages 1–7. Springer, 2010.
  • (23) R. Cowie, E. Douglas-Cowie, N. Tsapatsoulis, G. Votsis, S. Kollias, W. Fellenz, and J. G. Taylor. Emotion recognition in human-computer interaction. IEEE Signal processing magazine, 18(1):32–80, 2001.
  • (24) M. Csikszentmihalyi and E. Halton. The meaning of things: Domestic symbols and the self. Cambridge University Press, 1981.
  • (25) R. De Sousa. The rationality of emotion. MIT Press, 1990.
  • (26) P. Desmet. Designing Emotions. Pieter Desmet, 2002.
  • (27) P. Desmet and P. Hekkert. Framework of product experience. International journal of design, 1(1), 2007.
  • (28) O. Dieste, N. Juristo, and F. Shull. Understanding the customer: what do we know about requirements elicitation? Software, IEEE, 25(2):11–13, 2008.
  • (29) A. Dix, J. E. Finlay, G. D. Abowd, and R. Beale. Human-Computer Interaction (3rd Edition). Prentice-Hall, Inc., 2003.
  • (30) S. W. Draper. Analysing fun as a candidate software requirement. Personal Technologies, 3(3):117–122, 1999.
  • (31) S. Y. Eric. Social modeling and i*. In Conceptual Modeling: Foundations and Applications, pages 99–121. Springer, 2009.
  • (32) S. Y. Eric. Introduction to social modelling in re. In Y. Eric, G. Paolo, M. Neil, and J. Mylopoulos, editors, Social modelling for requirements engineering. The MIT Press, 2010.
  • (33) B. Friedman. Value-sensitive design. interactions, 3(6):16–23, 1996.
  • (34) B. Friedman, P. Kahn, and A. Borning. Value sensitive design: Theory and methods. University of Washington technical report, pages 02–12, 2002.
  • (35) R. Fuentes-Fernández, J. J. Gómez-Sanz, and J. Pavón. Understanding the human context in requirements elicitation. Requirements engineering, 15(3):267–283, 2010.
  • (36) M. V. Giuliani. Theory of attachment and place attachment. na, 2003.
  • (37) W. D. Glanze, K. Anderson, L. E. Anderson, et al. Mosby’s medical, nursing, and allied health dictionary. Mosby, 1990.
  • (38) J. A. Goguen. Requirements engineering. chapter Requirements Engineering As the Reconciliation of Social and Technical Issues, pages 165–199. Academic Press Professional, Inc., San Diego, CA, USA, 1994.
  • (39) J. A. Gogueny. Requirements engineering as the reconciliation of technical and social issues. Requirements Engineering: Social and Technical Issues, edited with Marina Jirotka, Academic Press, pages 165–199, 1994.
  • (40) C. K. Gonzales and G. Leroy. Eliciting user requirements using appreciative inquiry. Empirical Software Engineering, 16(6):733–772, 2011.
  • (41) J. Gordijn and J. Akkermans. Value-based requirements engineering: exploring innovative e-commerce ideas. Requirements engineering, 8(2):114–134, 2003.
  • (42) P. C. Govers and J. P. Schoormans. Product personality and its influence on consumer preference. Journal of Consumer Marketing, 22(4):189–197, 2005.
  • (43) P. Guinan and R. P. Bostrom. Development of computer-based information systems: A communication framework. ACM SIGMIS Database, 17(3):3–16, 1986.
  • (44) F. Guo, W. L. Liu, F. T. Liu, H. Wang, and T. B. Wang. Emotional design method of product presented in multi-dimensional variables based on kansei engineering. Journal of Engineering Design, 25(4-6):194–212, 2014.
  • (45) M. Hassenzahl, A. Beu, and M. Burmester. Engineering joy. Ieee Software, 18(1):70, 2001.
  • (46) S. Hatton. Choosing the right prioritisation method. In Software Engineering, 2008. ASWEC 2008. 19th Australian Conference on, pages 517–526. IEEE, 2008.
  • (47) J. Hughes, J. O’Brien, T. Rodden, M. Rouncefield, and I. Sommerville. Presenting ethnography in the requirements process. In Requirements Engineering, 1995., Proceedings of the Second IEEE International Symposium on, pages 27–34. IEEE, 1995.
  • (48) J. Ingram. Designing the spatial experience. Design Studies, 5(1):15–20, 1984.
  • (49) P. W. Jordan. Designing pleasurable products: An introduction to the new human factors. CRC press, 2002.
  • (50) S. B. Klein and S. Nichols. Memory and the sense of personal identity. Mind, pages 677–702, 2012.
  • (51) S. S. Kleine and S. M. Baker. An integrative review of material possession attachment. Academy of marketing science review, 2004:1, 2004.
  • (52) S. S. Kleinef, R. E. Kleine III, and C. T. Allen. How is a possession” me” or” not me”? characterizing types and an antecedent of material possession attachment. Journal of Consumer Research, pages 327–343, 1995.
  • (53) M. Komssi, M. Kauppinen, H. Töhönen, L. Lehtola, and A. M. Davis. Integrating analysis of customers’ processes into roadmapping: The value-creation perspective. In Requirements Engineering Conference (RE), 2011 19th IEEE International, pages 57–66. IEEE, 2011.
  • (54) M. a. Krumbholz, J. Galliers, N. Coulianos, and N. Maiden. Implementing enterprise resource planning packages in different corporate and national cultures. Journal of Information Technology, 15(4):267–279, 2000.
  • (55) M. Lewis, J. M. Haviland-Jones, and L. F. Barrett. Handbook of emotions. In Handbook of Emotions, pages 68–88. Guilford Press, 2008.
  • (56) A. A. Lopez-Lorca, T. Miller, S. Pedell, L. Sterling, and M. K. Curumsing. Modelling emotional requirements, 2014.
  • (57) M. Maclachlan, D. Harrison, B. Wood, et al. Exploring the reflective and utilitarian benefits of product attachment. In DS 58-10: Proceedings of ICED 09, the 17th International Conference on Engineering Design, Vol. 10, Design Education and Lifelong Learning, Palo Alto, CA, USA, 24.-27.08., 2009.
  • (58) M. Maclachlan, D. Harrison, B. Wood, et al. Let’s get emotional: introducing undergraduate product design students to the concept of emotional design. In DS 59: Proceedings of E&PDE 2009, the 11th Engineering and Product Design Education Conference-Creating a Better World, Brighton, UK, 10.-11.09., 2009.
  • (59) A. Marcus. The emotion commotion. In HCI and User-Experience Design, pages 83–89. Springer, 2015.
  • (60) M. Mende and R. N. Bolton. Why attachment security matters how customers’ attachment styles influence their relationships with service firms and service employees. Journal of Service Research, 14(3):285–301, 2011.
  • (61) A. Mendoza, T. Miller, S. Pedell, L. Sterling, et al. The role of users’ emotions and associated quality goals on appropriation of systems: two case studies. In 24th Australasian Conference on Information Systems. Citeseer, 2013.
  • (62) A. Meschtscherjakov, D. Wilfinger, and M. Tscheligi. Mobile attachment causes and consequences for emotional bonding with mobile phones. In Proceedings of the 32nd annual ACM conference on Human factors in computing systems, pages 2317–2326. ACM, 2014.
  • (63) D. G. Mick and M. DeMoss. Self-gifts: Phenomenological insights from four contexts. Journal of Consumer Research, 17(3):322–332, 1990.
  • (64) T. Miller, S. Pedell, A. A. Lopez-Lorca, A. Mendoza, L. Sterling, and A. Keirnan. Emotion-led modelling for people-oriented requirements engineering: the case study of emergency systems. Journal of Systems and Software, 105:54–71, 2015.
  • (65) T. Miller, S. Pedell, L. Sterling, F. Vetere, and S. Howard. Understanding socially oriented roles and goals through motivational modelling. Journal of Systems and Software, 85(9):2160–2170, 2012.
  • (66) L. M. Monti Fonseca, T. Min Lun, D. M. Vilela Dias, C. G. Silvan Scochi, A. M. Fernandes, J. C. Amado Martins, and M. Alves Rodrigues. Emotional design and its contributions to digital educational technology in health and nursing: integrative review. Revista de Enfermagem Referencia, 4(6):141 – 149, 2015.
  • (67) R. Mugge. Product attachment. PhD thesis, Delft University of Technology, 2007.
  • (68) R. Mugge, P. C. Govers, and J. P. Schoormans. The development and testing of a product personality scale. Design Studies, 30(3):287–302, 2009.
  • (69) R. Mugge, H. N. Schifferstein, and J. P. Schoormans. Product attachment and satisfaction: understanding consumers’ post-purchase behavior. Journal of consumer Marketing, 27(3):271–282, 2010.
  • (70) R. Mugge, J. P. Schoormans, and H. N. Schifferstein. Product attachment-17: Design strategies to stimulate the emotional bonding to products. 2008.
  • (71) R. Mugge, J. P. Schoormans, and H. N. Schifferstein. Emotional bonding with personalised products. Journal of Engineering Design, 20(5):467–476, 2009.
  • (72) E. Mumford. Values, technology and work, volume 3. Springer Science & Business Media, 2013.
  • (73) W. L. Neuman. Social research methods: Quantitative and qualitative approaches, volume 13. Allyn and Bacon Boston, 2005.
  • (74) D. A. Norman. Emotional design: Why we love (or hate) everyday things. Basic books, 2005.
  • (75) D. A. Norman. The design of everyday things: Revised and expanded edition. Basic books, 2013.
  • (76) A. Ortony, G. L. Clore, and A. Collins. The cognitive structure of emotions. Cambridge university press, 1990.
  • (77) F. Paetsch, A. Eberlein, and F. Maurer. Requirements engineering and agile software development. In Enabling Technologies: Infrastructure for Collaborative Enterprises, 2003. WET ICE 2003. Proceedings. Twelfth IEEE International Workshops on, pages 308–313. IEEE, 2003.
  • (78) T. Page. Product attachment and replacement: implications for sustainable design. International Journal of Sustainable Design, 2(3):265–282, 2014.
  • (79) W. G. Parrott. Emotions in social psychology: Essential readings. Psychology Press, 2001.
  • (80) D. S. Platt. Why Software Sucks–and what You Can Do about it. Addison-Wesley Professional, 2007.
  • (81) R. Plutchik. Emotions and life: Perspectives from psychology, biology, and evolution. American Psychological Association, 2003.
  • (82) R. Proynova, B. Paech, S. H. Koch, A. Wicht, and T. Wetter. Investigating the influence of personal values on requirements for health care information systems. In Proceedings of the 3rd Workshop on Software Engineering in Health Care, pages 48–55. ACM, 2011.
  • (83) I. Ramos and D. M. Berry. Is emotion relevant to requirements engineering? Requirements Engineering, 10(3):238–242, 2005.
  • (84) T. Rosqvist, M. Koskela, and H. Harju. Software quality evaluation based on expert judgement. Software Quality Journal, 11(1):39–55, 2003.
  • (85) L. M. Rulla. Depth psychology and vocation: A psycho-social perspective. Gregorian Biblical BookShop, 2003.
  • (86) P. Sable. Pets, attachment, and well-being across the life cycle. Social work, 40(3):334–341, 1995.
  • (87) K. Saeed, R. Chaki, A. Cortesi, and S. Wierzchoń. Computer Information Systems and Industrial Management: 12th IFIP TC 8 International Conference, CISIM 2013, Krakow, Poland, September 25-27, 2013, Proceedings, volume 8104. Springer, 2013.
  • (88) K. Salen and E. Zimmerman. Rules of play: Game design fundamentals. MIT press, 2004.
  • (89) H. Salzer. Atrs (atomic requirements) used throughout development lifecycle. In 12th International Software Quality Week, 1999.
  • (90) L. Scannell and R. Gifford. Defining place attachment: A tripartite organizing framework. Journal of environmental Psychology, 30(1):1–10, 2010.
  • (91) K. R. Scherer, A. Schorr, and T. Johnstone. Appraisal processes in emotion: Theory, methods, research. Oxford University Press, 2001.
  • (92) H. N. Schifferstein, R. Mugge, and P. Hekkert. Designing consumer-product attachment. In Design and Emotion, pages 378–83. CRC Press, 2004.
  • (93) A. M. Schmitter. 17th and 18th century theories of emotions. In E. N. Zalta, editor, The Stanford Encyclopedia of Philosophy. 2014.
  • (94) M. J. Sirgy. Self-concept in consumer behavior: A critical review. Journal of consumer research, pages 287–300, 1982.
  • (95) J. S. Slater. Collecting brand loyalty: A comparative analysis of how coca-cola and hallmark use collecting behavior to enhance brand loyalty. NA-Advances in Consumer Research Volume 28, 2001.
  • (96) H. Smith and A. Schneider. Critiquing models of emotions. Sociological Methods & Research, 37(4):560–589, 2009.
  • (97) I. Sommerville, T. Rodden, P. Sawyer, R. Bentley, and M. Twidale. Integrating ethnography into the requirements engineering process. In Requirements Engineering, 1993., Proceedings of IEEE International Symposium on, pages 165–173. IEEE, 1993.
  • (98) A. Sutcliffe and S. Thew. Analysing people problems in requirements engineering. In 2010 ACM/IEEE 32nd International Conference on Software Engineering, volume 2, pages 469–470. IEEE, 2010.
  • (99) F. Sweet. Frog: form follows emotion. Thames & Hudson London, 1999.
  • (100) S. Thew and A. Sutcliffe. Investigating the role of’soft issues’ in the re process. In International Requirements Engineering, 2008. RE’08. 16th IEEE, pages 63–66. IEEE, 2008.
  • (101) S. Thew and A. Sutcliffe. Value-based requirements engineering: method and experience. Requirements Engineering, pages 1–22, 2017.
  • (102) P. A. Thoits. The sociology of emotions. Annual review of sociology, pages 317–342, 1989.
  • (103) M. Thomson, D. J. MacInnis, and C. W. Park. The ties that bind: Measuring the strength of consumers’ emotional attachments to brands. Journal of consumer psychology, 15(1):77–91, 2005.
  • (104) L. Tichy and T. Bascom. The business end of it project failure. Mortgage Banking, 68(6):28, 2008.
  • (105) W. M. Trochim and J. P. Donnelly. Research methods knowledge base. 2001.
  • (106) M. Van Harmelen. Interactive system design using OO & HCI methods. Addison-Wesley, 2001.
  • (107) S. Viller and I. Sommerville. Social analysis in the requirements engineering process: from ethnography to method. In Requirements Engineering, 1999. Proceedings. IEEE International Symposium on, pages 6–13. IEEE, 1999.
  • (108) C. Whan Park, D. J. MacInnis, J. Priester, A. B. Eisingerich, and D. Iacobucci. Brand attachment and brand attitude strength: Conceptual and empirical differentiation of two critical brand equity drivers. Journal of marketing, 74(6):1–17, 2010.
  • (109) B. Whittaker. What went wrong? unsuccessful information technology projects. Information Management & Computer Security, 7(1):23–30, 1999.
  • (110) T. Winfree, P. Goldacre, M. Sherkat, P. Graham, A. Mendoza, and T. Miller. Learning for low carbon living: The potential of mobile learning applications for built environment trades and professionals in australia. Procedia Engineering, 180:1773–1783, 2017.
  • (111) D. Zowghi and V. Gervasi. The three cs of requirements: consistency, completeness, and correctness. In International Workshop on Requirements Engineering: Foundations for Software Quality, Essen, Germany: Essener Informatik Beitiage, pages 155–164, 2002.
Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
""
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
   
Add comment
Cancel
Loading ...
238727
This is a comment super asjknd jkasnjk adsnkj
Upvote
Downvote
""
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters
Submit
Cancel

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test
Test description