Human-Centered Software Engineering

Redesign proposals have been suggested as means to improve the feedback from usability evaluation to software development. Yet redesign proposals are usually created by usability specialists without any involvement of the software developers who will implement the proposals. This paper reports from an exploratory study where redesign proposals were created in an active and collaborative learning process that involved both software developers and usability specialists. The focus was on the support that the developers needed in order to contribute constructively to improve the usability of the system. The findings show that this process had a considerable impact on the developers’ understanding of the usability problems, especially the weaknesses of the system. They were able to contribute constructively to create redesign proposals, and they found the workshop very useful for their future efforts to eliminate the usability problems that have been identified.

A fair amount of literature has been published concerning the gaps between HCI and software engineering. However, most of it tends to look at the effects of these gaps rather than their causes. We argue that the use of autoethnographic methods would help us in identifying the root causes of these gaps and can bring us closer to finding potential solutions. In this paper, we focus on issues associated with effort estimation for HCI activities in three projects in three typical engagement models for outsourced software development projects in a mainstream IT company in India, namely Fixed Price model, Mixed model, and Time & Material model. We found that the HCI practitioner needs to negotiate her position with several members of the team, both within the vendor and client organisations. At times, a foot-in-the-door project turns out to be a foot-in-the-mouth project. At other times, it leads to inefficiencies and imbalance of work load. The autoethnographic approach led to reflexive thinking by the HCI practitioner, helping her to develop a deeper understanding of all aspects of a problem, and bringing her closer to potential solutions in some cases. The paper also brings to light several aspects of autoethnography as a method, which can influence effort estimation of HCI activities for future projects.

With respect to modeling the context of interaction, two different research communities consider the context from different viewpoints. The user-centered view which prevails in the HCI and the device-centered view which is dominant in the mobile and ubiquitous computing. Despite existing advances, context modeling and user interface (UI) design methods are still poorly integrated, making it difficult to use the contextual elements directly in UI design. This paper focuses on bringing user-related aspects of the interaction context in UI design. We propose a model-driven framework for the development of adaptive user interfaces. The framework describes the interaction context by integrating contextual factors from different context perspectives in a unison manner. Then it provides formal semantic relations between contextual and UI elements. The framework has been used in the data visualization domain, particularly in the design of the software instrument table for UAV (Unmanned Aerial Vehicle) that takes into account user context, namely human perceptual abilities.

Agile development has positive attitudes towards continuously improving work practices of IT professionals and the quality of the software. This study focuses on value adding activities such as user involvement and gathering metrics and non-value adding activities, such as correcting defects. Interviews were conducted with 10 IT professionals working with agile development in Iceland. Results show that IT professionals emphasise communication with users both through direct contact and using email, but they rarely use metrics to make improvements measurable. The most serious non-value adding activities are: partially done work, delays and defects. The core reason is that long lists of defects in the projects exist, which means that the software is partially done and the defects cause delays in the process. There are efforts to reduce non-value adding activities in the process, but IT professionals are still confronted with problems attributed to miscommunication and the impediments by the external environment.

Integrating practices and methods of Interaction Design and Usability into Software Engineering processes has posed some challenges. In this paper we extend a SE process to enable its instantiation as user centered in order to improve the usability level reachable by the final system. Also, we suggest a kind of a road map that enables software organizations to instantiate cumulative versions of this process to grow in their capability regarding the usability practices. The paper is organized in two parts. First, we describe de open source version of the Unified Process (OpenUP) and the ISO Usability Maturity Model (UMM-ISO) and present the results of an assessment made on the first to determine its conformity with the latter. In the second part we present an extension to OpenUP to fill the gaps discovered and report the highlights of an implementation of these contributions in a real project and the lessons learned.

The aim of this paper is to explore and evaluate the role that can be played by Usability Maturity Models (UMMs) in integrating agile processes and User Centred Design (UCD). UMMs can be utilised in Agile projects as a diagnostic tool to assess the extent to which UCD is systematically and consistently implemented and the extent of effective implementation of UCD in development projects. This paper investigated the suitability of Nielsen Corporate Usability Maturity Model for utilisation in the Agile domain in order to assess the organisation’s UCD capability. It reported on applying Nielsen Model in five case studies that performed Agile and User Centred Design Integration (AUCDI) and utilising the model in assessing their usability maturity level. The results revealed the existence of a correlation between the success of AUCDI attempts and the AUCDI case study’s usability maturity level. These results can have positive implications on AUCDI practice since practitioners who aim to achieve the integration can utilise Nielsen model to identify their strengths and weaknesses in regards to UCD related aspects and accordingly plan for improvement.

Whilst task models are perceived as critical artifacts within User Centered Design methods, task models development is often considered as a resource and time consuming activity. Structuring techniques can support handling issues such as reuse and scalability and can improve analysts’ productivity and the overall quality of models. In this paper we propose (based on the notation of the HAMSTERS project) several means to structure task models and present how they can be used in order to increase reusability and scalability in task models. Besides sub-models and sub-routines, generic components are described. These mechanisms are duly illustrated within a project for the ground segments of satellite missions. This paper shows, by example, how such elements look like and how both readability and quality of models is improved by their use.

In today’s rapidly developing Internet, the web sites and services end users see are more and more composed of multiple services, originating from many different providers in a dynamic way. This means that it can be difficult for the user to single out individual web services or service providers and consequently judge them regarding how much they trust them. So the question is how to communicate indicators of trustworthiness and provide adequate security feedback to the user in such a situation. Contemporary literature on trust design and security feedback is mostly focused on static web services and, therefore, only partially applicable to dynamic composite web services. We conducted two consecutive studies (a qualitative and a quantitative one) to answer the questions of how and when security feedback in dynamic web service environments should be provided and how it influences the user’s trust in the system. The findings from the studies were then analyzed with regards to Riegelsberger and Sasse’s ten principles for trust design [24]. The outcome we present in this paper is an adapted list of trust principles for dynamic systems.

We are using our mobile devices in an increasing number of dynamic contexts, thus we need more and more context-dependent applications. However, only end users can know the most appropriate ways their applications should react to contextual events. In order to facilitate end user development of context-dependent applications in smartphones a first generation of mobile environments has been proposed in the market. In this work we analyse three such Android applications in terms of their ability to express the relevant concepts and their usability, also through a user study. We indicate some limitations of the current solutions and provide indications that can support future work for providing more effective results.

It has been argued that too much research on usability engineering is incoherent with the processes, and settings being the realities for practitioners. In this paper we want to extend the existing knowledge about usability engineering in the wild. Through 12 semi-structured interviews we wanted to get an understanding of how usability is perceived, and practiced in reality. We found that our participants primarily focus on upfront work to support the design, and implementation process. They implement usability engineering through informal evaluations, and by following a set of local de facto standards. We want to extend the existing body of knowledge about usability engineering in practice, to support the development of methods aimed at practitioners.

With an increasing number of supported devices, usability evaluation of websites becomes a laborious task. Therefore, usability evaluation should be automated as far as possible. In this paper, we present a summative method for automated usability evaluation of websites. The approach records user actions and transforms them into task trees. The task trees are then checked for usability smells to identify potential usability issues. The approach was applied in two case studies and shows promising results in the identification of four types of usability smells.

Common characteristics of older adults have led to the rise of recommendations to conduct user research and testing with this particular age group. Even though guidelines exist regarding human-centred design with older adults, there are not many reports on experiences with creating and maintaining elderly user groups for design and research purposes. This paper reviews previous reports in the literature about user groups and adds i) results of qualitative research about the experience of researchers who built their entire professional experience with user groups of older adults, along with ii) the authors’ own account of building and maintaining an elderly user group for years. The paper provides recommendations for recruiting, maintaining and motivating elderly user groups towards participation in HCI design and research activities, along with suggestions of strategies to use during field work.

Current user interface modeling languages usually focus on modeling a single user interface and have a fixed set of user interface components; adding another user interface component requires an extension of the language.

In this paper we present CASSIS, a concise language that supports creation of user interface components using models instead of language extensions. It also allows the specification of design-time and runtime user interface variations. The support for variations has been used to generate constraints for custom user interface components, to specify design patterns and design decisions. CASSIS has been used in several projects including a multi-disciplinary applied research project.

Personas is a technique that supports designing and engineering interactive systems with the focus on the end-users. This paper reports from a case study, where we interviewed four software developers about their usage of personas in software development practice. The purpose of was to identify the practices of personas development in the software development industry. How the respondents perceive personas and its use does not always correlate with what is described as best practice in the literature. We found that practitioners are not using personas as stated in the literature but are developing their own practices both in regards to when and how personas are created.

This paper presents a three-year follow-up study considering the improvement process of user experience work in a software company utilizing Scrum. Problems encountered in the organization included managing the product vision, timing of UX, lack of cooperation among disciplines, and understanding user needs. We also observed changes in the organization over two years. They included ceasing the centralized UX team and dividing the UX specialists over business lines. UX specialists were given influential roles in regard to product decisions – such as nominated as product owners.

Self-Service Systems are technically complex and provide products and services to end users. Due to the heterogeneity of the users of such systems and their short residence time, the usability of a system’s user interface is of great importance. Currently, an intuitive and flexible usage is often limited because of the monolithic system architecture of existing Self-Service Systems. Furthermore, today’s Self-Service Systems represent the one-and-only endpoint of communication with a customer when processing a transaction. The integration of the customer’s personal computing devices, like desktop PC, notebook, and smartphone is not sufficiently covered yet. In order to tackle these problems, we have established a methodology for developing adaptive UIs for Multi-Channel Self-Services where a customer may, for example, start a transaction on a PC at home, modify it with the smartphone, and finally finish it at a Self-Service terminal. In this paper we describe our integrated model-based approach for the development of adaptive user interfaces for distributed Multi-Channel Self-Service Systems.

Recommendation is a popular feature of social software. Recommendations could be made by the software autonomously or by social contacts who are often aided by the software on what to recommend. A great deal of emphasis in the literature has been given to the algorithmic solution to infer relevant and interesting recommendations. Yet, the delivery method of recommendation is still a widely unexplored research topic. This paper advocates that the success in deducing recommendations is not the sole factor for “recommendees” to consider. Users have their own requirements on the way a recommendation is made and delivered. Failure in meeting user expectations would often lead to the rejection of the recommendations as well as the violation of user experience. In this paper, we conduct an empirical research to explore such user’s perspective. We start with qualitative phase, based on interviews, and confirm and enhance the results in a quantitative phase through surveying a large sample of users. We report on the results and conclude with a set of guidelines on how recommendations delivery should be designed from a user’s perspective.

Human centricity refers to the active involvement in the overall product lifecycle of different human actors including end-users, stakeholders and providers. Persona is one of the different tools that exist for human centricity. While marketing is the original domain in which persona was introduced, this technique has also been widely used in user-centered design (UCD) design. In these two perceptions, persona has demonstrated its potential as an efficient tool for grouping the users or customers and focusing on user or customer needs, goals and behavior. A segmentation technique is generally used with persona in order to group individual users according to their common features, identifying within these groups those that represent a pattern of human behavior. This paper investigates how persona has been used to improve the usability in the agile development domain, while studying which contributions from marketing and HCI have enriched persona in this agile context.

Designing with the community brings about a number of benefits, including tacit and contextual knowledge about the problem domain; this is especially apparent in rural settings. However, designing for problems that have yet to embed themselves in the fabric of society (i.e. future problems) poses a number of challenges, as they typically present intangible scenarios and concepts that have yet to be experienced by the wider-community. Using the OnSupply project as a case study, we share our experience in working with the Tiree community to address a future problem through a technology-mediated enquiry. Furthermore, we present a novel process that uses creative workshops augmented with physical artefacts to inform and learn from the community about a problem space, and to seed the design of a system that addresses it.

With the introduction and adoption of tabletop technologies a need for different user authentication mechanisms has arisen. Tabletops support close collaboration between users, typically in close physical proximity and such settings are more vulnerable to shoulder surfing attacks compared to desktops settings where users are more distantly located. Previous studies on desktop interfaces have shown that multi factor authentication provides a higher level of security than single factor authentication. This study extends previous work by comparing the usability of several authentication methods applied in tabletop settings. The aim of the study is to contribute with proposals on which authentication methods to apply when engineering user interfaces for tabletop devices. We compare single factor and multi-factor authentication mechanisms from a usability perspective.

User participation is a key element in user centered design of interactive systems. However, applying established methods is not straightforward while realizing a system for people with cognitive impairment due to their specific, heterogeneous needs and abilities. This paper presents experiences from conducting a user centered inclusive design process within an ongoing project. It aims at the development of a web application for people with dyslexia caused by cognitive impairment. A distinctive feature is the intended use of the application: It must support the counseling interviews for planning the real life inclusion of the target group. In addition, it must enable cognitively impaired as well as non-impaired people using it cooperatively. One of the most challenging issues was how and when to involve cognitively impaired users.

This paper summarizes the keynote address on the future of end-user software engineering. We believe the future that we envision has implications for not only end-user software engineering, but also for “classic” software engineering.

Software forms an integral part of the most complex artifacts built by humans. Communication, production, distribution, healthcare, transportation, banking, education, entertainment, government, and trade all increasingly rely on systems driven by software. Such systems may be used in ways not anticipated at design time as the context in which they operate is constantly changing and humans may interact with them an unpredictable manner. However, at the same time, we are able to collect unprecedented collections of event data describing what people and organizations are actually doing. Recent developments in process mining make it possible to analyze such event data, thereby focusing on behavior rather than correlations and simplistic performance indicators. For example, event logs can be used to automatically learn end-to-end process models. Next to the automated discovery of the real underlying process, there are process mining techniques to analyze bottlenecks, to uncover hidden inefficiencies, to check compliance, to explain deviations, to predict performance, and to guide users towards “better” processes. Process mining reveals how people really work and often reveals what they would really like to do. Event-based analysis may reveal workarounds and remarkable differences between people and organizations. This keynote paper highlights current research on comparative process mining. One can compare event data with normative process models and see where people deviate. Some of these deviations may be positive and one can learn from them. Other deviations may reveal inefficiencies, design flaws, or even fraudulent behavior. One can also use process cubes to compare different systems or groups of people. Through slicing, dicing, rolling-up, and drilling-down we can view event data from different angles and produce process mining results that can be compared.