ReviewEssays.com - Term Papers, Book Reports, Research Papers and College Essays
Search

Total Control of User Experience in Software Development - a Software Engineering Dream?

Essay by   •  February 4, 2011  •  Research Paper  •  3,209 Words (13 Pages)  •  2,405 Views

Essay Preview: Total Control of User Experience in Software Development - a Software Engineering Dream?

Report this essay
Page 1 of 13

Total control of User Experience in Software Development - a Software Engineering dream?

Joerg Doerr

Fraunhofer IESE

Fraunhofer Platz 1

Joerg.Doerr@iese.fraunhofer.de Daniel Kerkow

Fraunhofer IESE

Fraunhofer Platz 1

Daniel.Kerkow@iese.fraunhofer.de

ABSTRACT

In this paper we present a preliminary version of a software engineering approach to gain control over the User Experience (UX) during development time. We show results of an exploratory study with 59 subjects, discovering correlations between quality attributes as introduced in ISO9126 and our construct of UX, which is derived from the construct of Quality in Use as described in ISO9126.

Author Keywords

Position paper, User Experience; ISO9126/25000; Software Quality; Nonfunctional requirements, Quality in Use

ACM Classification Keywords

H5.m. Information interfaces and presentation (e.g., HCI): Miscellaneous.

INTRODUCTION

In this paper we present a preliminary version of a software engineering approach to gain control over the User Experience (UX) during development time. We define the construct of UX in terms of a quality goal of a product, which manifests itself during the usage of the product. In ISO9126, as well as in the newer version the ISO 25000, this kind of quality is referred to as Quality in Use (QiU). We claim UX to be closely related to User Perceived QiU.

Software engineering aims to make the quality of software systems predictable in early phases of development and repeatable during the usage of these systems. Since we consider UX as a set of quality characteristics, these characteristics should be on the one hand measurable and on the other hand controllable during development time. ISO9126 [5] and [8] state that software quality can be measured and described via 1. Static measures of the software (internal quality), 2. behavior of code when executed (external quality) and 3. in use (quality in use). Basically, QiU- characteristics can not directly be manipulated during development; developers have to measure and manipulate them indirectly via the internal and external quality (I&EQ) characteristics of the objects available during development. I&EQ and QiU or UX in software and system development are heavily intertwined. We experience that even though standards like ISO 9126 seem to involve both views in one standard, there is no simple relationship between these two aspects. UX can only be constructed into software and systems, if basic non-functional requirements (NFRs) that are expressed over the I&EQ attributes are measurably specified. The relationship between NFRs and UX, or in other words, the means to reach a desired UX by implementing NFRs remains unknown.

In Section 1 we present our definition of the UX construct, closely related to the QiU-model suggested in ISO9126 and 25000.

In prior publications [1,2,3] we have presented an approach that allows to master the measurable expression of NFRs on the I&EQ of software development artifacts. We will shortly present the concepts of our method in Section 2. In Fig 2 we present an example of the quality models suggested in ISO9126. We want to discuss some related terms, such as dependencies, Quality attributes, and Non-functional requirements.

We will discuss first hypotheses about probable and improbable relationships and present a first exploration of a dataset gained with 59 users of two webbased software systems in Section 3 and raise the question, whether such relationships can ever be generalized for all software products, or at least be explored for specific domains.

Finally in Section 4 we claim the position that UX can be mastered by construction (at least to some extent) using systematic methods, exploring the relationships among quality characteristics and capturing them in experience based quality models.

UX = QIU: A DEFINITION

In our understanding, UX is the perceived quality of a product that is being used within a sociotechnical environment. A similar concept can be found in ISO9126/ISO25000 [5] and [8] under the name Quality in Use (QiU). QiU according to the standard is a composition of the aspects: Effectiveness (ability to achieve specified goals), Productivity (appropriate amount of effort spent to achieve specified goals), Safety (acceptable level of risk) and Satisfaction (satisfy the user in a specified context).

Our initial UX model is a derivative of the QiU model, but further it is influenced by the concept of Acceptance in the Technology Acceptance Model (TAM) by Davis [6]. Thereby Acceptance signifies the overall evaluation of a product regarding attitude conformity and usage expectation from the user's perspective. Davis concept seems to be closely related to the concept of UX.

From these considerations we derive the following initial model (Figure 1), whose content and structure is currently being analyzed empirically. The results are not yet available at the time, we are writing this paper.

Figure 1: UX model derived from QiU in ISO9126

UX in our hypothesis is a composition of the aspects effectiveness, productivity, satisfaction and safety.

We expect effectiveness to influence UX rather indirectly through productivity, but all other factors to influence UX directly.

At this point in time we cannot provide empirical evidence about the correctness of these assumptions.

MASTERING INTERNAL/EXTERNAL QUALITY: A WELL ESTABLISHED METHOD

In this section, we give an overview on an industrial-proven [2] method to achieve a minimal and sufficient set of measurable NFR on the basis of I&EQ. We use this method to lay the foundation to construct quality into a software-based product. For this NFR method, the quality-needs, specified as NFRs, are on the level of I&EQ, so the link to UX was so far not explicitly addressed by our NFR method. The main features of the method are:

* a process for a common treatment of the various I&EQ.

* experience-based quality models (QM) that capture experience with general characteristics

...

...

Download as:   txt (22.2 Kb)   pdf (231.7 Kb)   docx (18.4 Kb)  
Continue for 12 more pages »
Only available on ReviewEssays.com