MUICSER: A MULTI-DISCIPLINARY USER-CENTERED

SOFTWARE ENGINEERING PROCESS TO INCREASE THE

OVERAL USER EXPERIENCE

Mieke Haesen, Kris Luyten, Karin Coninx, Jan Van den Bergh and Chris Raymaekers

Hasselt University - tUL - IBBT, Expertise Centre for Digital Media, Wetenschapspark 2, B-3590 Diepenbeek, Belgium

Keywords:

User-Centered Design, Software Engineering, Model-Driven.

Abstract:

In this paper we present an incremental and user-centered process to create suitable and usable user interfaces.

Validation is done throughout the process by prototyping, the prototypes evolve from low-ﬁdelity to the ﬁnal

user interface. Applications developed with this process are more likely to correspond to users’ expectations.

Furthermore, the process takes into account the need for sustainable evolution often required by modern soft-

ware conﬁgurations, by combining traditional software engineering with a user-centered approach. We think

our approach is beneﬁcial in its scope, since it considers evolving software beyond the deployment stage and

supports a multi-disciplinary team.

1 INTRODUCTION

In this contribution we present an overall approach for

the (re-)design of interactive applications that takes

into account the needs of the end-user as well as the

needs of the designer and developer. Our approach

can be used to design new applications that require in-

put from many different sources (e.g. domain experts)

or even to redesign legacy applications to improve the

user experience they offer. This approach builds on

existing models and notations in both the software

engineering and human-computer interaction (HCI)

community, providing extensions where necessary.

Based on our experiences and observations when

working with multi-disciplinary teams, we are gradu-

ally introducing an agile model-based approach in ap-

plied research and software development projects. We

will be using different models throughout the process,

where each model describes a speciﬁc aspect of an in-

teractive system and represents the viewpoint of one

or more speciﬁc roles in the multi-disciplinary team.

User-Centered Design (UCD) approaches are

meant to support the development process in such

a way that software quality attributes like usability

and overall compliance to the user needs are taken

into account. Therefore, combining a model-driven

approach and a user-centered development process

seems a natural way to ensure beneﬁts for the end-

user as well as the development team. The arti-

facts created using the proposed user-centered devel-

opment cycle contribute to well-considered and trace-

able design decisions. The need for communication

with end-users or customers results in additional mod-

els or artifacts (e.g. low-ﬁdelity and high-ﬁdelity

prototypes) on top of the commonly used models

in a model-driven approach, which are very helpful

to support the communication in a multi-disciplinary

project team (e.g. task model, presentation model

etc.).

In this paper, we present a user-centered software

engineering proces and detail the models and artifacts

that are at the center of the approach. A discussion of

our current and future work, as well as related work,

conclusions are presented.

2 MuiCSer

Our Multi-disciplinary user-Centered Software

engineering process, MuiCSer, embodies UCD

with a structured Agile Software Engineering

(ASE)((Larman, 2003)) process. Figure 1 gives an

overview of the proposed process. At ﬁrst this may

seem as a waterfall type of process model. However,

the development team can choose which stages are

tackled during one timeboxed iteration. At the end

of each iteration, there will be artifacts which can be

of various types (e.g. low-ﬁdelity prototypes, UML

diagrams, executable code,. . . ) and will be stored in

371

Haesen M., Luyten K., Coninx K., Van den Bergh J. and Raymaekers C. (2008).

MUICSER: A MULTI-DISCIPLINARY USER-CENTERED SOFTWARE ENGINEERING PROCESS TO INCREASE THE OVERAL USER EXPERIENCE.

In Proceedings of the Tenth International Conference on Enterprise Information Systems - HCI, pages 371-374

DOI: 10.5220/0001707303710374

 SciTePress

Figure 1: The MuiCSer Process.

a central repository ready for use in the next iteration.

The next paragraphs discuss the properties of the

process we propose in more detail.

The process typically starts with an analysis phase

in an initial iteration where the users tasks, goals and

the related objects or resources that are important to

perform these tasks are speciﬁed. If the user experi-

ence of a legacy system needs to be optimized, the

functionality of such a system can be often found

in existing manuals thus it can also contribute to the

analysis.

During the structured interaction analysis, the re-

sults of the ﬁrst stage are used to progress towards

system interaction models and presentation models.

These models are often expressed using the UML no-

tation, thus being kept in pace with the traditional SE

models.

Since both user needs and functional information

are speciﬁed, they can serve as input for the low ﬁ-

delity prototyping stage. User interface designers cre-

ate mockups of the user interface, based on the in-

formation contained in the task and interaction mod-

els, while using design guidelines and their experi-

ence. In subsequent phases, we go from low-ﬁdelity

to high-ﬁdelity prototypes, which on their turn evolve

into the concrete user interfaces. So each stage grad-

ually evolves towards the ﬁnal user interface, but is

still related to the artifacts created in a previous stage.

An overview of the artifacts involved can be seen

in Figure 2. It shows the artifacts created or updated

during the design and the relations between them.

By evaluating the result of each stage, the support

Figure 2: Artifacts created during the design process.

for user needs and goals and the presence of required

functionality is veriﬁed. If possible, an evaluation

with target users can be very useful to get feedback

from the end user directly. Because most of the ar-

tifacts do not present a fully functional system, part

of the testing takes place in a usability lab. To evalu-

ate some advanced prototypes, ﬁeld tests can examine

the user interface in more realistic situations. If the

results of a phase are too complex to involve an end

user during evaluation, it is still necessary to evaluate,

verify or validate the models or prototypes.

MuiCSer as depicted in Figure 1 shows only one

iteration (or increment) in the development of the

complete system. The back arrows at the upper side

of the ﬁgure emphasize that one can iterate over the

previous phase, depending on the results of the evalu-

ation of the current artifact.

3 HCI ENGINEERING MODELS

The proposed notations are built on existing model

notations used in software engineering and interface

design and speciﬁcation. The following sections de-

scribe different models being created, changed and

transformed during the execution of the process in or-

ICEIS 2008 - International Conference on Enterprise Information Systems

372

der to support a smooth transfer towards the ﬁnal user

interface. These models and their interrelations are

shown in the left part of Figure 2.

3.1 The User Task Model

User task models specify requirements for an appli-

cation from a users’s point of view, in a hierarchical

way. We use the ConcurTaskTrees (CTT) notation

(Mori et al., 2002) in the second stage of MuiCSer. It

has proven to be easy to understand and usable, even

for complex systems. In our previous work, we in-

dicated that working prototypes with inter- and intra-

dialog navigation can be generated by combining this

notation with high-level user interface descriptions

(Luyten et al., 2003).

Van den Bergh and Coninx (Van den Bergh and

Coninx, 2006) showed a correspondence between the

CTT notations and UML 2.0 activity diagrams. This

mapping results in the system interaction model as de-

scribed in section 3.3.

3.2 The Application Model

The application model is traditionally used to describe

the objects and commands provided by the applica-

tion logic. The model is represented using the UML

class diagrams and can be considered most closely

related to the data or object model. The application

model is useful to describe the interface between the

user interface and the application logic.

Furthermore, the application model is an impor-

tant model for legacy systems, since the application

logic is already available when the design of a (new)

user interface starts.

3.3 The System Interaction Model

The system interaction model offers a system-centric

description of action ﬂows that can be performed by

the system, by a user through the system, or that

have an inﬂuence on the system. The model is ex-

pressed using UML 2.0 activity diagrams, extended

with new stereotypes corresponding to the task types

in the CTT notation.

3.4 Abstract Presentation Model

The abstract presentation model deﬁnes the structure

of the user interface in a platform independent man-

ner describing the logical and temporal structure of

the user interface through a hierarchy of abstract in-

teraction components. The model is expressed using

the UML class diagram, in which each component is

represented by a stereotyped class

A complete formal mapping to the Canonical Ab-

stract Prototypes (CAP) (L. Constantine, 2003) is not

possible due to the informal semantics, rudimentary

rules were however identiﬁed. The notation will be

used as a front-end for a designer to add additional

information, such as screen organization.

In MuiCSer, we employ high-level user interface

descriptions to deﬁne the high-ﬁdelity prototype. We

prefer UIML (Luyten and Coninx, ) and XForms (Van

den Bergh et al., 2007) to render the prototypes, but

other user interface description languages are also

possible. These languages should be XML-based and

independent of the application logic, to smooth the

transformation between prototypes and the ﬁnal user

interface.

4 ONGOING AND FUTURE

WORK

The process presented in this paper has been tested on

software projects of limited complexity and, by con-

sequence, with a development team of limited size.

Although our tests did not include any larger soft-

ware projects, the process should be ﬂexible enough

to support the increased complexity and team size.

One of our main contributions is that different do-

main experts can use their own notations to create

models which can relate to models of other domain

experts, in order to obtain a complete and usable in-

teractive system with respect to the requirements. We

are testing this approach in different application do-

mains (applications for logistics, community websites

etc.), requiring different experts to collaborate.

We plan to gradually improve the relation between

the different types of artifacts, which is a key fac-

tor for success. The combination of HCI models

and UML models guarantees a smooth integration of

the user interface and application logic. Putting for-

ward the combination of models explicitly also pre-

vents mismatches between the functionality provided

by the application logic and the functionality accessi-

ble through the user interface.

5 RELATED WORK

This paper is not the ﬁrst to look into the relation be-

tween user-centered design and both HCI and soft-

Stereotypes (Object Management Group, 2004) are

standardized means to extend the UML language.

MuiCSer: A MULTI-DISCIPLINARY USER-CENTERED SOFTWARE ENGINEERING PROCESS TO INCREASE

THE OVERAL USER EXPERIENCE

373

ware engineering models. Hudson (Hudson, 2004)

describes a way to use UML in combination with

user-centered design, but does not address how user

interfaces can explicitly be modeled in UML.

A study on the state of UCD by Mao et al. (Mao

et al., 2005) showed among other things that although

many parts of UCD are practiced in actual develop-

ment in industry, end-to-end deployments are seldom

due to limitations of current practice.

Wisdom (Nunes, 2001) deﬁnes a set of models

using the standard UML-diagrams and stereotypes

to create more appropriate notations for a variety of

models that can be used to design interactive systems

in small teams. Wisdom also provides a user-centered

methodology to effectively use the notation.

Campos et al. (Campos and Nunes, 2005) pro-

posed a workstyle model, Galactic Dimensions, that

focuses on workstyle transitions rather than a design

process. The workstyle model allows to evaluate tools

according to different dimensions that characterize

different workstyles.

6 CONCLUSIONS

In this paper we introduced a novel process, prac-

ticing User-Centered Software Engineering in such

ways that methodologies used by developers as well

as the creativity of designers are included and a posi-

tive user experience is more likely to be obtained. Al-

though both HCI and Software Engineering models

are integrated into one process, it is up to the multi-

disciplinary project team which models will be used.

To provide even more ﬂexibility, we also included ex-

plicit support for legacy systems being updated or ex-

tended.

Whenever possible, an evaluation with end users

is recommended in each step. However since end

users are not always available or should not under-

stand all models, validation or veriﬁcation by the

project team is suggested. This approach also encour-

ages the development of software in several iterations

and increments.

ACKNOWLEDGEMENTS

Part of the research at EDM is funded by the

ERDF (European Regional Development Fund) and

the Flemish Government. The VIP-lab project

(4-BMG-II-2=37), is ﬁnanced by the “Interreg

Benelux-Middengebied” authorities and co-ﬁnanced

by Province of Limburg (B), Province of Limburg

(NL), Ministry of Economic Affairs (NL) and Min-

istry of Flemish Government/Economic Affairs (B).

The MuiCSer process is also based on our ex-

periences in IWT projects Participate (with Alcatel-

Lucent) and AMAS++ (IWT 060051).

REFERENCES

Campos, P. F. and Nunes, N. J. (2005). Galactic Dimen-

sions: a Unifying Workstyle Model for User-Centered

Design. In INTERACT’05, volume 3585 of LNCS,

pages 158–169. Springer.

Hudson, W. (2004). Object Modeling and User Inter-

face Design, chapter Toward Uniﬁed Models in User-

Centered and Object-Oriented Design, pages 313 –

362. Addison Wesley.

L. Constantine, L. (2003). Canonical abstract prototypes for

abstract visual and interaction design. In Proc. DSV-IS

2003, volume 2844 of LNCS, pages 1–15. Springer.

Larman, C. (2003). Agile and Iterative Development: A

Manager’s Guide. Addison-Wesley.

Luyten, K., Clerckx, T., Coninx, K., and Vanderdonckt, J.

(2003). Derivation of a Dialog Model from a Task

Model by Activity Chain Extraction. In Proc. DSV-IS

2003, volume 2844 of LNCS, pages 203–217.

Luyten, K. and Coninx, K. Uiml.net: an open uiml renderer

for the .net framework. pages 257–268.

Mao, J.-Y., Vredenburg, K., Smith, P. W., and Carey, T.

(2005). The state of user-centered design practice.

Communications of the ACM, 48(3):105–109.

Mori, G., Patern

o, F., and Santoro, C. (2002). CTTE: sup-

port for developing and analyzing task models for in-

teractive system design. IEEE Transactions on Soft-

ware Engineering, 28(8):797–813.

Nunes, N. J. (2001). Object Modeling for User-Centered

Development and User Interface Design: The Wisdom

Approach. PhD thesis, Univ. da Madeira.

Object Management Group (2004). UML 2.0 Superstruc-

ture Speciﬁcation.

Van den Bergh, J. and Coninx, K. (2006). Cup 2.0: High-

level modeling of context-sensitive interactive appli-

cations. In Proc. MoDELS 2006, volume 4199 of

LNCS, pages 140–154.

Van den Bergh, J., Huypens, S., and Coninx, K. (2007). To-

wards Model-Driven Development of Staged Partici-

patory Multimedia Events. In Proc. DSV-IS, volume

4323 of LNCS, pages 81–94.

ICEIS 2008 - International Conference on Enterprise Information Systems

374