Towards an Approach for Incorporating Usability Requirements into

Context-Aware Environments

Dorra Zaibi, Meriem Riahi and Faouzi Moussa

LIPAH Laboratory, Faculty of Science of Tunis, University of Tunis El Manar, Tunis, Tunisia

Keywords: Ubiquitous Computing, Context-Aware Systems, Usability Rules, Inference System, Model-Driven

Architecture.

Abstract: With the considerable advancement of technologies and the proliferation of mobile devices, evaluating the

usability of software applications has become an emerging research area. Hence, improving the quality of

software applications is crucial in context-aware environments. For that reason an increasing attention is

drawn towards the development and the adoption of appropriate research proposals able to evaluate the

mobile application usability. This article is a contribution to proposing a methodology for the development

of context-aware systems based on usability requirements during the user interface design stage. In

particular, this approach focuses on how to infer consistent context-based usability requirements and how to

incorporate these requirements into a user interface development process. As a proof of the proposal

concept, we have applied our methodology to an illustrative case study. More, experiments with end users

have been carried out.

1 INTRODUCTION

Nowadays, ubiquitous computing is indispensable in

the improvement of customized access to a wide

range of mobile devices and services (Pathan and

Reiff-Marganiec, 2009). In his/her ordinary activities

and on a daily basis a user engages simultaneously

with many smart devices. Thus, the concept of

context of use plays a paramount role in the

development of ubiquitous systems (Dey and Abowd,

2000). It involves information related to various

aspects such as screens, locations, users, etc.

Actually, it is proved in literature studies that the

context of mobile devices is highly dynamic when

compared with the traditional desktop computers

(Harrison et al., 2013). For that purpose, considerable

challenges and attention have been raised with respect

to usability inconsistencies of context-aware systems.

Accordingly, in order to provide better usability

quality in ubiquitous environment, the evaluation

process from the earlier stages of the user interface

design process is crucial; mainly in the case of non-

tolerant interactive systems failures in critical

domains (Serral et al., 2010). User Interface (UI) has

to show information in the best possible way in

order to minimize error risks and erroneous

manipulations. In this context, we believe we need a

new generation of methodologies which allows

context-aware systems developers to work while

taking into consideration usability requirements right

from the early UI design phase.

Actually, MDA (OMG MDA, 2003) is being

largely explored. It has been proven that it is quite

appropriate therefore becoming an essential paradigm

for the software system design and development. In

the last decade, the Human-Computer Interaction

community (HCI) has highlighted the benefits of the

MDA technology and accordingly is moving towards

it (Oliviera et al., 2013). Generally, in such a process,

attention is focalized in data and functional modelling

neglecting usability aspects in context-aware

environments. Hence, there is a need to require the

MDA process extension for the purpose of supporting

context-based usability as a first class entity in the

user interface development process.

The present article aims to propose a model-

driven approach for the integration of usability

requirements into context-aware systems. The remain-

der of this paper is structured as follows. Section 2

describes the related works. Section 3 presents the

proposed approach. Section 4 presents the case study.

Section 5 describes the experiments of the adaptive

system with end-users. Finally, the conclusion and

further works are presented in Section 6.

Zaibi, D., Riahi, M. and Moussa, F.

Towards an Approach for Incorporating Usability Requirements into Context-Aware Environments.

DOI: 10.5220/0006702404630474

In Proceedings of the 20th International Conference on Enterprise Information Systems (ICEIS 2018), pages 463-474

ISBN: 978-989-758-298-1

463

2 RELATED WORKS

In this section, we first present some approaches

dealing with context-aware systems development.

Second, we present some proposals that focus on

usability evaluation. Finally, we present a third

collection that brings these two topics together.

2.1 Context-Aware Approaches

Since the emergence of ubiquitous computing, great

research efforts have been devoted to the subject of

context-aware systems and several approaches have

been proposed. We focus our attention on the studies

that deal with context-awareness in Model-Driven

environment. A notable example is of (Paterno et al.,

2009) who suggest a universal model-based

language for UI, called MARIA, to support the

development of UI for interactive applications based

service-oriented architectures in ubiquitous

environments. The authors focus on exploiting the

novel model-based language to provide useful

support at both design and runtime. (Serral et al.,

2010) propose a specification of context-aware

pervasive systems through Model-Driven approach.

PERVML, a domain specific modelling language,

has been defined to describe the system functionality

in a platform and technology independent way in

order to model the pervasive system. (Oliveira et al.

2013) present a Model-Driven approach which takes

into account the content personalization of the UI

since earlier design stages.

As in the first group, these approaches, in spite

of focusing on context-aware UI adaptation, neglect

usability measures. They give support for the

development of context-aware systems; yet, do not

tackle the problem of preserving usability when

adapting the UI.

2.2 Usability Evaluation Approaches

Numerous are the research studies that deal with

usability in Model-Driven environment. For

instance, a model-driven approach to evaluate the

usability of multi-platform graphical UI is proposed

by (Aquino et al., 2010). The usability has been

quantified in terms of satisfaction, effectiveness and

efficiency. The graphical UI is evaluated using

small, standard and large size screens. (Gonzalez-

Huerta et al., 2010) propose an architecture to

support model-driven development process which is

guided by quality attributes. Model transformations

are specified and executed. The alternative

transformations are selected considering the desired

qualities of a particular target model. A method is

proposed by (Panach et al., 2014) for evaluating

internal usability. The authors deﬁne metrics for

conceptual primitives. Based on conceptual models,

the evaluation can be performed automatically and

applied to any Model-Driven Development method.

Nevertheless, these proposals do not support

usability evaluation with regard to changing context

requirements that involves a considerable challenge

to mobile devices’ effectiveness. Currently, some

research studies of usability-based context-aware

approaches have been found. Next, we deal with

these studies.

2.3 Usability-based Context-Aware

Approaches

(Ormeno et al., 2013) present a method to facilitate

the usability requirements of capture process at early

stages. The method consists of defining a tree

structure including interface design guidelines and

usability guidelines. A question-answer format was

used to enable the capture of requirements, through

an interview with the end-user. Using model to

model transformations, the usability requirements

are transformed into a conceptual model of any

existing Model-Driven Development method. The

result of the interview is a set of designs that the

system must satisfy. (Ben Ammar et al., 2015)

propose a Model-Driven method for integrating

usability guidelines into model transformation

process. The proposed method aims to obtaining a

UI which fulﬁls the desired usability attributes.

Usability properties are associated with the possible

alternative transformations, and parameterized

transformations are executed. (Hentati et al., 2016)

propose an MDE approach for optimizing usability

in interactive systems generation process. Three

main stages are fulfilled in their proposal: (1)

generating all the possible concrete UI from a given

abstract UI, (2) optimizing the usability by means of

metrics to measure the user interface usability value

considering a given context of use, (3) selecting the

alternative model transformation in order to generate

the optimal usability UI. A context based evaluation

method for the assessment of the quality in use of

mobile systems is proposed by (Ben Ayed et al.,

2017). The authors implement the Evaluation

Support System (ESS) in order to capture interaction

data and contextual information when using mobile

application. In addition to that, it allows the

quantitative measurement of criteria, defined by the

standard ISO/IEC 25010.

ICEIS 2018 - 20th International Conference on Enterprise Information Systems

464

The study of the aforementioned proposals

allows us to underline some limitations:

 Lack of dynamic selection of usability

requirements in context-aware environment:

All proposals are incapable to select

dynamically the usability requirements with

regard to the change of context of use. Only

static adaptation is provided.

 Difficulty to select consistent usability

guidelines: Applying a set of usability

guidelines may have negative impact and

provide conflicting influence in the whole

usability of the UI. In most proposals, there

doesn’t exist any system able to select

appropriate usability guidelines in order to

perform conflict resolution. We can notice as

an example of conflicting usability guidelines,

that a usability guideline, which improves the

user control, generally adds undo/cancel

buttons to the UI. The application of this

guideline may have negative impact in the

information density attribute (e.g. limiting the

amount of information in small screen size).

 Lack of usability guidelines in context-aware

environments: no existing approaches have

proposed novel usability guidelines for

context-aware environments. All of them

focus on existing context-based usability

standards and recommendations issued from

literature such as (ISO, 2010).

 Lack of empirical validation: Scientific

methods need to be empirically validated in

order to provide evidence about their

effectiveness. Only a minority of the proposals

has been empirically validated (Ben Ammar et

al., 2015; Ben Ayed et al., 2017).

Although dealing with usability in context-aware

systems is not novel, we can notice that this

important research field is still in its early stage and

many more research studies are recommended.

To cover this need, we propose a methodology

that focuses on usability requirements during the

adaptation of context-aware UI. The goals of our

proposal are: 1) the selection of consistent usability

requirements in context-aware environment is

proposed, 2) novel generation of context-based

usability rules are defined into a specific application

domain, 3) the model-driven development (MDD)

process is fulfilled during the context-aware UI

modelling, and 4) experiments with end-users are

carried out.

3 PROPOSED APPROACH

The main goal of this approach is to include

usability guidelines in the context-aware process

from the beginning of the UI modelling stage. We

would like, while designing a UI, to set which

consistent usability guidelines should be provided

for each context situation.

Our approach is summarized in Figure 1. The

system is developed by performing two stages;

taking into consideration our knowledge of regular

context information. In this sense, a specific system

is held at runtime, updating context information

constantly. During design time, we use this

information when designing the UI.

To define our approach, two main stages are

considered:

Stage 1: Selecting consistent usability rules. A

set of usability guidelines is established to define

consistent usability requirements considering current

context of use.

Stage 2: In this stage, Model-driven development

is adopted first, which uses the MDA approach and

produces transformation models. We specify three

MDA transformation levels namely: Computational

Independent Model (CIM as CIM), Contextualized

Platform Independent Model (CPIM as PIM) and

Contextualized Platform Specific Model (CPSM as

PSM). The context and the usability guidelines

established in stage one are taken as input in order to

generate a concrete model with the consistent

usability guidelines. At a second stage, a model-to

code transformation which allows the generation of

adaptive UI is processed. Following a detailed

explanation of each of these stages.

3.1 Stage 1: Selecting Consistent

Usability Rules

The aim of this stage is to provide the right usability

guidelines for a particular user considering its

context information. To achieve this purpose, an

intelligent inference system is provided.

A set of usability guidelines specific to our

application domain (detailed in the case study

section) is identified, initially, and stored in a

usability knowledge base. These guidelines are

presented in form of production rules. Each with a

premise and a conclusion. The context-based

usability rules are of the following type: IF (context

* context

) THEN usability_guideline.

where (context

* context

) are called

Premise (set of conditions) and usability_guideline

is called Conclusion.

Towards an Approach for Incorporating Usability Requirements into Context-Aware Environments

465

Figure 1: Overview of the proposed approach.

Earlier, the usability engineer attributes a priority

index to usability rules with respect to the popula-

tion characteristics and/or the task requirements.

In experts’ experiences, there is always a

possibility of conflicts. However, with traditional

inference systems, it is difficult to treat conflicting

rules and thus the deduction of erroneous knowledge

conducts to mistakes in decision making. There may

exist varied conflicting rules:

 Conflict between conclusions: For example, a

conclusion that provides different steps to

inform and guide a user (prompting attribute)

and another conclusion that reduce the set of

an action steps (brevity attribute). These

conclusions contribute to a conflict.

 Conflict between premise and conclusion: For

example, Rule

: IF environment is noisy

THEN display visual notifications (feedback

attribute) and Rule

: IF user is visually

impaired THEN apply vocal mode (flexibility

attribute). We can notice that the premise

(environment is noisy) has a negative impact

on the conclusion (apply vocal mode).

In order to perform the conflict resolution, we

adopt a decision-making strategy. Thus, a decision

matrix is used to allow the identification of

relationships between the set of rules. As we deal

with two kinds of conflicting rules, two kinds of

decision matrixes are, initially, specified by the

usability engineer to indicate whether a conflict

between two rules exists (Figure 2):

 Type decision matrix 1: a decision matrix

which identifies the semantic relation between

different rules conclusions. The elements of

the matrix are either zero or one. For example,

the usability engineer assigns the value 0, if

there is a conflict between the conclusion of

the rule 1 (Rule

(Conclusion)) and the

conclusion of the rule m (Rule

(Conclusion)),

otherwise 1;

 Type decision matrix 2: a decision matrix

which identifies the semantic relation between

premise and conclusion of different rules. The

elements of the matrix are either zero or one.

For example, the usability engineer assigns the

value 1, if there is a conflict between the

premise of the rule 1 (Rule

(Premise)) and the

conclusion of the rule 2 (Rule

(Conclusion)),

otherwise 0;

Figure 2: Decision matrixes.

Concerning context instances, a context meta-

model is defined at the beginning, which would be

exploited in the next stage all along the model-

driven development process. By this way, a set of

context instances are provided to describe different

context situations.

3.1.1 Inference System

An inference system is proposed in this section

which aims to infer consistent context-based

ICEIS 2018 - 20th International Conference on Enterprise Information Systems

466

usability rules. The different steps of our inference

system are as follows:

 Step1: Filtering rules

 Step2: Sorting filtered rules

 Step3: Managing conflicts between

conclusions

 Step4: Managing conflicts between premises

and conclusions.

Next, we detail each of these steps.

Step 1: Filtering Rules

A directed graph (or digraph) representation which

represents context-based usability rules in order to

handle the step of filtering rules is used in this paper.

As previously mentioned, a rule is in form of: IF

premise THEN conclusion, where the premise is a

set of conditions. Figure 3 illustrates the structural

dependencies of the digraph. It is constructed by: (1)

a set of nodes or vertices; (2) a set of arcs or directed

edges: all arcs have arrows that give direction. The

graph can only be traversed by the direction of the

arrows.

The proposed graph has three levels. The first

level identifies the context dimension (i.e. user,

environment or platform). The second level

identifies all the possible conditions alternatives of

rules. The third level includes all the possible

conclusions of rules. For each conclusion node, one

or a set of conclusion nodes are associated,

occurring in the premise part of a rule. Thus, a

completed rule is identified by the path from a leaf

to a root.

By using a filtering algorithm, we can evaluate

the premise with the set of the current contextual

parameters and check whether all the condition

elements of the premise part in a rule are satisfied.

Coming up next is a description of the filtering

algorithm:

Input: G: digraph, Parameters[]: input contextual

parameters.

Output: Filtered_List[] : a set of the filtered rules

(1) Iterate and check the context type in the graph G.

(2) Traverse through a sub-graph and check if

condition node exist in Parameters[].

(3) Traverse through the satisfied condition node and

check if the conclusion node has not yet been tested.

(4) Check if the conclusion node has one or more

successors.

(5) Iterate and go to the Step 1 until Parameters[] is

totally tested.

Step 2: Sorting Filtered Rules

After the filtering rules step and with the purpose of

managing the conflict set between rules, the set of

the satisfied rules are sorted in order of priority. We

first sort the rule with the highest priority index. An

existing hybrid algorithm is used as an efficient

implementation combining an efficient algorithm for

large data sets (the merge sort) with insertion sort for

small data sets.

Step 3: Managing Conflicts Between Conclusions

In order to make the decision for the consistency of

the rules, we used the type decision matrix 1 which

allows handling conflicts between conclusions. The

steps of this algorithm are: (1) comparing each

conclusion of the highly priority rule with the other

conclusions of the sorted rules, (2) if the conclusion

of the rule j have a negative impact on the

conclusion of the rule i, remove the rule j from the

filtered list.

Figure 3: Example of a digraph.

Towards an Approach for Incorporating Usability Requirements into Context-Aware Environments

467

Step 4: Managing Conflicts Between Premises

and Conclusions

In this step, we used the type decision matrix 2

which allows handling conflicts between premises

and conclusions. The steps of the algorithm are: (1)

comparing each premise of the highly priority rule

with the other conclusions of sorted rules, (2) if the

conclusion of the rule j has negative impact on the

premise of the rule i, remove the rule j from the

filtered list.

At the end of this step, a set of consistent

context-based usability rules are provided.

3.1.2 Dealing with Usability Guidelines in

Model-Driven Transformation Process

In order to be maximized and incorporated in the

next stage (the model-driven development process),

a usability model which contains the consistent

usability rules (previously resulted) should be

provided.

The proposed usability model extends the one

described in (Harrison et al., 2013). In such model,

usability is classified into seven groups and is

presented in Table 1.

To define what should be associated with each

usability group, we did a detailed literature review

about usability modelling. As a result of this review

(Table 1), we identified for each usability group the

opted usability attributes which we consider

appropriate due to their frequent use in the most

cited usability model like (Abrahao and Insfran,

2006; Ben Ammar et al., 2015; Scapin and Bastien,

1997; Seffah et al., 2006; Zhang and Adipat, 2005).

In the literature of model transformation, in order

to use a model, the definition of the meta-model is a

precondition. For that purpose, we define usability

meta-model in order to formalize the approach.

Figure 4 shows the usability meta-model. It is

composed of a UsabilityGroup, which defines a set

of features used to evaluate the quality of user

interface; UsabilityAttribute, which is a refined sub-

group of the usability group and UsabilityRule,

which includes ConditionGroup describing a set of

condition groups (i.e. WHEN clause) and

Conclusion, defining the usability guidelines (i.e.

THEN clause).

In this manner, a usability model, conforms to its

meta-model and enables the definition of the

context-based usability rules. Using converting

algorithms, the set of the consistent usability rules

has been converted into the usability model referred

to its usability meta-model, in order to be later

incorporated in stage 2 (Figure 5).

Table 1: Decomposition of usability groups.

Group

Attribute

Reference

Learnability

System

feedback

(Ben Ammar et al.,

2015)

Grouping

(Scapin and Bastien,

1997)

Legibility

(Scapin and Bastien,

1997)

Prompting

(Ben Ammar et al.,

2015)

Cognitive

Load

Brevity

(Scapin and Bastien,

1997)

Information

Density

(Ben Ammar et al.,

2015)

Navigability

(Ben Ammar et al.,

2015)

Satisfaction

Flexibility

(Scapin and Bastien,

1997; Seffah et al.

2006)

Error

Error handling

(Seffah et al. 2006)

Error

prevention

(Abrahao and

Insfran, 2006)

Efficiency

Speed of

accessing data

(Zhang and Adipat,

2005; Seffah et al,

2006)

Fast access to

common tasks

(Abrahao and

Insfran, 2006)

Effectiveness

Completeness

(Zhang and Adipat,

2005: Seffah, 2006)

Memorability

Time to

remember

(Zhang and Adipat,

2005; Abrahao and

Insfran, 2006)

Accuracy

(Abrahao and

Insfran, 2006)

Figure 4: Usability meta-model.

ICEIS 2018 - 20th International Conference on Enterprise Information Systems

468

Figure 5: Transformation of consistent usability rules into

usability model.

3.2 Stage 2: Model-Driven

Development

In this stage, we consider the MDA structure. Before

presenting the model transformation process, we will

explain the main models.

 Computational Independent Model

In order to deﬁne the CIM model, we use Business

Process Modelling Notation (BPMN). As cited in

(Brossard et al., 2007), the business process

becomes the central part of interaction modelling

since it is able to: 1) model all the business goals of

the user in an application, 2) define the tasks

concerning the specification of each business goal to

accomplish it. These tasks are interactive tasks (e.g.

data entry), non-interactive tasks and manual tasks,

and 3) consider all information flows made between

several actors (human or machine) in a single

business process.

A better known formalism, the (BPMN, 2006), is

chosen which allows the definition of business

processes. We justify the use of the BPMN by its

ability to model the functional tasks and to describe

the business logic of the application. Furthermore,

this formalism is especially able to model all

information flows between tasks which is

particularly important for the integration of the

content adaptation. Moreover, additional extensions

have been performed to the BPMN. Indeed, each

BPMN element in the BPMN model has been

annotated with interaction element types in order to

describe the type of interaction with the user such as:

types of input information, output or grouping

information.

 Usability Model

Details about usability meta-model have been

already explained in the previous section.

 Context Model

A defined context meta-model is used during the

model transformation process. The context contains

a description of user, environment and platform

dimension. To identify the information about those

elements, a literature review has been fulfilled

(Jumisko-Pyykko et al., 2010).

 Interactor Model

The interactor model is used during the CPIM-to-

CPSM transformation to define design elements of a

target interface. It is composed of the interactors

which are usually found in tool boxes such as

SWING or HTML. A set of containers and contents

are specified. The proposed interactor meta-model

was initiated by (Sottet, 2008) and then extended.

 Contextualized Platform Independent

Model

The CPIM model is the output of the first

transformation (CIM-to-CPIM transformation) and

the input of the second transformation (CPIM-to-

CPSM transformation). It allows the description of

the user interaction independently of any platform.

It is specified in Generic UIML (User Interface

Markup Language) meta-model (UIML, 2008). We

used UIML as a language for CPIM and CPSM level

representation because it allows, in an abstract way,

the applications development for the specification of

the UI elements. Moreover, we are able to provide

the transformation from UIML to source code in

diverse platforms (for example, the Eclipse plugin

Acceleo allows the conversion to HTML, Java, and

Android). The generation of the CPIM model allows

the definition of the structure, behavior and style

parts of the interface component to be generated.

 Contextualized Platform Specific Model

The CPSM model is the output of the second

transformation (CPIM-to-CPSM transformation). It

is a refinement of the CPIM. Indeed, it allows the

description of the user interaction for a specific

platform. It is specified in UIML meta-model. The

definition of the presentation, logic and style parts of

the interface component will be generated.

The model transformation definition consists

of a set of transformation rules. We notice that

model transformations are implemented with

ATLAS Transformation Language (ATL) (ATL,

2006). This language (ATL) allows developers to

define transformation rules in order to describe how

source model elements are matched and navigated in

order to create and produce the target model

elements. By applying the CIM to CPIM

transformation (T1), the BPM model is associated

with usability and context model to allow the

content and behavior adaptation. The generation of

the structure, behavior and style parts are as follow:

 For the structure part, the transformation is

accorded to the BPMN element, to which we

associate a type of interaction element that

describes a specific task. It allows specifying

Towards an Approach for Incorporating Usability Requirements into Context-Aware Environments

469

the interface with the general information. We

adopt a generic vocabulary of UI elements,

used in conjunction with UIML and which can

specify any user interface for any platform

(Zaibi et al., 2016). (Ali et al., 2002) provide a

more detailed description of it.

 For the behavior part, in order to allow

content and behavior adaptations, the user

interaction is specified with the UI by defining

rules. The content adaptation of interactive

tasks is applied by using the auto-fill form

method for the input information extracted

from the context model. The behavior

adaptation of interactive tasks is applied by

incorporating the usability guidelines.

 For the style part, a UIML code is defined

manipulating the properties correspondent to

content and to specific properties for each UI

element, associated to non-interactive tasks

(from usability guidelines for graphical UI

independently of a particular platform).

By applying the second transformation (T2), a

target concrete CPSM model is generated. A set of

transformation rules is, thus, established. For each

transformation rule, the context, usability and

interactor model are taken into account. The

presentation, logic and style parts are generated as

follows:

 By using the interactor model, the designer

can specify each particular platform. The

interactor model has a major advantage since

it allows developers to avoid implementations

at the code generation phase. In this way, the

presentation part is specified to join generic

UIML classes with a specific platform through

the interactor model.

 A logic part statement will be included

containing match between the methods used in

the behavior part and those defined in the

interactor model for a target platform.

 The style part is generated, containing content

and properties specific for each UI element

associated to non-interactive tasks

dependently of a specific platform.

Code Generation. The last model-to-text

transformation is made, to produce the adaptive UI.

(Acceleo, 2006) a tool which is built on an MDA

based generator, has been supported to accomplish

model-to text transformation. Acceleo is chosen

because of (1) its adequacy for quickly writing rules

for the generation of UI prototype (Model to Text)

and (2) its easiness to integrate the existing ATL

code in the template.

4 CASE STUDY

In order to expose its applicability, we have applied

our methodology to an illustrative case study. In this

paper, the object of the case study is the work order

management. The scenario is the following:

Following equipment failure detection, authorized

users should be able to connect to the application’s

database to transfer their work requests. This

functionality enables requester, in industrial fields,

to create a work request by providing information

about it. The system will review the work requests

and accordingly either approve or reject the

submission. Only approved submissions are

converted into work orders and attributed to

available staffs into a job process. Then, to proceed

on curative interventions, the technician can retrieve

the information about the work that has to be done

by consulting all information on the work form.

After terminating the work, the technician can send a

description of the work he processed.

Given that the work order management system

is large, we focused our interests on the generation

of adaptive UI for the <create work request> and

<prepare work orders> tasks. The Business Process

Model illustrates our scenarios in the left part of

Figure 6.

Different contexts of use have been presented in

order to implement our tasks. Table 2 presents the

various contexts of use.

Table 2: Example of different contexts of use.

User profile

User

Environment

Platform

C1:

Requester

Wearing

gloves

Silent

Tablet

C2:

Requester

Wearing

gloves

Noisy

Smartphone

C3:

Technician

Novice

In the car

Smartphone

C4:

Technician

Expert

In the car

Smartphone

4.1 Selecting Consistent Usability Rules

In this study, we are focused on the definition of

usability rules regarding the mobility and its

consequence. Considering the special circumstances

of an industrial environment, we have defined

context-based usability rules. An example of

usability rules in industrial context-aware

environment is presented in Table 3. Beforehand,

priority index is attributed by the usability engineer

to each usability rule (e.g. R1: (priority=2), R4 and

R5: (priority=1)). Besides, two different decision

ICEIS 2018 - 20th International Conference on Enterprise Information Systems

470

matrixes are specified by the usability engineer in

order to manage the conflict resolution steps.

In order to infer consistent usability rules, the

proposed inference system is used by performing the

different steps previously explained. The different

algorithms are implemented in java.

Once all the different inference system steps

have been accomplished, consistent usability rules

are resulted. Table 4 shows the consistent usability

rules for each particular context of use. We note for

example that in context C1, R1 and R5 are

conflicting rules since the premise of R5 has a

negative impact in the conclusion of the rule R1.

The XML file presented in part (a) of Figure 7

shows the consistent usability rule related to the

context C1. In order to be maximized in the model

transformation process, algorithms are implemented

in java in order to handle XML data of consistent

usability rules and retransform them into a new

XML-based file which is correspondent to the

usability meta-model. Part (b) of Figure 7 presents

the usability model.

Table 3: Example of usability rules in industrial fields.

Usability rule

Signification

R1: IF platform=

smartphone/tablet THEN

use tactile interaction

Using tactile interaction

with smartphone/tablet

platforms.

R2: IF user = novice and

location = in the car

THEN provide the GPS

functionality

Providing GPS

functionality for novice

workers in order to get to

the work location.

R3: IF platform =

smartphone THEN

provide the relevant

information.

Providing relevant

information in small

screen size (e.g.

smartphone).

R4: IF user = wearing

protective gloves and

noise=high THEN scan

the equipment code.

Scanning the equipment

code for workers who have

got protective gloves and

are in noisy environment.

R5: IF user = wearing

protective gloves and

noise= low THEN apply

vocal interaction.

Switching to vocal mode

for workers who have got

protective gloves and are

in noiseless environment.

Table 4: Example of consistent usability rules.

Context

Consistent usability rules

R3, R1

R2, R1

4.2 Model-Driven Development

This stage fulfils the model transformation process.

4.2.1 CIM to CPIM Transformation

A first model transformation T1 allows the generation

of the CPIM. The BPM model is in conformance with

the BPM meta-model. The execution of the ATL

transformation rule is a target CPIM. The CPIM is in

conformance with the UIML meta-model. Part (a) of

Figure 6 presents the structure part of the generated

CPIM. Part (b) of Figure 6 presents the style part. It

describes an example of the integration of usability

requirements with a non-interactive task. Parts (c) and

(d) of Figure 6 illustrates the behavior part for the

<create work order> task. These parts show

correspondently content (e.g. form auto-filling) and

behavior (e.g. usability integration with interactive

tasks) adaptation.

4.2.2 CPIM to CPSM Transformation

In the second model transformation (T2), a transition

is implemented in order to obtain the concrete CPSM

considering the context, usability and interactor

model. The complete UIML code, which is defined to

a specific platform, describes the code generated for

the presentation, logic and style parts of the generated

CPSM. An example of the generated target model

(CPSM) is showed in the right part of Figure 6.

4.2.3 CPSM to Code Transformation

Adaptive UI are generated according to Acceleo

templates. Figure 8 shows the generated UI for the

<create work request> task accordingly to context

C1 and C2. The generated UI are web-based. In the

case of context C1 (left part of Figure 7), a requester

wearing protective gloves and sitting in a noiseless

environment is using the application through a

tablet. Flexibility is considered by switching to vocal

mode in order to enable the work request creation. In

the case of context C2 (right part of Figure 7), a

requester wearing protective gloves, and sitting in a

noisy environment is using the application through a

smartphone. Flexibility is considered by allowing the

scan of the code of the malfunctioning equipment.

Figure 9 shows the generated UI for the <process

work orders> task. In the case of context C3 (left

part), an expert technician, in the car, is consulting

the work order information. Added to providing

tactile interaction (R1), Brevity is considered by

providing the relevant information. In the case of

context C4 (right part), a novice technician, in the

car, is consulting the work order information in

order to get to the work location. Added to providing

tactile interaction (R1), prompting is considered by

displaying GPS functionality.

Towards an Approach for Incorporating Usability Requirements into Context-Aware Environments

471

Figure 6: Model-driven transformation process.

Figure 7: Consistent usability rules: (a) XML file, (b)

converted XML-based file.

Figure 8: The generated UI for contexts C1 and C2.

Figure 9: The generated UI for contexts C3 and C4.

ICEIS 2018 - 20th International Conference on Enterprise Information Systems

472

5 EXPERIMENTS AND RESULTS

In order to validate our proposal, we have evaluated

our work management mobile application. We

consider a set of experimental subjects. Thirty two

participants were invited in the experiment. Table 5

describes the different profiles that participated.

We consider a questionnaire to gather end-users’

opinions and perceptions. A twenty question (ISO,

2010; Bastien and Scapin, 1997) ergonomic criteria

based questionnaire is handed to the participants.

This questionnaire considers the following criteria

(adaptation, ease of use, efficiency, effectiveness,

and satisfaction) as presented in Table 6. The final

results are illustrated in Figure 10.

According to these results, we can conclude that

the majority of the participants are satisfied (87.5%).

This is due to the quite satisfaction of their

requirements with regard to the changing context of

use. Concerning the efficiency, almost 90% of the

users affirmed that they accomplished the work

rapidly. This is due to the easy access to the system

functionalities. Furthermore, 81% of the users

asserted the adaptive UI ease of use.

Despite of the participants’ heterogeneity and the

highly dynamic context of use, almost all users

asserted that the use of the adaptive application was

very effective in term of reaching targets and

responding to users’ preferences and requirements.

Table 5: Profiles of participants.

Profile

Age

Number

User experience

Profile 1

30-35

Novice

Profile 2

35-40

Expert

Profile 3

40-45

Novice

Profile 4

45-50

Expert

Table 6: Defined criteria questions.

Criteria

Question

Efficient

Can the user finish the work quickly

during the interaction with the

interface?

Effectiveness

Have you achieved what you intended

to do with the interface?

Adaptation

Are your requirements sufficiently

taken into account while exploiting

the interface?

Satisfaction

Working with the interface is

satisfying?

Ease of use

The interface manipulation is easy?

According to the users’ responses, we have

perceived that: (1) manipulating the adaptive system

is greatly efficient and the users have quickly

accomplished their objectives, (2) the users’

objectives were effectively, (3) the system adapted

to the users profiles in different ways, (4) the users

declared satisfaction for using the adaptive system

and, (5) the users can use the adaptive system easily.

As researchers, we concluded that the proposed

context-aware application ensures an optimal degree

of usability. Changing the context of use will allow

the variation of the system’s behaviour and

presentation. Our proposed adaptive system will

enable users to support activities, anticipate their

needs without assistance and experience with more

independence while working and communicating

too.

Figure 10: The system evaluation: a questionnaire.

6 CONCLUSION

This paper presented an approach that addresses

context-awareness and usability issue as parts of an

adaptive UI development process. The main

motivation of the contribution is to combine these

issues in a conformity process to provide a more

reliable design of UI. The proposed methodology is

defined through two stages namely: the selection of

consistent usability rules and the model-driven

development. Firstly, our proposal is based on an

inference system for deducing consistent usability

rules with regard to the context of use. Secondly, a

model-driven transformation process has been

fulfilled in order to incorporate both context-

awareness and usability requirements into the

Model-Driven Architecture.

With regard to the existing proposals, the

usability-based context-aware concept initiated in

this paper presents the following benefits: (1)

usability requirements are consistently selected in

context-aware environments; (2) usability of

context-aware systems is processed since an early

design stage; and (3) experiments with end-users

have been carried out.

The continuity of our research work leads

directly to the investigation of more context-based

usability rules with the perspective of the users in

Towards an Approach for Incorporating Usability Requirements into Context-Aware Environments

473

order to depict other requirements and to continue

the evaluation of our approach and incorporate

findings into future enhancements.

ACKNOWLEDGEMENTS

This project has been developed in the scope of a

MOBIDOC doctoral thesis of the PASRI program

financed by the European Union and administered

by the ANPR.

REFERENCES

Abrahao, S. M., and Insfran, E., 2006. Early usability

evaluation in model driven architecture environments.

In QSIC, pages 287–294.

Ali, M.F. Perez-Quinones, M.A., Abrams, M., and Shell,

E. 2002. Building multiplatform user interfaces with

uiml, In Proceedings of Computer Aided Design of

User Interfaces, pages 255-266.

Aquino, N., Vanderdonckt, J., Condori-Fernendez, N.,

Tubio, O. D., and Pastor, O., 2010. Usability

evaluation of multi-device/platform user interfaces

generated by model-driven engineering, ESEM

ATLAS group LINA & INRIA, ATL, Atlas

Transformation Language, 2006. ATL User Manual -

version 0.7

Ben Ammar, L., Trabelsi, A., and Mahfoudhi, A., 2015.

Incorporating usability requirements into model

transformation technologies, Requirement Engineer-

ing, 20(), pages 465-479.

Ben ayed, E., 2017. Une approche pour l'évaluation des

systèmes d'aide à la décision mobiles basés sur le

processus d'extraction des connaissances à partir des

données : Application dans le domaine médical. PhD

thesis, National school of Engineers of Sfax and

University of Valenciennes and Hainaut-Cambrésis.

Brossard, A. Abed, M., and Kolski, C., 2007.

Modélisation conceptuelle des IHM : Une approche

globale s'appuyant sur les processus métier, Ingénierie

des Systèmes d'Information (ISI) – Networking and

Information Systems, vol 12, pages 69-108.

Dey, A.D., and Abowd, G.D., 2000. Towards a Better

Understanding of Context and Context-Awareness,

CHI Workshop on the What, Who, Where, When, and

How of Context-Awareness.

Eclipse Acceleo, https://eclipse.org/acceleo/, last visited

on 30 July 2017

Gonzalez-Huerta, J. Blanes, D. Insfran, E., and Abrahمo,

S., 2010. Towards an Architecture for Ensuring

Product Quality in Model-Driven Software Develop-

ment, PROFES

Harrison, R., Flood, D., and Duce, D., 2013. Usability of

mobile applications: literature review and rationale for

a new usability model, Journal of Interaction Science.

Hentati, M., Ben Ammar, L., Trabelsi, A., and Mahfoudhi

A., 2016 An Approach for Incorporating the Usability

Optimization Process into the Model Transformation,

ISDA.

ISO: ISO 9241-210, 2010, Ergonomics of human-system

interaction

Jumisko-Pyykko, S., and Vainio, T., 2010. Framing

the Context of Use for Mobile HCI, International

Journal of Mobile Human Computer Interaction, 2(4),

pages 1-28

Oliveira, K.M., Bacha, F., Mnasser, H., and Abed, M.,

2013. Transportation ontology definition and

application for the content personalization of user

interfaces, Expert Systems with Applications, 40(8),

pages 3145-3159.

OMG, BPMN – Business Process Modeling Notation

Specification version 1.0, 2006. OMG Available

Specification.

OMG, MDA Guide Version 1.0, 2003, http://omg.org/

mda/mda_files/MDA_Guide_Version1-0.pdf

Ormeno, Y. I. ,Panach, J. I., Condori-Fernandez, N., and

Pastor, O., 2013. Towards a proposal to capture

usability requirements through guidelines, RCIS.

Panach, I. J., Aquino, N., Pastor, O., 2014.

A proposal for modelling usability in a holistic MDD

method. Sci. Comput. Program. 86, pages 74-88

Paterno, F., Carmen, S., and Lucio, D. S., 2009. MARIA:

A universal, declarative, multiple abstraction-level

language for service-oriented applications in

ubiquitous environments. TOCHI. 16(4),

Pathan, K.T., and Reiff-Marganiec, S. 2009. Towards

Activity Context using Software Sensors, YR-SOC,

pages 27-35

Scapin, D. L., and Bastien, J. M. C., 1997. Ergonomic

criteria for evaluating the ergonomic quality of

interactive systems, Behaviour & Information

Technology, 16(4), pages 220-231.

Seffah, A., Donyaee, M., Kline, R. B., and Padda, H. K.

2006. Usability measurement and metrics: A

consolidated model, Software Quality Control, 14(2),

pages 159–178.

Serral, E., Valderas, P., and Pelechano, V., 2010. Towards

the model driven development of context-aware

pervasive systems, Pervasive and Mobile Computing,

pages 254-280.

Sottet, J. S. 2008. Mega-IHM: malléabilité des Interfaces

Homme Machine dirigées par les modèles, PhD

Thesis, University of Joseph Fourier.

User Interface Markup Language (UIML), Version 4.0.

Oasis, 2008, http://docs.oasisopen.org/uiml/v4.0/cd01/

uiml4.0- cd01.html

Zaibi, D., Riahi, M., and Moussa, F., 2016. Formalization

of ergonomic knowledge for designing context-aware

human-computer interfaces, ICSEA

Zhang, D. and Adipat, B., 2005. Challenges, methodolo-

gies, and issues in the usability testing of mobile

applications, International Journal of Human-

Computer Interaction, 18(3), pages 293-308

ICEIS 2018 - 20th International Conference on Enterprise Information Systems

474