Model-driven Engineering for Optimizing the Usability of User

Interfaces

Marwa Hentati

, Lassaad Ben Ammar

, Abdelwaheb Trabelsi

and Adel Mahfoudhi

University of Sfax, National School of Engineering, CES Laboratory, Sfax, Tunisia

University of Sfax, Sfax, Tunisia

College of Computer and Information Technology, Taif University, Taif, Saudi Arabia

Keywords: MDE, User Interface, Usability Optimization, Transformation Process.

Abstract: Usability is considered to be one of the most important quality factors that determine the success or failure

of an interactive system. This can be explained by the ever-increasing number of studies addressing the

integration of the usability evaluation at the development process. However, most of these proposals aim to

guide the user interface transformation process according to a set of usability criteria allowing the

generation of user interfaces which meet the usability requirement. In fact, the selection of the adequate

alternative transformation depends on the usability attributes that will be met. This paper proposed an

approach that considers the user interface generation process as a usability optimization problem according

to given usability optimization objectives. The aims to generate all possible concrete user interfaces from a

given abstract user interface. Then, the usability optimization process selects the optimal concrete user

interface for a specific context of use.

1 INTRODUCTION

The model-driven engineering (MDE) is currently

being adopted in the Human-computer Interaction

(HCI) field to support user interface specification and

engineering activities (Hussmann et al., 2011). The

MDE paradigm is proved to be quite appropriate

(Favre, 2004). This approach specifies an automated

process of developing user interfaces through the

definition of models and their transformation from

high-level models to code generation in the target

platform (OMG, 2003). A renowned approach in this

context is the Cameleon Reference framework

(Calvary et al., 2003). It provides a unifying reference

framework that structures the user interface

development taking the context of use (user, platform

and environment) into account.

In this framework, focus is generally placed on

data and functional modelling, disregarding usability

aspects. Therefore, there is a need to extend the

MDE process in order to promote usability as a first

class entity in the development process.

The present paper aims to delineate an approach

that addresses these weaknesses by extending the

user interface process in order to optimize the user

interface usability. The proposed approach is

intended to optimize the usability from the

conceptual model in a user interface (UI). It should

be noted that the conceptual model covers the

abstract user interface (AUI) model and the concrete

user interface (CUI) models. Therefore, the selection

of the adequate alternative transformation depends

on the usability attributes, which are able to convey.

The proposition is structured in three main

contributions: (1) generating all the possible

concrete user interfaces from a given abstract user

interface, (2) formulating the usability optimization

function for a given context of use and (3) selecting

the alternative transformation able to generate the

optimal usability user interface.

We structure the remainder of this paper as follows:

section 2 presents some related studies. Section 3

describes the different stages of the proposed approach.

In order to show the usefulness of our proposal, a case

study is presented in Section 4. Finally, section 5

presents the conclusion and the future research work.

2 RELATED WORKS

There are currently several research studies that have

dealt with usability in MDE environment have been

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

Model-driven Engineering for Optimizing the Usability of User Interfaces.

In Proceedings of the 18th International Conference on Enterprise Information Systems (ICEIS 2016) - Volume 2, pages 459-466

ISBN: 978-989-758-187-8

459

proposed. The main objective of these methods is to

propose a set of usability attributes in order to drive

the selection of adequate alternative transformations.

In (Panach et al., 2013), the authors have

addressed the usability features related to the system

functionality, which may involve important changes

in the system architecture. They use the term FUFs

(Functional Usability Features) to indicate this type

of usability features. The FUFs are abstractly

represented by means of conceptual primitives that

will extend the conceptual model of an existing

model-driven development (MDD) method. Then,

the conceptual model can be seen as the input of a

model compiler that can generate the software

application automatically (or semi-automatically). In

(Panach et al., 2014) a set of the well-known FUFs

in the human–computer interaction (HCI)

community are gathered in a usability model that is

included from an early stage of a holistic MDD

method.

In (Huerta et al., 2010), the authors have

proposed an architecture to perform the quality of

the model-driven transformation. The main goal of

this architecture is to deﬁne a set of artifacts and a

process for specifying and executing model

transformations. The selection of the alternative

transformations is guided by quality attributes. The

feasibility of the proposed architecture is shown

using a case study which transforms a requirements

model into a UML class model.

(Ammar et al., 2014) have suggested an approach

to integrate usability issues as a part of a user

interface development process from the conceptual

models. The proposed approach is structured in two

main stages: (1) the deﬁnition of the model

transformation and (2) the execution of a

parameterized transformation. Concerning the ﬁrst

stage, it establishes a set of transformation rules

describing all the possible alternative transformations

for a given domain. As for the second one, it executes

a model driven transformation parameterized by the

usability model in order to select the adequate

alternative transformation.

Although the previously mentioned works are

useful in the interplay between the usability and

MDD paradigm, they lack precise details about the

quality of the alternative selection process. In fact,

all of them try to select the alternative

transformation without addressing the problem of

the choice optimization. This problem is the object

of some initiatives in the research field among which

we quote (Gajos et al., 2010), (Petter et al., 2008)

and (Raneburger et al., 2011).

In (Gajos et al., 2010), the authors implemented

the SUPPLE system that automatically generates

multi-target UI using the user interface specification

as input. The user interface adaption is treated as an

optimization problem related to a user and device

specific cost function. Therefore, the SUPPLE

system renders each interface element from the

functional interface specification into an appropriate

concrete widget according to a matching function.

To find the optimal one, SUPPLE relies on a cost

function that provides a quantitative metric, such as

the speed of use.

In the same context, (Petter et al., 2008)

proposed to optimize the usability of the graphical

user interface (GUI). They suggest an extension to

the QVT standard in order to consider the usability

aspects. The main goal is to transform each element

from the user interface specification model to an

optimal (GUI) component. The optimization

function has been formulated on the basis of the

manipulation and navigation time.

Face to the increasing use of small screen

devices, (Raneburger et al., 2011) suggested an

approach for extending and incrementing the

Discourse-based communication models

transformation process. This approach introduces

straight-forward optimization techniques into the

model-driven generation of GUIs to reduce some

usability problems. This allows the optimization of

the generated GUIs for devices with small screens in

such a way as to utilize the given space and to

minimize navigation and scrolling.

The analysis of the aforementioned studies

allows us to detect some limitations related to the

usability optimization problem. In fact, a limited

usability aspect may not ensure the usability of the

final application. It is within this context that our

research work lies to introduce a framework

addressing the problem of usability optimization in

an MDE-compliant method.

3 USER INTERFACE

GENERATION AS USABILITY

OPTIMIZATION PROBLEM

3.1 Overview

Since its first beginning HCI has been concerned

with the evaluation of interactive systems, and

consideration of usability. Recently, the MDE has

been adopted in the HCI domain to support

interaction systems specification and engineering

ICEIS 2016 - 18th International Conference on Enterprise Information Systems

460

activities (Hussmann et al. 2009). In this context, the

approach is completely based on the Cameleon

framework which presents a unifying framework for

the development of multi-target user interfaces

supporting multi-context of use (Calvary et al.,

2003). In the development process, the unifying

framework structures the user interface development

process into four levels of abstraction, starting from

task specification to a running interface: The Task

and Concepts (T&C) level, the Abstract User

Interface (AUI) level, the Concrete User Interface

(CUI) level and the Final User Interface (FUI) level.

It should be noted that the proposed approach is

intended to improve the usability at the conceptual

model which covers the AUI model and the CUI

model. In fact, the conceptual model represents an

abstraction of the user interface. The proposed

approach is made up of three main stages:

 Generating all the possible CUI from a given

AUI,

 Optimizing the usability at the conceptual

model,

 Selecting the alternative transformation able to

generate a user interface with an optimal

usability.

In the first stage, all the possible generation of

CUI are established from a given AUI. The second

stage performs a usability optimization process

taking as input each CUI established in stage one. A

profit usability function is formulated as an

optimization function according to a set of usability

attributes. With each one of them is associated at

least one metric which provides a numerical value

interpreted by a mechanism of indicators. In the

third stage, the user interface with the optimal value

of usability will be selected as a result of the

optimization process in a particular context of use.

Figure 1 shows an overview of the proposed

approach.

Figure 1: The proposed approach to optimize the usability

of UI with model-driven development process.

Each stage of this approach will be described in the

following sections.

3.2 The Concrete User Interface

Generation

Model

transformation is a key mechanism when

building software systems with MDE approach. The

model transformation expresses how one or more

source models are transformed into one or more target

models. It consists of a set of transformation rules,

each of which describes how a construct of the source

model can be transformed into another in the target

model. The elaboration of the transformation rules

identifies the corresponding construct in the CUI

model for each construct from the AUI model. There

may be more than one construct in the CUI model that

can be associated with the one from the AUI model.

Transformation associated with an abstract

component are equivalent. However, they may differ

from the non-functional perspective and do not

satisfy the same usability attributes. For example, in

Figure 2, the first solution displays the ranges of the

accepted values. Hence, the user is well guided to

insert the expected doors number, and therefore, the

Prompting attribute is well fulfilled. In the second

and third solutions, the user is generally intended to

select from a set of possibilities, and thus the Error

Prevention attribute is well fulfilled. By considering

this example, we can notice that each transformation

optimizes at least one usability property.

In general, we can calculate the number of CUIs

(CUI

num

) that can be generated for a given AUI as:

CUI

num=

∏

_







(1)

where n is the number of abstract components that

are matched in the AUI and the

transformation_number is the number of all possible

transformations associated with an abstract

component i.

Figure 2: Example of different model transformations for

the input data element.

Task and Concepts model

(T&C)

Abstract User Interface

(AUI)

Optimal Concrete User

Interface (OCUI)

Final User Interface (FUI)

CUI Generation

CUI

Selecting optimal

CUI

Cameleon Framework

Artifact

Activity

Data flow

Usability measurement

Contexte of user

Usability Model

Select Metrics

Select Attributes

Grouping Function

Establish Rating

Levels

Usability optimization process

Model-driven Engineering for Optimizing the Usability of User Interfaces

461

3.3 Usability Measurement

Usability is a difficult concept to quantify. It

involves several dimensions, mainly the process and

the product dimensions. In this paper, we will focus

on the product dimension since our objective is to

evaluate usability at an early stage. Therefore, we

opt to use a usability model which extends the one

presented in the ISO/IEC 9126-1 standard (ISO,

2001). This view is more deal with the

characteristics of the product itself and can be used

to evaluate the intermediate artifacts produced in an

MDE development process.

A usability optimization approach is

implemented to optimize the usability of the targeted

user interfaces against the defined usability

attributes by means of usability metrics carried out

at the conceptual model. To do so, we propose a

usability optimization function that was formulated

according to a set of usability attributes through the

usability metrics. Consequently, this stage involves

four main steps: (1) formulating a profit function, (2)

selecting metrics, (3) establishing rating levels, and

(4) grouping function. The last three steps are

adapted from the evaluation method proposed by

(Ammar et al., 2015).

3.3.1 Formulation of the Profit Function

This step aims to formulate a profit function which

plays the role of an optimization objective function.

The profit function is intended to select the optimal

CUI generated in the first stage (section 3.2) of the

proposed approach.

Our optimization objectives reflect which aspect

to be considered in our approach and in which order

we consider their importance. In this context,

believe that

the usability optimization can be

improved by means of optimizing the performance

and the appearance of the user interface. Hence, we

formulated our usability optimization problem

through two optimization objectives which are:

 Maximizing the user performance P( ),

 Maximizing the user interface appearance

A().

We start by defining the CUIprofit of a specific

combination of transformation (t

) to be of the form:

CUIprofit (t

)= P(t

)+A(t

) (2)

where P(t

) represents the performance value of a

user interface generated with a specific

transformation t

and A(t

) represents the appearance

value of a user interface generated with a specific

transformation t

Since our objective is to optimize the usability

before the implementation of the user interface, we

believe that the user performance is optimized when

the user is well informed, guided and prevented from

making mistakes during the interaction with the final

user interface. In addition, optimizing the user

performance needs to minimize the number of steps

required to accomplish a task and an accepted level

of the control action which includes cancellability,

undoability and explicit user actions. Indeed, we

define that the user performance value of a given

CUI which represents objective 1 as the following

formula:

P(t

)= W

* BR(t

)+ W

ERP

* ERP(t

)

+ W

PR(t

)+ W

UCA

* UCA(t

)

(3)

where BR(t

) represents the brevity value, ERP(t

) is

the error prevention value, PR(t

) is the prompting

value, UCA(t

) is the user control action value and

, W

ERP

, W

ERP

represent the additional

weight values.

While the user performance is of high

importance, the user interface appearance is an

equally essential factor for determining the user

interface usability that gives the user a comfortable

feeling during the interaction with the user interface

(Marcus, 2011). Therefore, the usability is enhanced

by improving the appearance of the user interface

(or synonyms and related concepts, such as

“aesthetics”, “attractiveness”, “beauty” or “form”).

(Ngo-2000).

We believe that obtaining an optimized user

interface appearance needs an appropriate number of

elements per window by keeping a good equilibrium

between information density and white space in a

specific computing-device. Additionally, the best

appearance can be ensured by the consistency of the

user interface elements and the presence of an

acceptable number of navigation elements. Thus, the

appearance A () value of a specific CUI which

represents objective 2 is calculated by the following

formula:

A(ti)=W

*ID(t

)+W

*CN(t

)+W

NAv

*NAV(t

) (4)

where ID(t

) represents the information density

value, CN(t

) is the consistency value, NAV(t

) is

the navigability value and W

NAv re

present

the additional weight values.

It should be noted that each additional weight

(i.e., W

, W

ERP

, W

UCA

NAv

) is

used to influence the impact of a specific addend on

the overall usability profit value. The importance of

the usability attributes varies according to systems

and target population characteristics (e.g., small

ICEIS 2016 - 18th International Conference on Enterprise Information Systems

462

screen device, user experience). This work was

described with details in (Hentati et al., 2015).

3.3.2 Selecting Metrics

We need to associate with each usability attribute at

least one metric. In what follows, we illustrate for

each usability attribute the opted metric and its

generic description:

Brevity (BR): The user interface should allow

users to interact with the UI element in a few

number of action steps. The brevity can be

quantified by the number of steps (counted in

keystrokes) required to accomplish a task.

BR = distance(x,y) (5)

where distance(x,y) represents the distance between

source screen (x) and the target screen (y).

Error Prevention (ERP): can be quantified by

the percentage of the list primitive used instead of

text field when the input element has a set of

enumerated values:

ERP = list(x)/n (6)

where list(x) represents the number of the list

primitive and n is the number of input data elements

with limited possible values.

Prompting (PR): the prompting can be

measured in terms of the proportion of label that

displays supplementary information.

PR = StaticField()/n

(7)

where StaticField() represents the number of labels

which display the supplementary information and n

is the total number of static field (label).

User Control Actions (UCA): We propose to

quantify the user control actions according to the

degree of control assigned by the system which

includes cancellability, undoability and explicit user

actions.

UCA = Σxi /n (8)

where 

 represents the action elements and n

represents the total number of elements.

Information Density (ID): The information

density of a user interface can be measured in terms

of the number of elements per user interface.

ID = 

(9)

where n represents the total number of UI elements

per interface.

Consistency (CN): This factor includes the

element consistency attribute that can be measured

by means of the percentage of the repeated elements

and the label consistency which can be quantified by

the means of the percentage of the repeated label at

the same user interface.

CN= r (10)

where r represents the proportion of repeated elements

with the same label.

Navigability (NV): The navigability measures

the level of facilities that the system will provide to

navigate throughout several interfaces. We propose

the average of navigation elements per UI as an

internal metric to measure the navigability attribute.

NV= n (11)

where n represents the number of navigation elements.

3.3.3 Establishing Rating Levels

The previously defined metrics provide a numerical

value that necessitates having a meaning in order to

be interpreted. The mechanism of indicator is

restored in order to reach such goal. It consists in the

attribution of qualitative values to each numerical

one. In (Ammar and Mahfoudhi, 2013), the

qualitative values was summarized in: very good

(VG), good (G), medium (M), bad (B), and very bad

(VB). According to (Panach et al., 2014), these

indicators do not differentiate correctly between the

values (VG) and (G), and the values B and VB. In

our approach, we propose to attribute three

indicators (B), (M) and (G). For each qualitative

value, a numerical range was assigned. For example,

the prompting value (PR) may be considered as

good (G) when at least 95% of input element labels

should display additional information (e.g. the

required data format). Table 1 shows the list of

indicators that have been defined.

Table 1: Proposed indicators for usability metrics.

Metric G M B

PR PR≥0.95 0.75≤PR<0.95 PR<0.75

BR BR≤2 2<BR≤5 BR >5

ERP ERP≥0.9 0.7≤ERP<0.9 ERP<0.7

UCA UCA≥0.9 0.5≤UCA<0.9 UCA<0.5

ID ID<15 15≤ID<25 ID≥25

CN CN≥0.85 0.6≤CN<0.85 CN<0.6

NV NV<7 7≤NV<12 NV≥12

3.3.4 Grouping Function

The aim of this stage is to put metrics and attributes

together in order to obtain a single usability profit

measure. While executing the usability evaluation,

the previously defined metrics are applied. The

obtained numerical values are converted into their

corresponding qualitative ones. Next, each

categorical value is converted to numerical values

Model-driven Engineering for Optimizing the Usability of User Interfaces

463

with respect to the following hypothesis. (G)

⇒3,

(M) ⇒2, (B) ⇒1. The final result is calculated by a

sum function.

The last step in the grouping function is to

convert the obtained numerical value into an ordinal

value. We assign B to the value between 1 and 2, M

to the value between 2.1 and G to value lower or

equal to 3. The three steps are applied for each

grouping. Groupings are performed bottom-up from

indicators until the overall user interface usability is

reached.

3.4 Selecting the Optimal CUI

The entire CUI generated in stage one has a final

usability evaluation value. The higher one is selected

as the optimal CUI. It should be noted that the

present stage is actually performed (semi)-

automatically. The selecting stage will be an

exhaustive task with the increasing number of CUI

possible combinations. That is why we hope to

automate this stage by implementing it in the future.

4 AN ILLUSTRATIVE CASE

STUDY

This section investigates a case study in order to

illustrate the feasibility of our proposal. The research

question addressed by this case study is: Does the

proposal ensure the usability optimization of the

generated user interface artifact?

The object of the experimentation is Rent-a-car

system. The scenario is adapted from (Bouchelligua

et al., 2010). The system will run on terminals of

customers (laptop, PDA, mobile phone, etc.).

Therefore, the user interface must be able to bring

not only a feeling of comfort when interacting with a

best appearance user interface but also a best

performance regardless the context of use.

Since the Rent-a-car system was large, we focused

our interests on the generation of the CUI for “the

renting period”, “customer personal information”

and “car preferences” tasks. The transformation of

the AUI allowed the generation of 216 CUIs. It

covered all the possible combinations between the

possible concrete components. It should be noted

that the number of all possible combinations was

obtained from a limited number of tasks (3 tasks and

8 sub-tasks), thus selecting the best alternative

among them is an exhaustive task which can be

solved by our approach according to the usability

profit function.

We suppose to have the abstract user interface

from Figure 3 as a result of the transformation of the

three tasks «the renting period”, «customer personal

information» and «car preferences» following the

model transformation explained in details in

(Bouchelligua et al.,2010). The result of the

transformation is an XML file which is in

accordance with the AUI metamodel (left part of

Figure 3). An editor with the Graphical Modelling

Framework (GMF) of eclipse was developed in

order to better clear up the user interface layout. The

sketch of the user interface presented by the editor is

shown in the right part of Figure 3.

The AUI contains a UIGroup called «Renting

Cars» which gives access to three UIUnitSuit called

«Renting Period», «Customer Personal information»

and «Car Preferences». The «Renting Period»

container gives access to the collection date and the

return date. The customer should specify his/her

name, surname and age whose acceptable value

should be more than 18. After that, the customer

should specify the car preferences as the doors

number, fuel and color.

To better explain our proposal,

we started the

case study by

a CUI model transformation illustrated

in Figure 4. The generated CUI did not take into

account our proposed usability optimization in a

specific context of use. The context is the following:

a laptop as an interactive device (normal screen

size), a customer with a low level of experience of

using the Rent-a-car system. Although the deducted

CUI filled their functional objectives, it did not

satisfy our proposed usability optimization profit

functions.

Figure 3: The abstract user interface model for Renting-a-

car system.

Each input element generated CUI illustrated in

Figure 4 was transformed into UIFieldEdit. Since

ICEIS 2016 - 18th International Conference on Enterprise Information Systems

464

novice users should be protected against errors while

entering data, input elements with limited values (eg.

Doors number, Fuel and Car color) should be displayed

with list box elements such as dropdown list and radio

button. Moreover, novice users need to be well

informed and guided by the presence of supplementary

information displayed in the labels UI StaticField in

order to ensure the best user performance.

Figure 4: Non-optimal concrete user interface for a normal

screen size.

The execution of a proposed approach, wherein

the usability optimization function is conveyed lead

to the generation of the user interface which contains

concrete elements that fulfil the desired usability

attributes. Figure 5 illustrates such transformation

taking into account the computing device and the

target population characteristics. In this example, the

target population is a customer with a low level of

experience of using the Rent-a-car system a laptop

as an interactive device (normal screen size) (15’’).

In this case, the optimized CUI illustrated in Figure

5 ensure the best performance by using list box

elements and by the presence of the supplementary

Figure 5: An optimal concrete user interface for normal

screen size.

information like the mask of date (DD/MM/YYYY),

the range of accepted values for the age value (>18

years) and the doors number (3 or 5). Hence, the

novice users were well informed to insert the correct

value. Furthermore, the UCA attribute is respected

by the presence of a Cancel and Validate buttons.

Figure 6: An optimal concrete user interface for a small

screen size.

The second transformation to be conducted takes

an «iPAQ Hx 2490 Pocket PC» as platform and a

customer with a low level of experience of using the

Rent-a-car system as a target population. The

migration to such platform raises a new

redistribution of the user interface elements. The

small screen size (240x320) is not sufficient to

display all information.

In Figure 6, the CUI with an optimal usability for

small screen group a limited number of concrete

components, whose maximum number of concepts

can be manipulated (5 in the case of «iPAQHx2490

Pocket PC»). To ensure the best user interface

appearance, the user interface elements are

redistributed on several windows. The redistribution

of interface elements on several windows will bring

more steps to reach the goal by the aims of

navigability elements (eg. Next, Return).

5 CONCLUSIONS

This paper presents an approach for optimizing

usability aspects as a part of a user interface

development process.

It can be used in any software

development proposal based on conceptual models.

The proposed approach extends the Cameleon

reference framework by integrating the usability

optimization process from the conceptual models.

This proposition gathers three main stages: the first

one aims to generate all possible concrete user

interfaces from a given abstract user interface. Then,

Model-driven Engineering for Optimizing the Usability of User Interfaces

465

a usability optimization process was performed. We

used a profit function by a set of usability attributes,

each of which was associated at least with a metric.

These metrics will be interpreted by defining a

mechanism of indicator taking into account the

target population and the computing device

characteristics. Finally, the usability results allowed

the selection of the optimal alternative

transformation.

Several research studies can be considered as a

continuation of this work. In fact, the present

approach can be defined as a linear optimization,

while respecting a set of optimization constants. An

empirical validation of the optimized user interface

is recommended to clearly demonstrate the

coherence between values obtained by our proposal

and those perceived by end-user.

REFERENCES

Aquino, N., Vanderdonckt, J., Fernandez, N.C., Dieste, O.,

Pastor O., (2010) “ Usability evaluation of multi-

device/platform user interfaces generated by model-

driven engineering”, In: Proceedings of the 2010

ACM-IEEE international symposium on empirical

software engineering and measurement ESEM ‘10, pp.

30: 1-30. DOI 10.1145/1852786.1852826.

Ammar, L. B., Trabelsi, A. and Mahfoudhi, A.,(2014)

“Incorporating usability requirements into model

transformation Technologies”, Requirements

Engineering: 1-15, 2014.

Ammar, L.B., & Mahfoudhi, A. (2013). Early usability

evaluation in model driven framework. In ICEIS 2013:

Proceedings of the 15th international conference on

enterprise information systems, Volume 3 (pp. 23–30).

France: Angers, 4–7.

Ammar, L. B., Trabelsi, A., & Mahfoudhi, A. (2015). A

model-driven approach for usability engineering of

interactive systems. Software Quality Journal, 1-35.

Bouchelligua,W., Mahfoudhi,A., Mezhoudi,N., Dâassi, O.,

and Abed, M. (2010). User interfaces modelling of

workflow information systems. In EOMAS, pages

143–163.

Calvary, G., Coutaz, J., Thevenin, D. (2003) “A unifying

reference framework for multi-target user interfaces”,

Interacting with Computers, vol.15, no 3, p. 289-308.

Favre,J.M. (2004). Toward a Basic Theory to Model

Driven Engineering.

Gajos, K. Z., Weld, D. S., and Wobbrock, J. O. (2010)

“Automatically generating personalized user interfaces

with Supple. Artificial Intelligence”, 174(12), 910-

950.

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

A. (2015). Towards optimizing the usability of user

interface generated with model-driven development

process. In Human System Interactions (HSI), 2015

8th International Conference on (pp. 206-212). IEEE.

Huerta, J.G., Blanes, D., Insfran E., and Abrahão,

S.,(2010) “Towards an architecture for ensuring

product quality in model-driven software development

”, Proceedings of the 11th international conference on

product focused software (PROFES ‘10). ACM,New

York, NY, USA, pp 28–31.

Hussmann, H., Meixner, G., & Zuehlke, D. (Eds.). (2011).

Model-driven development of advanced user

interfaces, studies in computational intelligence (Vol.

340). Berlin: Springer.

ISO/IEC: ISO/IEC 9126 (2001). Software engineering:

Product quality. ISO/IEC.

Marcus, A. (2011). Design, User Experience, and

Usability. Theory, Methods, Tools and Practice: First

International Conference, DUXU 2011, Held as Part

of HCI International 2011, Orlando, FL, USA, July 9-

14, 2011, Proceedings (Vol. 6770). Springer.

Ngo, D. C. L., Samsudin, A., & Abdullah, R. (2000).

Aesthetic measures for assessing graphic screens. J.

Inf. Sci. Eng, 16(1), 97-116.

OMG (2003) MDA guide version 1.0.1.

http://www.omg.org.

Panach J.I., Aquino N. and Pastor O.,(2014) “A proposal

for modelling usability in a holistic mdd method” Sci.

Comput. Program. 86, pp 74-88.

Panach, J.I., Juzgado, N.J.and Pastor, O., (2013)

“Including functional usability features in a model-

driven development metho”, Comput Sci Inf

Syst,10(3):999–1024.

Petter, A., Behring, A., and Zlatkov, M.,(2008) “Modeling

usability in model transformations”, In : Proceedings

of the 1st International Workshop on Non-functional

System Properties in Domain Specific Modeling

Languages, NFPinDSML. p. 1613-0073.

Raneburger, D., Popp, R., Kavaldjian, S., Kaindl, H., and

Falb, J., (2011) “Optimized GUI generation for small

screens”,Model-Driven Development of Advanced

User Interfaces, volume 340 of Studies in

Computational Intelligence, Springer Berlin

Heidelberg, pp. 107-122.

ICEIS 2016 - 18th International Conference on Enterprise Information Systems

466