You are on page 1of 11

Composite, Real-time Validation for Business Process

Modelling
Martin Smuts Clayton Burger Brenda Scholtz
Department of Computing Sciences Department of Computing Sciences Department of Computing Sciences
Nelson Mandela Metropolitan Nelson Mandela Metropolitan Nelson Mandela Metropolitan
University University University
Port Elizabeth, South Africa Port Elizabeth, South Africa Port Elizabeth, South Africa
Martin.Smuts@nmmu.ac.za Clayton.Burger2@nmmu.ac.za Brenda.Scholtz@nmmu.ac.za

ABSTRACT
Graduates with Business Process Management (BPM)
1. INTRODUCTION
Forecasts for the Business Process Management (BPM) market
competencies are highly sought after and Business Process
predict a growth from $2.8 billion in 2013 to $8.3 billion in
(BP) modelling skills form an integral part of the BPM skillset.
2019 [1]. This growth has led to a need for Business Process
One challenge facing modellers is that only a few existing BP
(BP) modelling skills which have become essential in various
modelling languages have intuitive validation techniques
fields, ranging from Information Systems (IS) development,
which check for syntactic issues or rule violations. Not having
Software Engineering and BPM practices [2, 3]. BP modelling
an intuitive validation technique to support the effective and
is used in organisations before they implement new systems or
accurate creation of BP models can affect the success of BP
make crucial decisions, and can assist with understanding
modelling projects and have a negative impact on the
workflows and identifying manual, repetitive and system
operations of the business. In this paper a composite, real-time
automated tasks within BPs [4, 5]. However, the objective of
validation technique is proposed which provides both graphical
BP modelling is not simply to configure manual or repetitive
and textual feedback. The Event-Driven Process Chain (EPC)
tasks, but to create a common understanding of the interactions
modelling notation is used to demonstrate the proposed
and dependencies among: people, organisational operations,
validation technique. Investigating the effectiveness and
the systems they use, and the information required to do their
usability of the approach when implemented in a Microsoft
tasks effectively [5]. Furthermore, BP modelling is used to
Visio environment reveals that a composite, real-time
identify process improvements and to demonstrate compliance
validation technique reduces the volume of issues that typically
with external regulations [6]. A BP modeller typically creates a
are made during the modelling activity and enables modellers
BP model, which consists of a network of BP diagrams that
to distinguish between the severities of the different types of
graphically depict descriptions and sequence of at least the
issues. The validation technique can also assist modellers in
activities, events/states and control flows that constitute the BP
learning the rules and syntax of a modelling language.
process [7]. BP modelling is an activity that is complex and
time-consuming as the core activities need to be arranged in
Categories and Subject Descriptors their execution order and exceptions dealt with if and when
K.3.2 [Computers and Education]: Computers and Information they occur [8]. Although various languages for BP modelling
Science Education exist, each language has its own unique notation and syntax
rules which vary in complexity and provide challenges during
General Terms training and education, which can often be attributed to the
Business Process Modelling, Business Process Modelling semantic and syntactical complexity [7, 9]. Notation typically
Languages, Validation. refers to a set of symbols or elements that are used to
graphically depict a process model as a visual model, whereas,
syntax provides a set of rules or constructs [7, 10].
Keywords
Event Driven Process Chain (EPC), Business Process Seethamraju [11] argues that there is a global need for BP
Modelling, validation, Business Process Management, education and that graduates have inadequate BPM
experiential learning and education. capabilities. As a result several Higher Educational Institutions
(HEIs) have introduced BPM courses in their IS programs [3,
11, 12]. The heavy demand for skilled BP modellers poses
several challenges for HEIs offering BPM courses [2]. These
HEIs often struggle to provide sufficient software and training
Permission to make digital or hard copies of all or part of this work for resources to support graduates during the course of their study
personal or classroom use is granted without fee provided that copies are [12]. The complexity and learning curve of a particular
not made or distributed for profit or commercial advantage and that notation and modelling tool requires a significant amount of
copies bear this notice and the full citation on the first page. Copyrights time and money for teaching modellers to perfect their BP
for components of this work owned by others than ACM must be honored. models [2, 12]. Students of BP modelling courses often
Abstracting with credit is permitted. To copy otherwise, or republish, to struggle with the complexity of modelling languages and the
post on servers or to redistribute to lists, requires prior specific associated rules, which highlights the need to investigate
permission and/or a fee. Request permissions from
Permissions@acm.org.
effective methods for teaching these. Furthermore, there exists
little insight into the way modellers interact with a BP
SAICSIT2014, September 29 - October 01 2014, Centurion, South Africa modelling tool, the pages they are using, what areas they get
Copyright 2014 ACM 978-1-4503-3246-0/14/09…$15.00 stuck on for longer than anticipated, or alternative ways that
http://dx.doi.org/10.1145/2664591.2664603 they find for addressing problems that repeatedly arise [5]. Not

93
addressing these issues will inevitably affect the performance 2.1 Business Process Management Lifecycle
and quality of the BPs. A BP consists of a set of tasks and activities within an
Individual stakeholders are not necessarily familiar with BP organisational structure that are linked with a logical flow
modelling [13]. Moreover, BPM tools heavily rely on experts using task dependencies and events. Each BP serves to produce
that model idealised “to-be” processes and do not help the a desired result which is valuable to both the organisation and
stakeholders to understand the “as-is” BP [9]. Therefore, it is its stakeholders [21]. BPM is a multi-disciplinary practice and
difficult for non-expert individuals to derive at a BP model that process-centric approach for improving organisational
expresses their specific systems requirements. Barjis [14] performance using a combination of methods, techniques and
reveals that an alarming amount of software projects fail due to software to assist organisations during the BPM lifecycle
poor BP modelling. The cost of fixing errors in the activities [5]. Moreover, Harmon [6] defines BPM as a
specification of a BP model is high and can have a significant management approach that typically focusses on process
impact on an organisation’s operations [15]. A reasonable optimisation, process analysis and operations management
explanation for this phenomenon is twofold. Firstly, a lack of using technology-enabled software that drives organisational
BP modelling skill is applied by the modeller. Secondly, agility and innovation. BPM software tools are often bundled
current BP modelling tools lack direct model checking features together as a complete software package known as a Business
or validation techniques [14]. Allowing novice modellers, such Process Management Suite (BPMS) that enables the BPM
as student modellers with virtually no practical modelling lifecycle activities for design, execution, monitoring and
experience, to create BP models without a sufficient validation controlling of BPs [22]. Within the scope of BPM, processes
technique increases the probability of making errors in their are modelled to bridge the gap between business audiences,
designs [16]. These issues are typically less evident in smaller technical audiences and technologies, thereby aligning the
models, but as the models expand and increase in complexity organisation’s resources according to its requirements and
modellers often lose track of the modelling notation’s syntax business rules [23]. BP modelling tools typically support
rules and interrelations due to their limited cognitive ability languages, such as, Business Process Modelling and Notation
[17]. The result is that modellers are often unaware of errors (BPMN), Unified Modelling Language (UML), Petri-net and
that exist in the created BP models as traditional modelling traditional flowcharts. Although BPMS do exist that provide
tools do not provide optimal support on issue feedback [10]. A functionality for each activity in the BPM lifecycle, smaller BP
real-time validation technique allows the modeller to detect modelling tools are typically used for the creation of BP
and fix issues during a modelling activity [18]. A validation models, which can be exported to external operational systems
technique should also provide feedback to the modeller by and execution engines. Popular BP modelling software tools
reporting the reason for the rule violation or issue [19]. are IDS Scheer ARIS Business Architect and Express,
ADONIS, IBM WebSphere Business Modeller and Microsoft
The main aim of this paper is to propose and evaluate a Visio [22, 24].
composite, real-time validation technique for BP modelling.
The evaluation will determine if the technique can assist Recker [15] describes BP modelling as an essential cornerstone
novice BP modellers with their knowledge of notational syntax of the initial capturing and documentation of existing BPs, as
rules, thereby improving the accuracy and quality of the well as the specification of the redesigned, automated or
produced BP models. This study therefore addresses gaps in amended BPs. The use of BP modelling is critical throughout
research related to BP validation techniques and tools. The the BPM lifecycle, as it forms the foundation of the
need for a validation technique has been motivated by various understanding and restructuring of process activities,
authors [7], [10, 20]. In an attempt to satisfy this need, information flows and roles of participating individuals and
literature provides evidence of a variety of validation applications for process optimisation [4, 7]. A lack of BP
techniques and tools being contributed. However, these tools modelling training has become a key issue in BP modelling
lack intuitiveness and are often applied after the BP model is [12]. Recker and Rosemann [3] suggest that these challenges
created, which has a great impact in terms of the BPM can be addressed by improving students’ technical knowledge
lifecycle (Section 2). The real-time validation technique is of BP modelling as well as the students’ actual modelling
applied to a popular modelling language, the Event-Driven skills. Students, therefore, require knowledge of the most
Process Chain (EPC) language (Section 3). The research popular modelling methods and practice with market leading
methodology followed for this study is a field study approach BP modelling tools.
whereby the proposed composite validation technique was However, BP modelling languages often lack sufficient
developed and evaluated at a HEI in South Africa (Section 4). understandability due to the complexity of BP modelling
The evaluation results suggest that the application of the methods, notation and syntax rules [25]. A study by Mendling,
composite validation technique was successful and that it was Reijers and Cardoso [26] confirms that students tend to
well received by students (Section 5). Several conclusions and improve their understanding of process models when more
recommendations from this study can be made (Section 6). time is spent on training with BP modelling tools.
Additionally, Mendling, Reijers and van der Aalst [27] explain
2. BUSINESS PROCESS MANAGEMENT that BP modelling tools that support the modelling process are
BPM is a wide management approach that evolved from favoured by modellers; for example, tools which provide
multiple disciplines and consists of a lifecycle that involves warnings when syntax rules are violated or when models
four generic phases (Section 2.1). The EPC modelling become too large.
language consists of several elements and syntax rules (Section
2.2). Several rules for the validation for BP models can be 2.2 The Event-Driven Process Chain (EPC)
identified (Section 2.3). The BP modelling tool, Microsoft Modelling Language
Visio, is the platform used to implement the composite The Event-Driven Process Chain (EPC) modelling language is
validation technique (Section 2.4). developed within the Architecture of Integrated Information
Systems (ARIS) framework by IDS Scheer [28]. The

94
framework provides a central method to conceptually integrate one branch amongst the multiple outgoing branches is allowed
various dependencies that exist between data, activities, to be activated. The XOR-join is required to wait for the
functions, organisational units and its outputs in terms of activated branch to complete in order to synchronise the
designing for information systems [29]. The advantage of branches into a single control flow.
using the EPC modelling language lies in its high user
acceptance and extensive tool support. EPC models are easily 2.2.2 Event-Driven Process Chain Rules
understood and to express BPs in an intuitive way [7, 30]. EPC According to various authors [21], 31] a strict set of syntax
models are most often used in practice with ERP systems such rules and recommendations apply to the EPC modelling
as SAP R/3 [31] and for teaching introductory BP modelling language. Examples of the main EPC rules are:
[32].
 Events must be used at either the start or end of a process
The importance of rule-adhering EPC models are emphasised and when important conditions exist during the process.
in situations when EPC models need to be interchanged with
 Events are passive elements and cannot make decisions.
respective software execution engines such as SAP R/3 ERP
Events can, therefore, only be followed by an ‘AND’
systems and other BPMS products. EPC models that embed
Logical Connector to trigger Functions in parallel and
issues hinder the smooth interchange and execution of BPs
not an ‘OR’ or ‘XOR’ Logical Connector. Figure 2
across the organisation. For this reason, EPC models often
depicts the allowable logical connections between Events
require intensive revision during the design activity before
and Functions.
being applied to software execution engines in the enactment
 An Event can only be followed by a Function, unless
activity [31, 33]. EPC models are used to logically construct
followed by a Logical Connector.
the business activities at a low, detailed, technical level to
enhance understanding (Section 2.2.1) and must adhere to the  Events and Functions should always alternate. It is still
set of EPC rules (Section 2.2.2). important to alternate between the Event-and-Function
patterns when intercepted by a Logical Connector.
 Functions are allowed to be followed by all the Logical
2.2.1 Elements of the Event-Driven Process Chain Connector types, as Functions are allowed to make
Notation decisions and trigger activities in parallel.
EPC models are constructed as a set of nodes and arcs as a  Every element used within the EPC model must be
directed graph. Nodes are either: Functions, Events or Logical connected using a control flow and provided with an
Connectors (Figure 1). A Function represents a task or an appropriate text label. Elements that are free-standing or
activity that needs to be executed within a process, which uses have no text label add no value to the created EPC
time and resources. An Event describes a condition or a status model.
that exists at a particular time within a process. Events trigger Single Event Single Function Multiple Event Multiple Function
Functions which are in turn triggered by Events. Three types of Trigger Trigger Trigger Trigger

Logical Connectors (AND, OR and XOR) exist, which each


A
can serve as a connector split or a connector join. N V
V
XOR
V
V
V V

V V
O V V

R Not Allowed
VXOR
V

X V
XOR
V
XOR
O Not Allowed
VXOR XOR

Figure 1. Elements of the EPC notation. R

Each of the Logical Connector types differs in their semantic


meaning. Arcs between these elements depict the sequence and Figure 2. Adapted Logical Connections between Events and
direction of control flows (inflows and outflows) in order to Functions [34].
represent the interdependencies that exist between them. The
AND-split logical connector enables a control flow to be split 2.3 Validation of Business Process Models
into two or more parallel outgoing branches. The AND-split An EPC model, whether it is a high-level or low-level
activates all the outgoing branches to execute concurrently. representation of a BP, should always adhere to the set of EPC
The AND-join is required to wait for all active branches to rules. For this reason, the importance of a validation technique
complete in order to synchronise the branches into a single for EPC modelling is emphasised. Having the correct
control flow. The OR-split enables one incoming control flow validation technique that can assist in the identification of
to be split into two or more outgoing branches. The OR-split issues is critical to the quality of the overall EPC model.
logical connector either activates a single branch or multiple Mendling [7] substantiates this need by stating that novice
outgoing branches. In doing so, a specific condition is satisfied modellers often create BP models of a low quality without
and the OR-split will activate at least one of the alternative truly understanding the impact that the models have on
branches and disable the rest. The OR-join is required to wait technologies that are intended to execute the modelled
for all the activated branches to complete in order to processes. It is therefore important to identify issues at the
synchronise the branches into a single control flow. The XOR- earliest possible stage during the design of a BP as the cost to
split logical connector enables one incoming control flow to be repair a BP at a later stage is extremely high [18].
split into two or more outgoing branches. The XOR-split Validating a BP model for compliance with respect to some
creates a situation where a decision needs to be made. Only syntax or business rule is critical and should be supported by

95
appropriate tools and concepts [35]. Furthermore, Kühne et al. 2.4 Microsoft Visio
[10] explain that a validation technique should follow a Microsoft Visio is an example of a simple modelling tool
framework pertaining to three quality goals namely: syntactic which is used both in industry for professional modelling and
quality, semantic quality and pragmatic quality. Syntactic in educational environments for modelling and diagraming
quality refers to the extent to which the process models adhere instruction [24, 25]. Users tend to be familiar with Visio’s
to the syntax rules for a given BP modelling language. simple drag-and-drop approach and the tool provides a pre-
Semantic quality is concerned with the completeness and defined set of templates for the most common modelling
validity of the process model, implying that a process model is languages. Visio has been reported to be easy to use and can
only valid when all statements made in the model are correct easily be customised. The modeller manipulates the model by
and that the model has no missing elements that give meaning selecting elements, also known as shapes, from a stencil and
to the process as a whole. Pragmatic quality refers to the structures those elements on a canvas. Microsoft Visio
assistance that is provided to the modeller by recommending a Professional 2013 offers a standard validation technique for
better alternative in the given situation. Pragmatic quality is BPMN, Cross-Functional flowchart, Basic Flowchart, Six
concerned with the selection of the most practical alternative to Sigma and Microsoft SharePoint 2013 Workflow templates
derive at a construct that depicts the same meaning. [38]. This validation technique is executed through a validation
In order to provide automated support for compliance engine, which is programmed in a declarative validation
checking, compliance rules have to be expressed in a formal language. The validation technique provides the modeller with
language [35, 36]. Declarative validation languages typically textual feedback on any issues that are embedded in the created
constrain the patterns in the model by using Boolean queries BP model. Visio provides a Validation Application
(If-clause) to suit the compliance rule, such as “A True value is Programming Interface (API), which enables programmers to
returned If Element A is connected to Element Z” to indicate develop their own rules for any Visio template [38]. The
that a fragment adheres to the rule [35]. Declarative validation default feedback provided for any models that fail validation is
languages enable queries to be executed rapidly as small provided as “issues” and provided as textual feedback only.
fragments are targeted in the created BP model. The Thus, graphical feedback or visual alerts are not natively
advantages of using declarative validation languages are supported in the Visio environment and validation is performed
twofold. Firstly, it supports the creation of complex syntax manually by the user through the use of a “check diagram”
rules that are applied to parts of the BP model. Secondly, the button. Furthermore, the standard EPC template supported in
majority of these languages have the ability to provide real- Visio does not offer a set of EPC rules to validate the created
time textual and visual feedback based on the implemented EPC model against.
rules [10, 20, 36, 37].
Transgressions of each EPC rule category can be classified into 3. A COMPOSITE VALIDATION
five logical categories of similar rules (Table 1). These logical TECHNIQUE
groupings can be further decomposed into two high-level A validation technique should not only detect existing issues in
groups: major and minor rule violations. Major rule violations a model, but it should communicate additional information to
are syntactic violations, while minor rule violations assist in the modeller about the particular cause and solution to the
the readability of the constructed model. Although minor rule issue, as well as, any warnings to prevent issues [18]. Wynn et
violations do not directly violate syntactical correctness of the al. [39] and Kühne et al. [10] suggest that a validation
EPC model, they can cause deadlocks when interpreted by technique should give instant and continuous feedback about
execution engines if not addressed. any issues that exist whilst modelling. This is supported by
Speck et al. and Awad et al. [18, 35] who propose that a
Table 1. Rule Categories (Based on Monk and Wagner validation technique should be interactive and real-time,
[34]) allowing the modeller to correct an issue as it occurs.
Rule Description Furthermore, a validation technique should provide assistance
Category and recommendations based on the solution to an issue [40].
Four main features of BPM validation techniques are therefore
Alternation Includes rules that govern the Event-
identified in this study and together comprise a composite
to-Function alternations throughout the
validation technique which serves as the key contribution of
created EPC model. This category also
this paper (Figure 3). The features were all incorporated into
includes rules that govern the
the Visio platform as a proof of concept (Section 3.1) and are:
alternation of Events and Functions
around Logical Connectors.  Rule set categorisation (Section 2.3);
 Textual feedback (Section 3.2);
Major

Logical Includes rules that govern the


Connectors allowable Logical Connections (Figure  Graphical feedback (Section 3.2); and
2). This category prevents Event  Real-Time validation (Section 3.3).
elements being used for decision
making.
Control Includes rules that prohibit elements
Flow from exceeding the allowable amount
of inflows and outflows.
Text Label Ensures that all elements are labeled.
Minor

Connectivity Includes rules ensuring no elements


are free standing and have at least a
single outflow or inflow.

96
differentiate between minor and major issues (Figure 4, notes a
and c).

a) Red graphical b) Visio c) Yellow


effect for major supported graphical effect
issues selection for minor issues

Figure 3. Composite Validation Technique. d) Textual descriptions


and categories

3.1 Application of the Composite Validation


Technique
This paper proposes the composite validation technique and Figure 4. Colour coding graphical (a, c) and textual
applies it to the Visio platform as proof of concept. The Visio feedback (d).
platform was therefore modified by changing the default If a serious modelling rule violation is detected, the element, or
canvas behaviour. Firstly, the set of EPC rules were subset of elements, is given a red glowing border for
programmed for the EPC template in Visio. Programming the immediate notice. Less serious issues are highlighted in a
EPC rules enabled the native Visio validation engine to similar fashion but with a yellow colour which indicates that
compare the rule-logic with the created model onscreen. In they are not as urgent as red issues. Three alternatives to
addition to the rule-logic, a description of each rule and the glowing effects were considered (Figure 5), namely: adding a
rule-category was programmed along with a possible solution glow effect around the border of offending elements, changing
to provide the modeller with textual feedback when a rule is the line weight of borders of offending elements and changing
transgressed. Visual alerts were incorporated to indicate the the actual colour of the elements. By consulting BPM students
exact element(s) onscreen that transgresses a particular during a focus group discussion, it was determined that the
validation rule, as well as, whether the rule transgression is of a glow effect would be most effective graphical effect and was
minor or major nature. Lastly, the validation process in Visio revealed to be the most visually striking and was thus selected.
was automated by incorporating continuous sectional checking
of subsets of the created diagram. Thus, the modeller receives
instant, real-time graphical and textual feedback as the
elements are manipulated onscreen. Providing multiple types
of feedback allows the modellers to decide which means of
feedback is most desirable for them at any given stage.

3.2 Textual and Graphical Feedback


When a model or part thereof fails validation due to a rule
violation, such as incorrectly linked EPC elements, the Figure 5. Candidate graphical compliance effects for rule
modeller should be informed with the cause together with a violations.
suggestion to solve the issue [10]. The textual feedback lists
descriptions for each incurred issue in a separate feedback sub- 3.3 Real-Time Validation
window (Figure 4, note d). The Visio environment provides To apply validation to a model, the Visio environment only
native support for the display of textual feedback to the supports user-requested validation through the use of a “check
modeller when a declarative rule indicates that a validation diagram” button [38]. Continuous, or real-time, validation is
issue has been detected. The modeller will typically click on an considered to be a favourable alternative to user-requested
issue’s text and expect the offending element to be selected validation [10]. The composite validation technique uses real-
(Figure 4, note b), which is natively supported by Visio. time validation to allow novice modellers to detect issues as
they are modelling and to thus be aware of any modelling
Graphical feedback is a feature that applies graphical effects to flaws before the model is finalised.
the elements that are involved with issues in a BP model.
Awad, Weidlich and Weske [35] apply the graphical By ensuring that novice modellers receive feedback in real-
highlighting of process paths in a BP modelling language time, the Visio environment was adapted through macros
similar to EPC. Graphical feedback is also enabled using a which trigger whenever an element is placed or interacted
combination of pattern recognition and declarative validation with. This allows the modeller to ignore an issue if they desire,
languages to highlight elements according to a colouring while the feedback about the issue will remain persistent. By
algorithm [41]. Kühne et al. [10] illustrate a different method applying persistent feedback, the risk of the modeller
for graphical feedback by annotating the elements involved in temporarily ignoring an issue and subsequently forgetting to
issues with markers. Graphical feedback features are rectify it is minimised [20, 35].
advantageous to modellers as it provides feedback in an
intuitive manner which is easy to understand to enhance direct 4. RESEARCH METHODOLOGY
understanding [20]. In addition to the textual feedback, The goal of this paper is to propose and evaluate a real-time
graphical feedback is provided using two different colours to composite validation technique for novice BP modellers.

97
Therefore the main research question this paper addresses is perceived impression levels and system preferences, and
“Can the proposed real-time, composite validation technique covers a variety of characteristics of the system under
successfully assist novice modellers with their knowledge of the evaluation, such as usefulness, ease of use and complexity
rules of a BP modelling language?” The proposed, real-time [45]. Each statement in the SUS is answered on a five-point
composite validation technique was implemented in the Visio Likert scale on which participants rate their level of agreement
environment and the success of the technique was evaluated by [46]. Half of the statements were positively worded and half
means of a field study involving an informal focus group and a were negatively worded. The five-point Likert scale ratings
usability evaluation (Section 4.1). A number of research ranged from strongly disagree (a rating of one) to strongly
instruments (Section 4.2) were used to measure the success of agree (rating of five). According to Tullis and Albert [46], it is
the technique. Students of a BPM course were used as not recommended to report on the ten statements individually,
participants in the study to evaluate the modified Visio but should rather be reported on as a combined rating for
software (Section 4.3). overall usability. Furthermore, it is convenient to think of the
SUS scores as a percentage, since they are on a scale of 0-100,
4.1 Research Design with 100 representing a perfect score. Each statement’s
In order to determine the preferred features for a BP validation contribution towards the combined SUS score differed and was
technique a focus group approach was used with BP modellers. calculated according the method suggested by Brooke [45].
A cognitive walkthrough was used in a focus group to identify
any usability problems or cognitive issues, by analysing the The third section of the questionnaire consisted of statements
mental processes, assumptions and goals required by users of regarding the validation technique’s perceived usefulness for
the validation technique [42]. Determining the tasks that preventing successive issues from occurring which participants
modellers perform during a modelling activity is particularly were required to rate on a five-point Likert scale. All
useful for selecting the most appropriate validation features to participants were required to complete the same task-list,
assist in those tasks. A single cognitive walk-through was which consisted of three major tasks with predefined subtasks.
performed using six participants, to determine broad issues and The subtasks instructed the participants to specifically structure
ideas. The participants were selected through purposive the EPC model in violation of the set of EPC rules. It was
sampling, ensuring that the participants had some experience necessary to include subtasks which triggered the validation
with BP modelling. These participants were all HEI students aspects of the prototype. The subtasks also assisted participants
that have either completed or were still attending a one- with rectifying the issues. Task 1 was simple with the goal of
semester BP modelling course. The cognitive walkthrough was getting participants acquainted with the EPC validation
initiated by presenting participants with a specific set of tasks technique in Visio. Tasks 2 and 3 evaluated the validation
that are typically performed during a modelling activity. Each technique’s ability to assist modellers in preventing and
of the proposed validation features was individually introduced rectifying rule violations throughout the modelling activity.
and discussed by demonstrating the functionality of each Tasks 2 and 3 included a single “problem-solving” subtask
feature using a combination of paper-based prototypes and which required participants to use their own initiative to rectify
Microsoft Visio software. This allowed participants to evaluate existing issues in the created EPC model. The problem-solving
the proposed validation features that appeared to be preferred subtasks required participants to rectify the issues from the
in terms of accomplishing modelling tasks. Any suggestions validation feedback provided on the screen. Participants were
and concerns surrounding the usefulness of the proposed also allowed to use their own knowledge to solve problem-
design were raised and discussed. The different validation solving tasks. The major tasks included rule violations relating
feedback methods were evaluated and the most preferred to all categories. It would have been impractical to test every
methods were incorporated into the design of the modified rule that was implemented in the prototype. A decision was
Visio software (the prototype) which implemented the real- made to include at least one rule from each category. Eye-
time, composite validation technique. tracking has become significantly more common over the past
few years as it shows researchers exactly where participants
Once development of the prototype was complete, a pilot test are looking at a specific point in time [46]. In order to analyse
was conducted with two modellers who had some BP the performance of participants’ task list actions, the onscreen
modelling experience, both external to the sampled users for video Pupil Centre Corneal Reflection (PCCR) technique was
the cognitive walkthrough and usability studies. Two minor used to analyse eye-movement, including gaze plots and
problems were identified and the changes were made to the fixation patterns, as well as mouse-movements involving
prototype. The final prototype was then evaluated according to mouse-clicks and mouse roll-overs [47].
its usability, since usability is increasingly recognised as an
important quality factor for interactive software systems. 4.3 Participant Profile
Usability focusses on measures for effective, efficient and Information Systems (IS) students from the Nelson Mandela
satisfactory task execution and aims to support the ordinary Metropolitan University (NMMU) in South Africa were
and uninterrupted interaction between the user and the system selected as a purposive sample for the focus group (n=6) and
[43]. Evaluating the usability of a system is particularly useful usability evaluation (n=14) with no overlap between samples.
when aiming to improve the user interface or to establish the The usability evaluation was conducted with participants who
quality in use within a given context. are representative of students that typically enroll in the BP
modelling course every year. A total of 14 participants were
4.2 Research Instruments selected for the usability evaluations with the majority (64%)
A post-test questionnaire was administered to participants being male (n=9) and the remainder (36%) were female (n=5).
which was based on the System Usability Scale (SUS). The The majority of the students (93%) were between the ages of
first section of the questionnaire addresses the participants’ 20 and 22 (n=13) and all of the students had prior experience
previous experience since this can have an effect on the results (1-6 months) with BP modelling and Microsoft Visio. The
of the usability evaluation [44]. The SUS is a popular standard, application of the composite validation technique is
ten-statement, usability measurement tool that measures users’ particularly suitable within this BPM course as the students are

98
expected to draw BP models with Visio. Since the validation progressed through the task-list. This result is supported by the
technique was tested with the EPC modelling language, aggregated heat map of participants (n=14) for task 1 (Figure
participants were required to have some knowledge of, and 6). The heat map indicates a variety of spots consisting of
practical experience with, the EPC modelling language. green, yellow or red colours. Green spots represent little focus
or a typical glance on the affected area onscreen. Yellow spots
represent a moderate amount of focus was applied on the
5. RESULTS OF USABILITY
affected area onscreen, whereas, red spots represent extensive
EVALUATION focus on the affected screen areas.
5.1.1 Effectiveness and Efficiency It can be deduced that from the diminishing awareness of
The effectiveness metric was assigned two sub-metrics: a) minor issues that the participants perceived the minor issues as
task-completion and b) issue correction rate. Task-completion a low priority. Participants eventually started to temporarily
indicated whether the validation technique assisted modellers ignore the minor issues and devoted less effort to read the
to successfully complete the modelling activity. Issue associated textual descriptions to solve the issue. The
correction rate indicate whether the modellers were able to decreased awareness in minor issues did not, however,
correct issues in order to derive at a successfully completed decrease the effectiveness of the validation technique as
and rule-adhering EPC model. The majority of the participants participants in due course fixed the minor issues successfully.
(n=12) were able to complete all the tasks successfully to The fact that participants started to temporarily ignore the
derive at the final EPC model. Task 3 was deemed minor issues to address the major issues supports the result that
unsuccessfully completed by only two (n=2) participants, be learning did occur during the course of the evaluation. This is
because they failed to place a matching “OR” Logical evident since participants could distinguish that the minor
Connector to merge the process in the created EPC model. issues were of a preventative nature. Participants learned that
They placed a Logical Connector in the correct position, but although minor issues were temporarily ignored, they would
placed an “XOR” and an “AND” Logical Connector not affect the core structure of the EPC model and could rather
respectively. A rule governing the matching Logical focus on major issues that would have a significant effect on
Connectors are typically related to the EPC modelling the model if not fixed immediately.
language’s semantics. The proposed validation technique does
not attempt to address the semantics of the EPC modelling
language. For this reason, the validation technique relies on c) Extensive
focus on major
modellers to use their own initiative to correct any issues issue
surrounding the matching Logical Connectors. Although the a) Less focus
scope of this study excluded the evaluation of matching on minor
Logical Connectors, this issue provides evidence that such a issues
validation rule is a requirement. Since Task 3 was the final task b) Less focus
d) Extensive on category
and two participants failed to successfully complete the final
focus on column
problem-solving task, two of the final EPC models were
first part of
declared to be unsuccessfully completed.
the textual
A total of nine rule violations were made by participants descriptions
during Task 3 of which only seven were successfully
corrected. The remaining two issues were made by same
participants that caused two of the final EPC models to be Figure 6. Visual heat map aggregated from all participants
declared unsuccessful. These seven corrected issues related to for task 1.
spelling, reading, partial disobedience and connecting multiple
inflowing or outflowing Dynamic Connectors to an Event or The major issues were represented by a large-sized, red glow
Function element. A total of 20 issues were made during the effect. The awareness of major issues remained constant
entire usability evaluation. The validation technique enabled throughout the evaluation. Elements that had undergone a red
participants to correct 90% of the total issues that were made, glow effect caused participants to generate extensive focus on
which supports the validation technique’s ability to effectively the elements involved in the issue (Figure 6, note b). An
enable participants to correct issues. The validation technique observation was made as participants were timeously inclined
is efficient since the mean time overall for participants to to execute a task onscreen and instantly read the list of task
complete the entire task-list was 16 minutes and 18 seconds, instructions once more. For this reason, some elements
which is less than the target time of 20 minutes. involved in major issues were temporarily unnoticed. These
elements were, however, instantly noticed when participants
refocused on the screen.
5.1.2 Issue Awareness
Issue awareness was used as a metric for measuring the support The participants were inclined to perceive the major issues as
for learning of BP rules provided by the software. Issue high priority. Another observation was made as the major
awareness was further categorised into a) awareness of minor issues encouraged participants to read the textual descriptions
and b) major issues, as well as, c) awareness of textual more thoroughly compared to minor issues. Participants were
feedback. The minor issues were represented by a medium- also inclined to devote more effort to solve the major issues
sized, yellow glow effect. Minor issues were typically noticed instantly. This was done by observing the surrounding EPC
for elements at the beginning of each task, such as, when a structure and related textual descriptions of the issues in the
newly added element required a text label or had no incoming EPC model. An observation was made from the eye-tracking
or outgoing Dynamic Connector. A common trend was video recordings that participants were more likely to identify
observed amongst participants (n=14) for the minor issues. The issues from the graphical feedback than the textual feedback.
awareness of minor issues gradually diminished as participants The eye-tracking indicated that focus was typically generated

99
on the glow effect before the participant focused on the textual “Real-time validation” and “Prevents successive issues” all
descriptions. The heat maps and eye-tracking videos depicted received the same rating (µ = 4.57). The lowest rated metric
that less focus was generated to read the entire textual was for the rule set categorisation feature (µ = 4.36); however
description of reoccurring issues (Figure 6, noted). This can be this is still a positive rating. Since all of the overall metrics
viewed from the heat map as extensive focus (large red spots) were in the positive range, it can be deduced that the
appear on the first parts of the descriptions, which fade to a implementation of the real-time, composite validation
lighter green spot towards the end of the textual description. technique in the EPC language in Visio was perceived as
Occasionally, participants found it sufficient to read the initial useful by the participants.
part of the textual description to gain an understanding. This
was particularly identified for participants that devoted less 5.1.4 System Usability Scale (SUS)
effort to read the textual descriptions associated with minor A mean overall SUS score below 60% is considered relatively
issues. Participants were also more likely to notice an update of poor, while one over 80% is considered good [46]. The mean
textual descriptions when the Issues Window was less overall SUS rating for the prototype (which implemented the
crowded. The heat map depicts that the participants did not proposed validation technique) was 87%. It can be deduced
regularly use the Category column to solve issues (Figure 6, therefore that the participants were satisfied with the usability
note c). Therefore, the textual descriptions of categories cannot of the validation technique (Figure 8). For demonstration
be proven to be as effective as the full textual descriptions for purposes, the ratings of negative worded statements are
each issue. depicted in a lighter colour. All of the positively worded
Participants also displayed trust in the validation technique, as statements in the SUS questionnaire had overall mean ratings
minor issues could be temporarily ignored and textual in the positive range. The statement “I would use the
descriptions could be read less thoroughly, as they progressed validation technique frequently” was the highest rated (µ =
through the task list, to gain an understanding of the issue. For 4.64) SUS statement. The second highest rated (µ = 4.57) SUS
this reason, participants became more confident with the statement was, “The system was easy to use”.
validation technique during the evaluation. A deduction can be
drawn from these results to support the fact that learning of the
validation technique did occur.

5.1.3 Usefulness
The perceived usefulness of the validation technique was
analysed according to six metrics: a) to d) the perceived
usefulness ratings of the features: textual feedback, graphical
feedback, real-time validation and rule set categorisation, and
e) the perceived usefulness rating for increasing a participant’s
knowledge of EPC rules and f) the perceived usefulness for
preventing successive issues from occurring. Figure 7 indicates
the mean score ratings for each perspective and includes error
bars indicating the standard deviation (σ) from the mean. Three
frequency ranges of the Likert scale item responses were
statistically identified and calculated in order to classify mean Figure 8. Mean score per SUS statement.
ratings as negative (1.0 ≥ µ < 2.6); neutral (2.6 ≥ µ ≤ 3.4) or
positive (3.4 > µ ≤ 5.0). The statement “I felt confident using the validation technique”
was also rated positively and had the third highest mean rating
(µ = 4.50). Four of the five (80%) of the negatively worded
statements had overall mean ratings in the positive range. The
statement “The system had a low complexity” was the highest
rated negatively worded statement (µ = 3.64). The second
highest rated negatively worded SUS statement was
“Inconsistency in the system is low” (µ = 3.64). From this it
can be deduced that the participants felt that the system was
not complex and was consistent.

The open-ended questions were analysed and grouped


according to themes. Comments from participants revealed that
textual feedback was thought to be easy-to-understand and
informative. The textual feedback could be easily read,
informing participants about the cause of the issue and possible
solutions. The graphical feedback was thought to be precise,
Figure 7. Usefulness of composite, real-time validation distinguishable, encouraging and visually appealing. The
features. precision of the glow effect around elements enabled
All six metrics had overall mean ratings in the positive range. participants to focus on the exact element involved in the issue.
The graphical validation feature had the highest mean rating (µ It was identified that the precision of the graphical effect
= 4.93). The second highest rated feature was “Enhanced served as an encouragement to correct issues instantly.
knowledge of EPC rules” (µ = 4.71). This supports the studies Participants stated that the yellow and red glow effects were
[20, 35] showing that graphical feedback can enhance the visually appealing and assisted with distinguishing between
understanding of BP rules. Three features “Textual feedback”, major and minor issues. The real-time validation feature

100
enabled the validation technique to be perceived as being BP modelling competencies of IS graduates, and in this way
reliable, preventative, responsive, and natural. The real-time can reduce the gap that exists between industry needs and the
validation enabled the textual and graphical validation features competencies of IS graduates. The validation technique can
to serve as reliable and natural guide to modellers that also be used by modellers working in industry to improve BP
informed them of any issues that existed within the created BP model accuracy and BP modelling project success. Future
model instantly. The qualitative data analysis revealed that research could be undertaken to investigate the effectiveness of
some participants perceived the re-occurring textual the proposed validation technique in other environments and
descriptions as cumbersome to read and required a sufficient on other platforms such as mobile devices.
description for the exact matching Logical Connectors.
Another limitation identified was that the participants
7. REFERENCES
perceived the Issues Window to be overcrowded with textual
descriptions at times. In terms of graphical feedback, some [1] Curtiss, E. T. 2013. Business Process Management (BPM),
participants stated that the meaning of the yellow glow effect Cloud, Mobile, and Patterns: Market Shares, Strategy, and
was initially confusing, but could assume that it represented Forecasts, Worldwide, 2013 to 2019.
minor issues.
[2] Recker, J. and Dreiling, A. 2007. Does It Matter Which
Process Modelling Language We Teach or Use ? An
6. CONCLUSIONS AND Experimental Study on Understanding Process Modelling
RECOMMENDATIONS Languages without Formal Education. In Proceedings of the
In this paper a real-time, composite validation technique is Australasian Conference of Information Systems
proposed which detects and to a certain extent prevents (Toowoomba, Australia, December5-7, 2007). 356–366.
modellers from incurring issues within the created BP model.
The validation technique provides textual descriptions of the [3] Recker, J. and Rosemann, M. 2009. Teaching Business
issues identified and possible solutions to correct the issue. Process Modelling: Experiences and Recommendations.
Furthermore, issues are detected in real-time and a distinction Communications of the Association for Information Systems.
is presented to the modeller between minor and major issues 25, 32 (September 2009), 379–394.
using graphical methods. The validation technique was
demonstrated with the EPC modelling language and evaluated [4] Ko, R. K. L. 2009. A Computer Scientist’s Introductory
by a group of BP modelling students. Guide to Business Process Management (BPM). Crossroads.
15, 4 (June 2009), 4.
The results of the field study evaluations were positive, since
the validation technique allowed modellers to successfully [5] van Greunen, D., van der Merwe, A. and Kotze, P. 2010.
detect and correct issues. Furthermore, modellers appeared to Factors Influencing BPM Tools: The Influence on User
quickly understand the modelling language’s feedback method. Experience and User Interfaces. International Journal of
The participants in the field study rated the graphical feedback Computing and ICT Research. 4, 1 (October 2010), 47–57.
to be the most useful feature of the proposed validation
technique for detecting issues. The real-time feedback was also [6] Harmon, P. 2009. Business Process Change: A Guide for
rated positively and enabled participants to detect problems business Managers and BPM and Six Sigma Professionals,
before the modelling activity was completed. The textual Second Edition. Morgan Kaufmann Publishers, Burminham.
feedback was also rated positively and textual descriptions
were used to assist participants in understanding the underlying [7] Mendling, J. 2007. Detection and Prediction of Errors in
causes and solutions of the issues. From this it can be deduced EPC Business Process Models. Doctoral Thesis. Vienna
that the proposed real-time, composite validation technique University of Economics and Business Administration.
significantly assisted modellers in terms of understanding the
[8] Li, Y., Cao, B., Xu, L., Member, S., Yin, J, Deng, S. and
rules of a BP modelling language. One limitation of the study
Yin, Y. 2014. An Efficient Recommendation Method for
was that the prototype does not fully offer support to deactivate
Improving Business Process Modeling. IEEE Transactions
all validation functions. In addition, the validation technique
on Industrial Informatics. 10, 1 (February 2014), 502–513.
was only evaluated with high-level BPs on a single modelling
canvas with novice modellers. [9] van der Aalst, W. M. P. 2013. Business Process
The results of this study reveal that HEIs can adopt a real-time, Management : A Comprehensive Survey. ISRN Software
composite validation technique to assist novice modellers in Engineering, 1,1 (September 2012), 1–37.
their BP modelling courses. The proposed rule set
categorisation incorporated in the technique provides a [10] Kühne, S., Kern, H., Gruhn, V. and Laue, R. 2010. Business
valuable contribution to the research field of real-time Process Modeling with Continuous Validation. Journal of
validation techniques for BP modelling. Whilst the field study Software Maintenance and Evolution: Research and
was limited to a small group of BP modelling students at one Practice. 22, 7 (October 2010), 547–566.
institution, this is a sufficient number of participants for
usability studies. The results of the field study provided a more [11] Seethamraju, R. 2012. Business Process Management: A
in-depth understanding of the challenges encountered by Missing Link in Business Education. Business Process
modellers in drawing BP models and trying to understand the Management Journal. 18, 3 (2012), 532–547.
rules of the modelling language. The benefit of implementing
[12] Bandara, W., Chand, D. R., Chircu, A. M., Hintringer, S.,
the proposed validation technique is that it can assist modellers
Karagiannis, D., Recker, J., van Rensburg, A., Usoff, C. and
with improving their knowledge of a particular BP modelling
Welke, R. J. 2010. Business Process Management Education
language’s syntax rules, thereby improving the overall quality
in Academia: Status, Challenges, and Recommendations.
of the created BP model. Thus, the real-time, composite
validation technique could contribute towards the attainment of

101
Communications of the Association for Information Systems. [27] Mendling, J., Reijers, H. and van der Aalst, W. M. P. 2010.
27, 1 (November 2010), 743–776. Seven Process Modeling Guidelines (7PMG). Information
and Software Technology. 52, 2 (January 2008), 127–136.
[13] Gröner, G., Bošković, M., Silva Parreiras, F. and Gašević, D.
2013. Modeling and validation of business process families. [28] Opocenská, K. and M. Kopecký. 2008. Incox–a language for
Information Systems. 38, 5 (July 2013), 709–726. XML integrity constraints description. In Proceedings of the
Dateso 2008 Workshop (Cerna Ricka, Czech Republic, April
[14] Barjis, J. 2008. The Importance of Business Process 16-17, 2008), 1–12.
Modeling in Software Systems Design. Science of Computer
Programming. 71, 1 (March 2008), 73–78. [29] Mendling, J. and van der Aalst, W. M. P. 2007.
Formalization and Verification of EPCs with OR-joins Based
[15] Recker, J. 2010. Opportunities and Constraints : The Current on State and Context. In Proceedings of the 19th International
Struggle with BPMN. Business Process Management Conference, CAiSE 2007 (Trondheim, Norway, June 11-15,
Journal. 16, 1 (2010), 181–201. 2007), 439–453.
[16] Reijers, H. and Mendling, J. 2011. A Study Into the Factors [30] Bögl, A., Schrefl, M., Pomberger, G. and Weber, N. 2009.
That Influence the Understandability of Business Process Semantic Annotation of EPC Models in Engineering
Models. IEEE Transactions on Systems, Man, and Domains to Facilitate an Automated Identification of
Cybernetics - Part A: Systems and Humans. 41, 3 (May Common Modelling Practices. In Enterprise Information
2011), 449–462. Systems. Springer, Berlin-Heidelberg, 155–171.
[17] Simon, H. 1996. The Sciences of the Artificial, Third [31] van Dongen, B. F. and Jansen-Vullers, M. H. 2005. EPC
Edition. The MIT Press, Cambridge. Verification in the ARIS for MySAP Reference Model
Database. In Proceedings of the Business Process
[18] Speck, A., Witt, S., Feja, S., Lotyzc, A. and Pulvermüller, E. Management with Event-Driven Process Chains Conference
2011. Framework for Business Process Verification. In (Hamburg, Germany, December 8-9, 2005), 24–40.
Proceedings of the 14th International Conference on
Business Information Systems (Pozna´n, Poland, June 15-17 [32] Laue, R. and Mendling, J. 2010. Structuredness and its
2011), Springer LNBIP, 50–61. Significance for Correctness of Process Models. Information
Systems and E-Business Management. 8, 3 (June 2010), 287–
[19] Ly, L. T., Rinderle-ma, S. and Dadam, P. 2010. Design and 307.
Verification of Instantiable Compliance Rule Graphs in
Process-aware Information Systems. In Advanced [33] van der Aalst, W. M. P. 2008. Challenges in Business
Information Systems Engineering. Springer, Berlin- Process Analysis. In Enterprise Information Systems.
Heidelberg, 9–23. Springer, Berlin-Heidelberg, 27–42.
[20] Laue, R. and Awad, A. 2011. Visual Suggestions for [34] Monk, E. and Wagner, B. 2009. Concepts in Enterprise
Improvements in Business Process Diagrams. Journal of Resource Planning: International Edition, Third Edition.
Visual Languages & Computing. 22, 5 (October 2011), 385– Course Technology, Boston.
399.
[35] Awad, A., Weidlich, M. and Weske, M. 2011. Visually
[21] Becker, J., Krugeler, M. and Rosemann, M. Process Specifying Compliance Rules and Explaining Their
Management: A Guide for the Design of Business Processes, Violations for Business Processes. Journal of Visual
Second Edition. Springer, Heidelberg. Languages & Computing. 22, 1 (February 2011), 30–55.
[22] Hallerbach, A., Bauer, T. and Reichert, M. 2010. Capturing [36] Knuplesch, D., Reichert, M., Ly, L. T., Kumar, A. and
Variability in Business Process Models: The Provop Rinderle-Ma, S. 2013. Visual Modeling of Business Process
Approach. Journal of Software Maintenance and Evolution: Compliance Rules with the Support of Multiple Perspectives.
Research and Practice. 22, 7 (October 2010), 519–546. In Conceptual Modeling. Springer, Berlin-Heidelberg, 106–
120.
[23] Rozman, T., Vajde-Horvat, R. and Draghici, A. 2011.
Business Process Management Certifications Overview and [37] Semmelrodt, F., Knuplesch, D. and Reichert, M. 2014.
Transfer of Innovation. In Proceedings of the Management, Modeling the Resource Perspective of Business Process
Knowledge and Learning International Conference Compliance Rules with the Extended Compliance Rule
(Romania, December, 2010). Graph. In Enterprise, Business-Process and Information
Systems Modeling. Springer, Berlin-Heidelberg, 48–63.
[24] Recker, J. 2012. Modeling With Tools is Easier, Believe Me
- The Effects of Tool Functionality on Modeling Grammar [38] Hopkins, C. R. 2013. Creating Custom Validation Rules in
Usage Beliefs. Information Systems. 37, 3 (May 2012), 213– Visio 2013. [Online]. Available:
226. http://blogs.msdn.com/b/chhopkin/archive/2013/01/03/creati
ng-custom-validation-rules-in-visio-2013.aspx. [Accessed:
[25] Rosemann, M. 2006. Potential pitfalls of process modeling: 06-Aug-2013].
Part A. Business Process Management Journal, 12, 2 (2006)
, 249–254. [39] Wynn, M.T., Verbeek, H. M. W., van der Aalst, W. M. P.,
ter Hofstede, A. H. M. and Edmond, D. 2007. Business
[26] Mendling, J., Reijers, H. A. and Cardoso, J. 2007. What Process Verification - Finally a Reality!. Business Process
Makes Process Models Understandable?. In Business Management Journal. 15, 1 (August 2007), 74–92.
Process Management. Springer, Berlin-Heidelberg, 48–63.

102
[40] Gruhn, V. and Laue, R. 2007. Checking Properties of Digital Libraries. Information Processing & Management.
Business Process Models with Logic Programming. In 44, 3 (May 2008), 1234–1250.
Proceedings of the 5th International Workshop on
Modelling, Simulation, Verification and Validation of [44] Effinger, P., Seiz, S., Jogsch, N. and Eberhard, K. 2011.
Enterprise Information Systems, MSVVEIS-2007, In Evaluating Single Features in Usability Tests for Business
conjunction with ICEIS 2007 (Funchal, Madeira, Portugal, Process Modeling Tools. In Proceedings of the Informatics
June, 2007), 84–93. 2011 conference (Berlin, Germany, October 4-7, 2011).

[41] van Dongen, B.F., Mendling, J. and van der Aalst, W. M. P. [45] Brooke, J. 1996. A Quick and Dirty Usability Scale. In
2006. Structural Patterns for Soundness of Business Process Usability Evaluation in Industry. Taylor & Francis Ltd,
Models. In Proceedings of the Enterprise Distributed Object London, 189–191.
Computing Conference, EDOC’06. 10th IEEE International
(Hong Kong, China, October 16-20, 2006), 116–128. [46] Tullis, T. and Albert, B. 2013. Measuring the User
Experiecne, Second edition. Morgan Kaufmann, Burlington,
[42] Lewis, C. and Wharton, C. 1997. Cognitive Walkthroughs. In 138–139.
Handbook of Human-Computer Interaction. Elsevier,
Amsterdam, 717–732. [47] Coltekin, A., Heil, B., Garlandini, S. and Fabrikant, S.I.
2009. Evaluating the Effectiveness of Interactive Map
[43] Tsakonas, G. and Papatheodorou, C. 2008. Exploring Interface Designs: A Case Study Integrating Usability
Usefulness and Usability in the Evaluation of Open Access Metrics with Eye-movement Analysis. Cartography and
Geographic Information Science. 36, 1 (August 2009), 5–17.

103

You might also like