Complex processes have become more common and include, for example, knowledge workers ie people who work with more complex and problematic tasks. Our interest is to better understand how a problematic and complex situation in an organization should be model. Four major process oriented modeling techniques has been analyzed using Moody's quality criterion for what makes a good model. The findings indicate that there are only one candidate, BPMN, that meets some of Moody's quality criterion, but there are also weaknesses in BPMN, such as for the criterions; expressiveness and emphasis.
Modelling has always been an essential part of organizational design as well as
information systems development. Models enable decision makers to filter out the
complexity of the real world; so that efforts can be directed toward the most important
parts of the system can be studied. [
Processes on the overall levels are considered more complex and more difficult to
analyze because they are at a higher abstraction level. Complex processes have
become more common and include, for example, knowledge workers i.e. people who
work with more complex and problematic tasks. When the activities of the processes
to be described will become more complex the demands on the modelling technique
to be used increases [
Various different models/diagrams have long been used for developing information
systems. Nevertheless, it is little research focusing on how to produce "good"
graphical models [
We are interested in business process modeling techniques and addressed the question: what constitutes a “good” graphical modeling technique according to Moody’s quality criterion? One major reason is our interest to better understand how a problematic and complex situation in an organization could be modeled.
Hence the research question addressed is: Which are the relevant graphical
business process modelling techniques to be used when modelling a problematic and
complex situation? The research approach is evaluation research. In this article we
will:
1. identify relevant graphical modelling techniques
2. identify relevant quality criterion for evaluation of graphical modelling techniques
3. evaluate graphical relevant modelling techniques, using the identified quality
criterion
One of the main motives behind the research is that graphical diagrams are believed
to be more effective than text in communication between end users and/or domain
practitioners [
The research started with a literature review with the aim to identify relevant graphical modeling techniques that could be useful when producing a business process model. We started by reviewing well known articles about graphical modelling techniques and identified references in order to distinguish the most prominent graphical business process modelling techniques that could be used for a a problematic and complex situation. Since our major focus is process modeling techniques we choose to exclude non process oriented graphical modelling techniques. We found four candidates: 1) Flowchart/nodemaps, 2) Business Process Modeling Notation (BPMN), 3) Event Driven Process Chains (EPC), and 4) UMLactivity diagrams. These four process oriented modeling techniques vill be evaluated based by Moodys quality criterions.
The paper is organized as follows. In the next section we will briefly present the identified graphical modeling techniques relevant our research question.
In Section 3 we provide an overview of quality criterions related to business process modelling techniques. In Section 4 we will provide our empirical data. In Section 5 an analysis of the modeling techniques is presented. We conclude by a discussion and identification of future research.
This section aims to present some major business process modeling techniques presented in the ISD-literature. The proposed modeling techniques are characterized as process oriented modeling techniques. The motive for the choice of the four process modeling techniques is that they are widely used in describing problematic and complex situations in organizations.
Event-driven process chains (EPC), business process modeling notation (BPMN)
[
As described by Miller and Mukerji [
BPMN has attracted considerable attention in the IS research field as a convenient
description technique for documenting and re-engineering processes [
UML-activity diagrams are typically used for business process modeling, for
modeling the logic or for modeling the detailed logic of a business rule (Figure 2). In
many ways UML activity diagrams are the object-oriented equivalent of flow charts
[
The main reason for the popularity of EPC is that it is a component in the SAP
Reference Model [
The SAP Reference Model contains almost 10000 individual models with 3000 of
them being EPCs [
A flowchart is a type of diagram that represents an algorithm or process, showing the steps as boxes of various kinds, and their order by connecting these with arrows. This diagrammatic representation can give a step-by-step solution to a given problem.
Process operations are represented in these boxes (Figure 4), and arrows
connecting them represent flow of control. Data flows are not typically represented in
a flowchart, in contrast with data flow diagrams; rather, they are implied by the
sequencing of operations. Flowcharts are used in analyzing, designing, documenting
or managing a process or program in various fields [
Graphical diagrams are believed to be more effective then text in the communication
between end-users and/or domain practitioners [
Moody [
Discriminability means easy to see and to differentiate from one another. It discusses the elements of a diagram. Moody distinguish between absolute and relative discriminability: Absolute discriminability: the ability to see diagram elements and separate them from the background. It is determined by three primary factors: size, contrast, and proximity. Relative discriminability: the ability to differentiate between different types of diagram elements. It is determined by the number and size of differences between symbols used to represent different constructs. The greater the perceptual variation between symbols used to represent different constructs, the faster and more accurately they will be recognised.
Human beings have both perceptual and cognitive limits. One of the most common mistakes in ISD diagramming practice is to show too much information on a single diagram. This results in a diagram that acts as a barrier instead of a communicative aid. Perceptual limits: the ability to visually discriminate between diagram elements decreases with their number and proximity. In general, the difficulty of discerning diagram elements increases quadratically with diagram size. Cognitive limits: the number of diagram elements that can be comprehended at a specific moment is limited by working memory capacity, which is believed to be “seven plus or minus two” concepts at a specific moment. When this is exceeded, a state of cognitive overload ensues and comprehension degrades. One of the most effective ways of reducing complexity is to divide them into smaller parts, decomposition or modularisation.
In most ISD diagrams, all elements look the same: there is no way of telling which are the most important. The most important concepts should be emphasized (highlighted) to bring them to the readers’ attention, while less important or background elements should be de-emphasized (lowlighted).
It is very common in ISD diagramming to use multiple diagrams. Many developers use for instance UML which consists of 13 different modeling notations. The notation should provide an explicit mechanism to support cognitive integration: Conceptual integration: enabling the reader to integrate information from separate diagrams into a coherent mental representation of the problem. Perceptual integration: providing perceptual cues (orienting, contextual and directional information) to aid navigation between diagrams.
Perceptually direct representations are representations whose interpretation is spontaneous or natural, in that their meaning can be extracted automatically by the perceptual system. Representation of constructs: icons are symbols which perceptually resemble the objects they represent. Using icons to represent constructs reduces cognitive load. Representation of relationships: perceptual directness also applies to representation of relationships among diagram elements. Certain spatial configurations of diagram elements predispose a person towards a particular interpretation even before the meaning of the elements is understood.
Organizing diagram elements by a structure into perceptual groups expands the number of elements that can be shown on each diagram without exceeding cognitive limits.
Moody divides the concept of identification into external identification and internal identification. External identification defines the correspondence between the diagram and the represented world. Diagram elements (both nodes and links) should be clearly labeled, using terminology familiar to domain experts to help trigger domain knowledge. Internal identification defines the correspondence between graphical conventions and their meaning. A key should be used summarizing the graphical convention.
Most graphical modelling techniques in ISD use a very limited graphical vocabulary (boxes and arrows) which implies a low expressiveness. We could use the full range of visual variables such as shape, colors, size, value, orientation and texture.
The number of graphical conventions used in a notation should be limited (use visual variables instead). Keep it simple, let simplicity rule.
The evaluation is based on the important findings in the quality criterion presented by
Moody [
The analysis started with a discussion about how to interpret and use the scale. We focused on the contrast between the different values in the scale. In order to give the reader a more thorough understanding of the analysis (and the use of the scale) we will present two examples from our discussion and analysis.
The first example relates to the researchers’ understanding of the concept discriminability in the process-oriented modeling techniques. If we have a model with boxes and connecting arrows/lines (frequently used in ISD) we looked on the arrows to see if there was a description on the arrow and a direction ( = +1), if the arrow did not have the description but had a direction ( = 0), or if the arrow did not have both a description and a direction ( = -1).
The second example relates to the researchers’ understanding of the concept perceptual and cognitive limits in the process-oriented modeling techniques. This started with the question: How well does the modelling technique support decomposition and modularisation? If the modelling technique supports decomposition and modularisation ( = +1), if the modeling technique partially supports decomposition and modularisation e.g. splitting and merging ( = 0), or if the model does not support decomposition and modularisation ( = -1). Our analysis followed the logic in the described examples.
Three researchers did the analysis and classification independently. The outcomes were compared. There was more than an 80% agreement between the three outcomes, which is an acceptable inter coder reliability [17]. Where non-agreement existed, the three researchers discussed the evaluations and final classification and evaluation decision was made.
The result from the analysis is presented in Table 1. The rows in the table are of the nine criterions presented by Moody and the columns are the four process-oriented modelling techniques.
Discriminability Perceptual and cognitive limits Emphasis Cognitive integration Perceptual directness Structure Identification Expressiveness Simplicity
Flowcharts The discussion will be structured according to our results presented in Table 1. The maximum sum that could be reached for the process-oriented modelling technique was 9 and the minimum was -9. Most notable is that only one of the modelling techniques (BPMN) has a positive sum (+2). The other three modelling techniques have a negative sum. Based on the used quality criterion, this indicates that there are problems to be solved when it comes to modelling techniques. BPMN reached the highest sum (+2), but it is far from the maximum sum which indicates that even BPMN could be improved. All four process modelling techniques have a negative value () on emphasis and expresivness, a neutral value () on perceptual directness and identification, and a positive value () on simplicity. These findings indicate that we need to rethink process modelling techniques as a tool for computer independent models is neeeded. First and foremost, process-oriented modelling techniques must start using visual variables to reach a higher expressiveness and emphasis (as described in Section 2). The different process-oriented modelling techniques used more or less the same graphical symbols (variants of boxes and arrows) over and over again and there is no way of telling which are the most important objects/processes in a model.
During the literature review we were also surprised by the fact that: issues of semantics and content (what constructs to include in a notation) are treated as matters of substance, but details of graphical syntax (how to visually represent these constructs) are treated with no or little consequences, Many graphical notations in ISD are not consistent with principles of graphical design (visual appearance or form). Graphical models are treated as standalone artifacts.
When we started the research process the underlying research question was: Which are the relevant graphical business process modeling techniques to be used when modeling a problematic and complex situation? The answer to that question is that we had four candidates of process-oriented modeling techniques: flowcharts, BPMN, EPC, and UML-activity diagrams. They satisfy (to some extent) the seven demands/requirements put forward by OMG. Using Moody’s quality criterions to anlyze the models, only BPMN reached an acceptable level.
There are strong indications that business process models should benefit from not only to be discussed from issues of semantics and content (what constructs to include in a notation) but also details of graphical syntax (how to visually represent these constructs) and how well they serve as communication catalyst in the modeling process. One major reason behind this statement is that process models at a high abstraction levels should first and foremost be used by domain practitioners as a communicative tool and for these models the principles of good graphical design should be applied. 17.Lewis-Beck, M.S., A. Bryman, and T.F. Liao, The Sage encyclopedia of social science research methods. Vol. 1. 2004: Sage Publications, Inc.