In process mining, declarative process discovery aims to extract declarative process models from event logs. Declare is a declarative process modeling language that provides an extensible set of standard templates for modeling declarative processes imposing temporal constraints on the sequence of process activities. TheeasyDeclare graphical notation has recently been developed to represDeenctlare models. This graphical notation has been proven to enhance human understandability of declarative process discovery results and reduce the cognitive load required to interpret these models compared to the originalDeclare graphical notation. This paper presents EDD, a web-based editor for declarative process models using theeasyDeclare notation. Researchers and practitioners in the process mining ifeld have evaluated the usability of EDD, resulting in a high usability rating. The tool can be used directly within a web browser or by cloning its GitHub repository. As an open-source project, EDD is available to the research community for use and further improvement.
CEUR ceur-ws.org
https://blasilli.github.io/EDD https://github.com/blasilli/EDD https://github.com/blasilli/EDD https://blasilli.github.io/EDD/tutorial.mp4 (T. Catarci)
Declarative process models capture the behavior of a business process by imposing temporal
constraints on the sequence of process activities. In process mining, declarative process
discovery aims to extract these constraints from event logDse.clare is a declarative process modeling
language 1[] that specifies an extensible set of standard templates that a process analyst can use
to model a processD.eclare has proven beneficial in modeling flexible processes where it is
hard to specify in advance a prescriptive order of process activities to be executed, particularly
in healthcare2[
Constraints are concrete instantiations Dofeclare templates.Declare is equipped with formal semantics based on Linear Temporal Logic on Finite Trac7e],sb[ut the use of templates makes model comprehension independent of the logic-based formalization. Analysts can work with the graphical representation of templates while the underlying formulas remain hidden. Graphically, Daeclare model is a diagram in which activities are represented as nodes (labeled rectangles) and constraints as arcs between activities.
Although the semantic aspects of declarative processes have been extensively investigated
[
This paper presents EDDe(asyDeclare Designer), an open-source web application developed in JavaScript. EDD is the first editor that allows working with declarative process models based on the noveleasyDeclare graphical notation. To assess its usability, we conducted a user study involving researchers and practitioners from the process mining field, using the System Usability Scale (SUS) questionnaire11[]. The results of this evaluation were highly positive, with the tool receiving an average grade of A+, indicating that users found EDD to have a high level of usability.
The rest of the paper is organized as follows. Secti2onpresents the visual components and features ofered by EDD to model declarative processes usinegasyDeclare. Section 3 discusses the results of a user study conducted to assess the tool’s usability, and discusses the maturity level of the tool. Finally, Sectio4nconcludes the paper.
EDD is an editor for declarative process models using ethaesyDeclare visual notation. We developedEDD as an open-source web application implemented in JavaScript, and usable with any browser. It can either be used directly in the brows1eorr by cloning the GitHub repository. 1https://blasilli.github.io/EDD
User Interface The user interface has been designed to be simple and eficient. As shown in
ifg. 1, it is composed of three panels: AC)anvas Pane, B) Details Pane, and C) Overview Pane.
The main pane is theCanvas (A) on the left side of the interface, which shows the model. This
pane also ofers an interactive interface for model creation and manipulation. Users can create
activities by double-clicking on an empty area of the canvas. To create binary constraints
between activities, users can mouse over the border of an activity and drag the appearing
arrow to another activity. Although the tool automatically arranges activities and constraints
on the canvas to maintain an organized layout, users can manually adjust the positions by
dragging these elements across the canvas, allowing for a custom layout. Furthermore, the
canvas supports zoom and pan functionalities, allowing users to manage models of various
sizes. On the right side of the interface is thDeetails Pane (B), which provides an overview of
the existing activities and constraints in the model. This pane also shows details about each
activity and constraint and allows the editing of their properties. In the bottom right corner of
the interface, there is theOverview Pane (C). This pane shows the overview of the entire model,
highlighting which part is currently visible in the canvas, following the focus+context visual
paradigm [
Templates EDD supports all the mainDeclare templates for constraints: 1In)it 2) End 3) Absence 4) Existence 5) Choice 6) ExclusiveChoice 7) Response 8) Precedence 9) AlternateResponse
Import and Export To ensure interoperability with other tools, EDD supports the import
and export of models in the d“ecl” file format [
To evaluate the usability oEfDD, we conducted a user study using the System Usability Scale
(SUS) [
A total of 18 users were involved in the user study, both researchers and practitioners in the process mining field. The group is composed of 1 associate professor, 3 assistant professors, 10 PhD students, 2 master students, 1 bachelor student, and 1 process mining consultant. Methodology After some profiling questions to the participants, we briefly recapped the basic concepts of Declarative Process Modeling and illustratedeathseyDeclare visual notation. Then we presentedEDDexplaining its functionalities. After, we allowed the participants to familiarize with the tool for ten minutes. To test the system functionalities, we designed two tasks (T1 and T2) for the users to complete. The first task has been designed to allow users to familiarize with the system by creating a simple model from scratch. The second task involved the editing of an existing and more complex model, which represents a booking procedure for flight tickets. After completing the tasks, participants were asked to fill out the SUS questionnaire. Finally, they were invited to share their comments and thoughts on the tool.
Results The SUS results show a highly positive evaluation of the tool usability, with an average grade oAf+ (87.64) and a median of 90. The standard deviation of= 7.54 suggests relatively low variability in scores, meaning that most users had a similar usability experience with EDD. These values indicate that users found the system very usable. The detailed results are shown in fig. 2.
Feedback Participants provided several positive comments and valuable suggestions for improving the tool. Users generally appreciated its visual design and usability. For example, one e r o c S Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10
TOTAL and constraints in a more organized and visually appealing way. Another planned feature is the ability to export models as images, supporting both SVG and PNG formats.Since we released EDD as an open-source project, contributions from the community are highly encouraged, and any development eforts to extend its functionalities are greatly appreciated.
This paper presentedEDD, an editor for declarative process models, using ethaesyDeclare visual notation. Through a comprehensive usability evaluation involving researchers and practitioners from the process mining fieldE,DD has demonstrated a high level of usability, as evidenced by its average SUS grade of A+. Even if the ofered functionalities have been positively evaluated, there is still space for further development, with opportunities to enhance its capabilities and improve the user experience. For this reasEoDnD, has been released an open-source project; contributions from the community are highly encouraged and appreciated. Acknowledgments. This work has been supported by the H2020 project DataCloud (Grant number 101016835), the Sapienza projects DISPIPE and FOND-AIBPM, and the PNRR MUR project PE0000013-FAIR.