The software development process (SDP) is complex, dynamic, and knowledge-intensive by its nature. Therefore, we need a suitable approach to model it. Some authors have proposed different business process (BP) modelling approaches, like variant-based, rule-based, declarative, workflow-based, and case handling for various BP, such as dynamic, knowledgeintensive, goal-oriented, context-sensitive BP. This paper aims to investigate which BP modelling approach is most suitable for SDP modelling. Consequently, a set of necessary criteria is defined, and several BP modelling approaches are assessed. The obtained results show that the most suitable BP modelling approach for SDP is case handling, as it focuses on what can be done to achieve a predefined process goal rather than using a predefined workflow of the process.
This study contributes to SDP research by investigating the BP modelling approaches suitable for SDP modelling. By choosing the most suitable BP modelling approach, we could get the more precise SDP model.
Note that in the practice there is plenty of software management approaches, such as Agile, Waterfall, Scrum, Kanban, and all others [23]. However, in this paper we concentrate on SDP modelling but not on management activities.
The rest of this paper is structured as follows. Section 2 introduces related works on the comparison of BP modelling approaches. Section 3 presents the BP types and modelling approaches. Section 4 introduces SDP and main requirements for its modelling Section 5 shows the comparison of various BP modelling approaches. Finally, Section 6 concludes the paper.
Various BP modelling approaches have been offered in the literature, based on the user’s needs and a BP type. In [11] authors analyze the declarative approach for BP modelling, which focuses on explicitly stating the business issues which control BP.
The authors in [18] compare two BP models that are based on two approaches: graphs and rules. The strengths and weaknesses of two approaches are examined in terms of: 1) expressiveness [18] (i.e., the ability of a process modelling language to communicate precise process requirements that represent the aim of process model-ling and execution determines its expressive power); 2) flexibility (i.e., the ability of a BP to function on a loosely or partially described model [18], with the full specification being made at runtime); 3) adaptability [18] (i.e., the ability to respond to unusual circumstances); 4) dynamism [18] (i.e., the workflow process' flexibility to adjust as the BP evolves); and 5) complexity [18] (i.e., the difficulties of modelling, analyzing, and deploying a process model, as well as support for dynamic and evolving BP).
Authors in [16] also compare three different modelling approaches: BPMN2.0, IDEF0 and IDEF3. The main goal of this research is to understand main differences and weakness of each modelling approach based on its usability, representation, communication and alignment, readability, and creativity. It is based on survey.
Other authors compare BP in a particular context. Authors in [
As can be summarized, in the literature we can find different comparisons of BP modelling
approaches. However, they are limited according to some features, like declarative BP [11], graphs or
rules [18], or a particular context [
Because of the changing environment, BP becomes dynamic. Therefore, a uniform BP model cannot be defined for a dynamic business process (DBP) in all cases [14], i.e., it could adapt to a continuously moving and changing environment. Dynamic and flexible systems offer advantages for businesses in addressing dynamic uncertain factors and implementing DBP [14]. DBP modelling and simulations allows the process operator to see the impact of possible solutions or changes on the process end, identify problems and preventive measures to these problems at a process in-stance run-time.
Processes, where knowledge is a key characteristic, are called knowledge intensive processes (KIP)
[
Goal-oriented process modelling is driven by the need to ensure congruence of BP and decisions
with the values and vision of the business while meeting continuous demands for increased business
productivity [
Context-sensitive processes use context to provide more relevant information to support users while performing their tasks. A context is any information that can be used to characterize the situation in which something exists or occurs [27]. Context-sensitive processes demand that designers consider new aspects and challenges in comparison to traditional approaches [26]. A context-sensitive process can adapt its instances at run-time to the changing context, defined by variables containing all relevant information about the design and execution of BP [32].
According to [33], authors model BP using families of BP variants, where a family of BP variants is presented via a single model, from which each variant can be de-rived through certain transformations (i.e., adding or deleting fragments) of the mod-el.
A BP model is called rule-based if the logic of its control flow, data flow and re-source allocation is declaratively expressed by means of business rules (BR). A rule-based process model represents BP as a description of its state space and a set of BR that constrain the valid movements in that state space. Rule-based process modelling allows to include a lot of useful information that otherwise would remain implicit in procedural process models. The advantages of rule-based process modelling manifest themselves during both the design and enactment phase of the BP management (BPM) lifecycle.
According to some [9] declarative models specify what should be done without specifying how it should be done. These models specify a set of constraints, BR, event conditions, or other expressions that define the properties of, and dependencies between, activities in BP [11]. So, BP instances are constructed according to predefined BR and do not violate them. These BR restrict the final execution path without defining the process model. In [9], the author states that declarative artifact-centric process models, which present knowledge-intensive processes, use BR that define how knowledge experts can make progress in a process. However, in many business situations knowledge experts must deal with uncertainty, which is modelled using fuzzy logic, and extends declarative artefact-centric processes models with fuzzy logic [9].
In a workflow-based BP, the business activity consists of atomic tasks, data types, and human actors [31]. A workflow-based modelling approach is focused on a static sequence of those tasks and their workflow scenarios.
Case management is a model for enabling flexible, knowledge-intensive BP. It is heavily reliant on
data. Case handling, unlike workflow management, focuses on what can be done to achieve a business
goal rather than using established process control structures describing what should be done during a
workflow process. When it comes to case handling, the knowledge worker in charge of a certain case
actively decides how that case's goal will be met [
In this paper, we are interested in SDP modelling. Therefore, in this section, we define main requirements for SDP modelling as follows:
1. Dynamics: the developing SDP model should be dynamic, i.e., its model and in-stances should support structural changes (i.e., there is no predefined sequence of activities) at SDP instance run-time according to its context and rules, and that can be implemented with minimal delay (adopted from [14]).
2. Visible context: in the SDP model should be possible to define and observe the context or a business environment of the SDP instance [19].
3. Knowledge-intensive: in the SDP model, the knowledge worker should be responsible for the sequence of activities and how the goal of SDP is reached [30].
4. A single process instance: the focus should be placed on a single process instance rather than on a workflow [17].
5. Parallel assignment or skip role [20]: a. The SDP model should support parallel modelling of activities in a SDP.
b. The SDP model should support parallel assignment of activities to one employee/resource in the same and/or in the separate process cases/instances.
6. Risk management: In the SDP model, it should be the possibility to assign risks for activities or an SDP instance/case [29].
Those requirements become the main comparison attributes for modelling approaches.
forA family of process variants
aBcPtMivNity, diagUraMmLCMMN
Bizagi Modeler TCraimsoutencdha,
tBaPbMleNs/,mDeactrisixioensConDec *
extensions (SBPM [15]),
* A rule-based approach for process modelling has no separate notation. Existing notations, like DMN, Semantics of Business Vocabulary and Business Rules (SBVR), BPMN, can be used to model rules in processes.
** The relationships between tasks in ConDec are understood as constraints, which represent policies (or business rules) [24].
From this table, the most suitable modelling approach is case handling as it meets all the comparison attributes. It could be used for DBP (Requirement 1). In this approach, the context can be visible (Requirement 2). There is nothing defined on how to specify a context in a declarative-based and a variant-based approaches. Since the primary purpose of those approaches is not related to the context, it is therefore not described in the analyzed papers. A knowledge-intensive feature is occurring in all analyzed approaches, except a workflow-based approach (Requirement 3). A case handling approach is also suitable because it focuses on a whole case (Requirement 4) and allows us to model activities and task in parallel (Requirement 5). The same functionality could be seen in workflow-based and variant-based approaches as well. What is more, risk management could be adaptable to the case handling approach (Requirement 6), where a case can depict the risk management. All modelling approaches have notations and/or modelling languages (Table 1, rows 8-9). Some notations and tools could be applicable for few modelling approaches as they have similarities in characteristics.
At the beginning of this paper, we have defined three research questions to answer. Here we present the answers. The analysis of the BP modelling approaches shows that BP is a collection of main business activities and tasks that should be arranged in a specific sequence. It is significant to understand how BP are working in a particular domain area. For this reason, BP modelling has become crucial, and various BP types have their modelling approaches.
The analysis of BP types shows that the main of them are as follows: dynamic, knowledge-intensive, goal-oriented, context-sensitive. The found BP modelling approaches are the following: variant-based, rule-based, declarative, workflow-based, and case handling. Those found BP types and their modelling approaches are focused on a particular entity in a process, like activity, task, context, etc. Moreover, those BP modelling approaches view BP from a particular perspective, such as variants, rules, contexts, cases, workflows, etc.
The analysis of the related works allows us to formulate and specify six requirements of the SDP. Those SDP requirements are based on the main SDP characteristics. Based on the proposed SDP requirements, a comparison of the found BP model-ling approaches is presented.
The obtained comparison results show and teach us that the most suitable BP modelling approach for SDP is case handling, as it meets all defined requirements for SDP. Moreover, it focuses on what can be done to achieve a predefined process goal rather than using a predefined workflow.
In future works, we plan to use the case-handling approach to model and simulate SDP. Moreover, since SDP is knowledge-intensive and knowledge workers should deal with uncertainty, we will apply fuzzy set theory for knowledge in SDP modelling. 7. References [8] Decker G, Dijkman R, Dumas M, García-Bañuelos L (2010) The Business Process Modeling Notation. Mod Bus Process Autom YAWL its Support Environ 347–368. doi: 10.1007/978-3-64203121-2_13 [9] Eshuis R (2018) Feature-Oriented Composition of Declarative Artifact-Centric Process Models.
Lect Notes Comput Sci (including Subser Lect Notes Artif Intell Lect Notes Bioinformatics) 11080 LNCS:66–82. doi: 10.1007/978-3-319-98648-7_5 [10] Goedertier S, Haesen R, Vanthienen J (2008) Rule-based business process modelling and enactment. Int J Bus Process Integr Manag 3:194–207. doi: 10.1504/IJBPIM.2008.023219 [11] Goedertier S, Vanthienen J, Caron F (2014) Declarative business process modelling: principles and modelling languages. http://dx.doi.org/101080/175175752013830340 9:161–185. doi: 10.1080/17517575.2013.830340 [12] Grambow G, Oberhauser R, Reichert M Knowledge Provisioning A Context-sensitive Processoriented Approach Applied to Software Engineering Environments. doi: 10.5220/0004083005060515 [13] Herrmann C, Kurz M (2011) Adaptive Case Management: Supporting Knowledge Intensive Processes with IT Systems. Commun Comput Inf Sci 213 CCIS:80–97. doi: 10.1007/978-3-64223471-2_6 [14] Kalibatiene D, Vasilecas O (2021) A survey on modeling dynamic business processes. PeerJ
Comput Sci 7:1–52. doi: 10.7717/PEERJ-CS.609/SUPP-1 [15] Kang G, Yang L, Zhang L, Liu J, Wen Y (2021) Organization-driven business process configurable modeling for spatial variability. China Commun 18:229–242. doi: 10.23919/JCC.2021.11.016 [16] Kelani KT (2018) MODELLING TECHNIQUES: comparison between BPMN2. 0 and IDEF (0&3) (Doctoral dissertation) [17] Lemos AM, Sabino CC, Lima RMF, Oliveira CAL (2011) Using process mining in software development process management: A case study. Conf Proc - IEEE Int Conf Syst Man Cybern 1181–1186. doi: 10.1109/ICSMC.2011.6083858 [18] Lu R, Sadiq S (2007) A Survey of Comparative Business Process Modeling Approaches. Lect Notes Comput Sci (including Subser Lect Notes Artif Intell Lect Notes Bioinformatics) 4439 LNCS:82–94. doi: 10.1007/978-3-540-72035-5_7 [19] MacCormack A, Verganti R (2003) Managing the Sources of Uncertainty: Matching Process and Context in Software Development. J Prod Innov Manag 20:217–232. doi: 10.1111/15405885.2003004 [20] Martin RH, Raffo D (2000) SOFTWARE PROCESS IMPROVEMENT AND PRACTICE A Model of the Software Development Process Using Both Continuous and Discrete Models Research Section. Softw Process Improv Pr 5:147–157. doi: 10.1002/1099-1670 [21] Pesic M, Schonenberg H, van der Aalst WMP (2008) DECLARE: Full Support for Loosely
Structured Processes. 287–287. doi: 10.1109/EDOC.2007.14 [22] La Rosa M, Van Der Aalst WMP, Dumas M, Milani FP (2017) Business Process Variability
Modeling. ACM Comput Surv 50. doi: 10.1145/3041957 [23] Stuart CE (2020) Cynefin-Framework as a Guide to Agile Leadership Which Project Management
Method for Which Type of Project? - Waterfall, Scrum, Kanban? [24] Tiam RT, Seriai A-D, Michel R An Operational Model of Variable Business Process. doi: 10.5220/0005372301620172 [25] Tiam RT, Seriai AD, Michel R (2015) VASCO: Variability Specification in Business Process
Models. Lect Notes Bus Inf Process 241:612–632. doi: 10.1007/978-3-319-29133-8_30 [26] Vieira V, Tedesco P, Salgado AC (2009) A process for the design of context-sensitive systems.
Proc 2009 13th Int Conf Comput Support Coop Work Des CSCWD 2009 143–148. doi: 10.1109/CSCWD.2009.4968049 [27] Vieira V, Tedesco P, Salgado AC (2011) Designing context-sensitive systems: An integrated approach. Expert Syst Appl 38:1119–1138. doi: 10.1016/J.ESWA.2010.05.006 [28] Westergaard M, Maggi FM Declare: A Tool Suite for Declarative Workflow Modeling and
Enactment [29] Ye T (2008) A study of software development project risk management. Proc - 2008 Int Semin
Futur Inf Technol Manag Eng FITME 2008 309–312. doi: 10.1109/FITME.2008.125 [30] Ye Y (2006) Supporting Software Development as Knowledge-Intensive and Collaborative Activity. Proc 2006 Int Work Work Interdiscip Softw Eng Res - WISER ’06 15–22. doi: 10.1145/1137661 [31] Zhang Y, Perry DE (2014) A data-centric approach to optimize time in workflow-based business process. Proc - 2014 IEEE Int Conf Serv Comput SCC 2014 709–716. doi: 10.1109/SCC.2014.129 [32] Zhou J, Gilman E, Palola J, Riekki J, Ylianttila M, Sun J (2011) Context-aware pervasive service composition and its implementation. Pers Ubiquitous Comput 15:291–303. doi: 10.1007/S00779010-0333-5/FIGURES/8 [33] (2007) Business Process Management Architectures. Bus Process Manag 305–343. doi: 10.1007/978-3-540-73522-9_7