30 pages

Seven Process Modeling Guidelines (7PMG)

of 30
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Seven Process Modeling Guidelines (7PMG)
  Seven Process Modeling Guidelines (7PMG) J. Mendling a , ∗ H.A. Reijers b W.M.P. van der Aalst b a Humboldt University, Unter den Linden 6, 10099 Berlin, Germany  b Eindhoven University of Technology, P.O. Box 513, 5600 MB Eindhoven,The Netherlands  Abstract Business process modeling is heavily applied in practice, but important quality issueshave not been addressed thoroughly by research. A notorious problem is the lowlevel of modeling competence that many casual modelers in process documentationprojects have. Existing approaches towards model quality might be of potentialbenefit, but they suffer from at least one of the following problems. On the one hand,frameworks like SEQUAL and the Guidelines of Modeling are either too abstractto be applicable for novices and non-experts in practice. On the other hand, thereare collections of pragmatic hints that lack a sound research foundation. In thispaper, we analyze existing research on relationships between model structure onthe one hand and error probability and understanding on the other hand. As asynthesis we propose a set of seven process modeling guidelines ( 7PMG ). Each of these guidelines builds on strong empirical insights, yet they are formulated to beintuitive to practitioners. Furthermore, we analyze how the guidelines are prioritizedby industry experts. In this regard, the seven guidelines have the potential to serveas an important tool of knowledge transfer from academia into modeling practice. Key words:  Business Process Modeling, Model Quality, Guidelines ∗ Corresponding author Email addresses:  (J. Mendling),  (H.A. Reijers),  (W.M.P. van der Preprint submitted to Elsevier July 29, 2009   1 Introduction Since the 1970s and 1980s, conceptual modeling is a major research area in theIS field. The main motivation to engage in conceptual modeling is to reducethe chances on developing faulty requirements in the early phases of systemdevelopment [1]. A recent empirical study has shown that  business processes  have become the central objects in many conceptual modeling efforts, support their documentation, improvement and automated enactment [2].This development can be explained by an increased focus of enterprises onthose same business processes: They are perceived as the most relevant entitiesto be managed towards enhanced organizational performance [3].Usability is an important quality issue of process documentations [4]. As un-derstanding the process is a crucial task in any process analysis technique[5], also the process model itself should be intuitive and easy to comprehend.Process modeling tools, like  ARIS  and  Casewise , have greatly eased the stan-dardization, storage, and sharing of diagrams of process. Many enterpriseshave adopted such tools as they are perceived as much better alternativesto the use of pen and paper, or even general graphical drawing tools, e.g.Microsoft’s  Visio  or  Powerpoint . But despite the support that is providedby tools, users get hardly any support in creating process models that busi-ness professionals can easily analyze and understand. Adequate guidance isof particular importance as large projects on process documentation heavilyrely on novices and non-expert modelers [6]. To appreciate the impact of amodel that is difficult to assess, it should be realized that in the execution of a single project dozens, hundreds or even thousands of process models maybe developed [7,8]. This clarifies why a process model which is immediatelyusable towards its purpose is of great economic benefit.Even though some theoretical frameworks and guidelines are available in thearea of process modeling, for instance SEQUAL or the Guidelines of Modeling Aalst). 2  [9,10], these typically require a certain level of modeling competence. Theydistinguish the major quality categories, but remain too abstract to be directlyapplicable by non-experts. In other words, such guidelines are hardly relatedto the concrete actions that process modelers undertake in capturing e.g. thesteps and actors in a process. More practice-oriented and -inspired guidelinesare available too, see e.g. [11]. The problem behind such guidelines is thathardly any empirical support is provided for them and, if so, it is anecdotic atbest. From a research perspective, it can be noted that much of the existingwork into process modeling does not focus on providing modeling supporteither. Rather the interest is with the more formal side of process modeling,see e.g. [12,13].This paper seeks to support the builders of business process models by pro-viding them with a set of seven modeling guidelines, called  7PMG . This set isthought to be helpful in guiding users towards improving the quality of theirmodels, in the sense that these are likely (1) to become comprehensible tovarious stakeholders and (2) to contain few syntactical errors. Each of theseguidelines gives directions on how a process model can be improved and whichalternative of a set of behavior-equivalent representations should be preferred.As such, the application of   7PMG  will improve the efficiency of projects withinenterprises that rely on the use of this particular type of conceptual models.The novelty of the presented work is that all the guidelines of   7PMG  build onsound scientific insights that have emerged over the past years into the re-lationship between process modeling styles on the one hand and both modelunderstanding and error-proneness on the other. As of yet, these insights havenot been synthesized into guidelines that are clear, practically applicable, andwell-motivated. In this way,  7PMG  not only contrasts other frameworks thathave been criticized for lack of empirical foundation [14] but it also offersguidance that practitioners can apply in their business-process centered ini-tiatives straightaway. Finally,  7PMG  provides a baseline for further researchinto process modeling to extend this set and to develop advanced tool support3  to facilitate modeling activities.Against this background, the paper is organized as follows. Section 2 outlinesthe background of our research, namely different approaches towards pro-cess model quality. Section 3 presents the seven process modeling guidelines(7PMG) that we synthesize from prior research. Section 4 presents indicationson how the guidelines should be prioritized. Section 5 contributes a discussionof the limitations and merits of these guidelines. Section 6 closes the paperwith a conclusion. 2 Background The roots of process modeling can be traced back to the early 20th century asa tool for organizational design (see [15]). It gained some attention as a sub- ject of information systems research with the invention of office automationsystems in the 1970s and 1980s (see [16,17]). The business process reengineer-ing boom of the early 1990s contributed to a consolidation of the field and thedefinition of process modeling languages such as Event-driven Process Chains(EPCs) [18]. At the core of such languages is a representation of control flowbetween different activities, which can be extended with different perspectivessuch as organizational responsibilities or object flow [19,20,21,22]. There aremainly four streams of work that discuss guidelines and quality issues for suchconceptual process models: top-down quality frameworks, bottom-up metricsrelated to quality aspects, empirical surveys related to modeling techniques,and pragmatic guidelines.One prominent  top-down quality framework   is the SEQUAL framework [9,23].It builds on semiotic theory and defines several quality aspects based on rela-tionships between a model, a body of knowledge, a domain, a modeling lan-guage, and the activities of learning, taking action, and modeling. In essence,syntactic quality relates to model and modeling language; semantic quality to4  model, domain, and knowledge; and pragmatic quality relates to model andmodeling and its ability to enable learning and action. Although the frame-work does not provide an operational definition of how to determine the vari-ous degrees of quality, it has been found useful for business process modelingin experiments [24]. The Guidelines of Modeling (GoM) [10] define an alterna-tive quality framework that is inspired by general accounting principles. Theguidelines include the six principles of correctness, clarity, relevance, com-parability, economic efficiency, and systematic design. This framework wasoperationalized for EPCs and also tested in experiments [10]. Furthermore,there are authors (e.g. [14]) advocating a specification of a quality frameworkfor conceptual modeling in compliance with the ISO 9126 standard [25] forsoftware quality. A respective adaptation to business process modeling is re-ported in [26]. Although these works offer a good insight into quality issuesof a model, they do not provide a straightforward method for implementationin a modeling project. A major problem in these projects is the sheer numberof models (often more than thousand) and the low level of competence thatcasual modelers have [6]. Therefore, easy-to-follow guidelines are needed inpractice.For these reasons, several recent works has tried to approach this problemby studying  bottom-up metrics related to quality aspects   of process models.This area is still fragmented and authors have partially worked isolated fromeach other (see for an overview [15]). Several of these contributions are the-oretical without empirical validation. Most authors doing experiments focuson the relationship between metrics and quality aspects:  Canfora et al.  studythe connection mainly between count metrics – for example, the number of tasks or splits – and maintainability of software process models [27];  Cardoso validates the correlation between control flow complexity and perceived com-plexity [28]; and  Mendling et al.  use metrics to predict control flow errors suchas deadlocks in process models [29,30]. The results reveal that an increase insize of a model appears to have a negative impact on quality. Further workby  Mendling, Reijers, et al.  investigate the connection between metrics and5
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks