Wednesday, 23 November 2011

Summary: Modeling languages for business processes and business rules: A representational analysis

M. zur Muehlen and M. Indulska. Modeling languages for business processes and business rules: A representational analysis. Information Systems, 35(4):379–390, Elsevier, 2010.

Process modeling languages and rules modeling languages are both used to document organizational policies and procedures. While process modeling languages typically describe a procedural sequence of activities, including decisions and concurrency, rules modeling languages often rely on a declarative description of conditions, and constraints that need to be followed. Understanding the relationship between the two languages types would allow organizations to maximize synergies, avoid content duplication, and thus reduce their overall modeling effort. Rules modeling languages has received less attention than process modeling languages. This paper is to investigate the capabilities of four rule modeling languages, namely: Simple Rule Markup Language (SRML), Semantic Web Rules Language (SWRL), Production Rule Representation (PRR), and Semantics of Business Vocabulary and Business Rules (SBVR). The paper looks into the representational capabilities of these languages, and if these capabilities are complementary or substitutive to those of process modeling languages. The paper also looked into integration with respect to four process modeling languages: Colored Petri nets (CPN), Event-driven Process Chains (EPC), Integrated DEFinition methodology (IDEF), and Business Process Modeling Notation (BPMN).

Business rules can be categorized into 5 different types: Integrity rules (the acceptable relationship between elements), Derivation rules (used to infer new facts based on known facts), Reaction rules (alternative action rules), Production rules (condition, action rules), and Transformation rules (restrict the state of changes). Process modeling languages can be: Activity centric, Process object centric, and Resource centric. Process languages appear as “graph-based languages (e.g. BPMN, EPC), net-based languages (e.g. Petri nets, flow nets), and workflow programming languages (e.g. Business Process Execution Language (BPEL))”. Work on integration of these two types of languages appeared after introducing rule modeling concept, but no theory-based evaluation of the usefulness of these combinations has been conducted so far. There is a need to augment existing research on the representational capabilities of process modeling languages with matching evaluations on the rule modeling side. Furthermore, there is a need to provide practitioners with guidance as to which rule modeling language and which combination of rule and process modeling languages will allow them to capture the most real-world details using language primitives.

The authors used Bunge-Wand-Weber (BWW) ontology to develop criteria construct. Based on these criteria the study showed that in general SRML would satisfy more constructs than other rule modeling language. Moreover, in combination with a process modeling language it was shown that the combination of SRML and BPMN would provide the best representation capability, thus the best combination of the eight examined languages.  The study then went a step further considering combining rule modeling language with rule specification beside the process modeling language. The paper considered two combinations SBVR with PRR and SBVR with SWRL, then compared these two combination in respect to all four process modeling languages. The study shows that Even thought the pare SBVR with PRR is a strong combination with BPMN than any other pare, but it is still better to use SRML with BPMN as it has better representation capabilities.

Zur Muehlen and Indulska in this paper compared between four ‘rule modeling languages’ (SRML, SWRL, PRR, and SBVR) in respect to integration with four different process modeling languages (CPN, EPC, IDEF, and BPMN) based on representation capabilities criteria. The study showed that the best combination is to have SRML as the rule language and combine it with BPMN as a process modeling language; this combination satisfies more representation capabilities than any other combination. The study was limited to only four rule languages and four process modeling languages (non of which is an automation language). And the study was also limited to comparing in regard to representation capabilities. So based on this study, if representation capability was the important aspect one is looking for, and the choice was between these four rule languages and these four process modeling languages, one showed chose SRML and BPMN.


No comments:

Post a Comment