Process Smell 2.0: um catálogo de bad smells para processo de software utilizando BPMN.

Abstract

An explicitly specified software process improves the quality of the generated product. The specification directs the path and steps to be followed so that the developed software reaches the quality desired by the project. The software process evolves along with the needs of the organization and the professionals who use it. It needs to be continuously monitored and evaluated to maintain its quality. To evaluate a software process, simula- tions or data obtained after execution are commonly used. However, in both cases, it is difficult to predict problems in the software process in a given development project before one or more executions. Therefore, Process Smells were proposed to enable problems in the design of software processes to be identified before they were executed. The presence of a Process Smell can negatively impact the quality of the Process, thus affecting the quality of the software product. Initially, Process Smells were specified using the Soft- ware Process Specification Metamodel (SPEM) notation. Although SPEM is a language for the domain of process modeling, Business Process Model and Notation (BPMN) has been widely used for processes in general, with high popularity both in the market and academia. In this sense, the specification of a new catalog can allow Process smells to be understood independently of the language domain. Still, in the context of process im- provement, just like SPEM, although BPMN is used to improve the understanding of a process, the specification of a process can be done inappropriately, injuring the necessary factors for the quality of a process of software. In this context, this research aimed to specify a Process Smells catalog to support the identification of anomalies in software processes specified with BPMN based on the (SANTOS; MACIEL; SANT’ANNA, 2018) proposal. Initially, Process Smells were specified in BPMN, thus originating a new ca- talog with 8 Process Smells, Process Smells 2.0. Thirty-two professionals evaluated the specification through an interview study, which indicated that these professionals accep- ted the proposed new Process Smells catalog. The results obtained in this study made it possible to verify that the proposal to identify smells regardless of the language domain makes sense. The second stage of the study proved to be more concurrent with SPEM. The catalog specification is expected to support the identification of Process Smells in software processes modeled using BPMN to indicate the points where the Process can be improved, even before its first execution, avoiding problems that negatively affect process quality attributes. Additionally, the new catalog is expected to expand the possibility of using Process Smells.