Um catálogo de extensões da linguagem de modelagem Kaos

Abstract

Knowledge Acquisition Automated Specification (KAOS) is a modeling language whose main focus is to explicitly represent the goals of systems, the conflicts and obstacles between these goals, which objects are responsible for satisfying them, and the consequences of the outcome of the interaction between them. The KAOS modeling language has been extended to various domains/application areas such as system security, adptive systems, and aspects. Identifying a KAOS extension or its constructs based on its characteristics, such as application area, for example, can be a complex task that requires a large amount of time and has a propensity to be unsuccessful when performed without proper support. We then noticed the need to create a catalog to store the extensions of the KAOS modeling language, in order to contribute to the identification of these extensions and their constructors. The catalog proposed in this paper aims to facilitate the identification of KAOS extensions and their constructors based on their characteristics, such as application area or type of constructor. It is based on the results of a Systematic Literature Review (SLR) performed in 2019 and updated in 2021 that identified and analyzed existing KAOS extensions and the results point to 42 identified extensions and 183 proposed builders. The catalog was further evaluated through an exploratory study based on survey that aimed to analyze what effect the use of the catalog has on the identification of KAOS extensions and their builders. In the evaluation a questionnaire was proposed to be answered by the participants of the study, where they would have to use the catalog to obtain the answers to the questionnaire with questions related to the identification of the extensions and their builders. As a result we have the catalog available at the link https://kaos-catalogue-d9e49.web.app/. The evaluation results show that at least 90% of the participants succeeded in answering at least twelve (12) of the fifteen (15) proposed questions, and that in approximately 75% of the questions the participants found it very easy or easy to get the answer using the catalog. Finally, this paper presents discussions and future work.