Gömülü yazılım geliştirme projelerinde gözlemlenen modelleme yaklaşımı kalıp ve kültürleri.


Tezin Türü: Doktora

Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Türkiye

Tezin Onay Tarihi: 2018

Tezin Dili: İngilizce

Öğrenci: Deniz Akdur

Danışman: ONUR DEMİRÖRS

Özet:

Due to their multiple constraints across different dimensions of performance and quality, the analysis, design, implementation and testing of software-intensive embedded systems are not trivial, which makes their development more challenging. To cope with these growing complexities, modeling is a widely used approach in this industry. However, the modeling approaches in embedded software vary since the characteristics of diagram development and usage (e.g., purpose, modeling rigor, medium type used, modeling stakeholder profile, target sector, etc.) differ among systems as well as among sectors. At one extreme, some stakeholders use software modeling informally, where they sketch the diagrams on a paper in order to communicate with other stakeholders. At the other extreme, modeling turns into programming with automated generation of some software development life cycle (SDLC) artifacts (i.e., code, documentation or test driver). Moreover, different stakeholders in the same software development project can use diagrams for different purposes within different SDLC phases. This PhD dissertation identifies and defines the modeling patterns and cultures of embedded software development projects. To achieve this, it firstly figures out the current state-of-practice of modeling to investigate the relations between the characteristics of diagram development and usage and also the significant parameters to identify modeling patterns. After identifying the modeling patterns and cultures, this study proposes a characterization model. This model not only identifies and defines modeling patterns and cultures of the modeling stakeholder in embedded software development projects, but also gives recommendations for commonsense modeling practices. Finally, this proposed model is validated by multiple case studies.