January 30, 2023 | 02:00 pm (CET) | Room: S.0.05 | | Alpen-Adria-Universität Klagenfurt
Dr.Phil. Volodymyr Shekhovtsov | Department of Informatics Systems
Abstract: Conceptual modeling activities need to integrate software and data quality at the same level of abstraction as the functionality of the system or the data itself (a conceptualization of quality). This need is related to the fact that looking at the problem domain to obtain its conceptual model only from the view of the system’s functionality or the structure of the associated data restricts the analyst and can be the source of mistakes. The reason for such mistakes is that many quality-related issues become only evident when the system is put into use; on the other hand, such mistakes are parts of the early design decisions and can be difficult and costly to fix. This thesis summarizes several selected publications of a research program towards elaborating conceptualizations of software and data quality on different stages of the software development process and for different domains. This program addresses five research challenges. The first challenge is related to the need for the unified classification of the quality conceptualization techniques based on the agreed-upon concept of the quality itself. Without such a classification it is difficult for analysts to decide which quality conceptualization solution is better suited for their problem. This is addressed in the thesis by proposing the classification of such techniques, which can serve as a foundation for the evaluation framework for quality conceptualizations; the rest of the thesis is aligned with this classification. The second challenge is positioned at the intersection of the research fields of static code analysis and requirement engineering. It is related to the problem that the rationale of applying code-checking rules in static code analysis is not captured explicitly, which leads to the problems of rule reuse in similar development contexts. In this thesis, it is proposed to trace possible sources of such rules back to design decisions and quality requirements and to conceptualize the qualityrelated rationale information along with specific code-checking rules to serve as a schema for a rule repository. The third challenge is positioned in the research field of quality-aware requirements engineering. It is related to the problem that more attention should be paid to conceptualizing software quality requirements before performing design-time activities. It is addressed in the thesis by the quality requirements conceptualization activity positioned between the activities of quality requirements elicitation and conceptual design. The results of this activity can be used at different stages of the software process, capturing the quality requirements semantics in a way that can be easily understood and verified by the system users and can be mapped into different design notations. The fourth challenge is positioned in the research field of software process improvement. It is related to the problem that organizing quality-related interaction between business stakeholders and software developers is difficult as they need a common set of concepts. In this thesis, this challenge is addressed by three approaches. First, the thesis proposes to conceptualize the stakeholder interaction process on two levels: a coarse-grained level defining the set of generic quality-related activities and the conditions of launching these activities and a fine-grained level describing specific interaction steps in detail. Second, the thesis proposes to conceptualize such a process as a process of quality view harmonization on three levels: terminology harmonization (agreeing on the common quality-related terminology), view harmonization (agreeing on the sets of objects and the types of their qualities to be assessed, as well as the assessment procedures), and quality harmonization (agreeing on the evaluation schemes and the specific qualities to be used). Third, the thesis introduces a model-centered software solution that facilitates quality-related stakeholder communication in the software development process based on these concepts. The fifth challenge is positioned in the research field of data modeling and management for a specific domain (biobanking). It is related to the problem that the quality of data and metadata in biobanks is not properly defined on the conceptual level and that such conceptualizations are not used to facilitate the biobank search. In this thesis, it is again addressed by three approaches. First, the thesis conceptualizes the data item quality in biobanks and defines a set of data item quality characteristics and metrics. Second, the thesis proposes a definition and a conceptualization for the metadata in biobanks, which covers stored data item quality values, and defines a set of metadata quality characteristics and metrics. Third, the thesis proposes an architecture to support researchers in identifying relevant collections of material and data with documented quality for their research projects while observing strict privacy rules. It is based on the conceptualization of the biobank metadata (including its data item quality values) to serve as a schema for a metadata repository to be queried in a search for collections.
Bio: Volodymyr Shekhovtsov is a postdoctoral researcher at the Department of Informatics Systems (ISYS), University of Klagenfurt. He received his PhD degree in Computerized Control Systems and Progressive Information Technology in 1998 from the National Technical University “Kharkiv Polytechnic Institute”, Kharkiv, Ukraine. As a researcher, since 2012 he participated in total in 5 research projects at the University of Klagenfurt.
He co-authored more than 45 papers, published in journals such as Applied Sciences, Internet of Things, International Journal of Web Information Systems, and gave 27 talks at venues such as ACM/IEEE International Conference on Model Driven Engineering Languages and Systems (MODELS), International Working Conference on Requirements Engineering (REFSQ), Dagstuhl seminars “Next Generation Domain Specific Conceptual Modeling: Principles and Methods”, and “The Evolution of Conceptual Modeling”. As a teacher, since 1994 he was responsible for 14 university lecture and lab courses in both Ukraine and Austria. His research interests include software, data, and metadata quality, Internet of Things and Industry 4.0, behavior modeling and prediction, methods and tools for information system modeling, analysis, and design, source code analysis and quality, and applying decision theory and optimization techniques to software engineering problems.