Source code is a part of enterprises intellectual value. Today, the amount of source code in projects grows into millions of lines of code. This makes gaining overview and insights about software development projects barely impossible. Even a Smartphone runs at 12 million lines of code. Additionally, more and more microchip driven devices hit the market every day and the amount of source code is continuously growing. Today, we cannot estimate if this development will ever hit a limit. Therefore, reducing development times, improving software reuse and enhancing quality control of software is of crucial interest.
Current research shows new approaches for software development and provides custom source code analysis infrastructures. In order to apply such code analysis infrastructures, new and custom academic tools without professional vendor support would need to be introduced in enterprises. Therefore, a new source code analysis approach is needed that integrates into the current system landscape of a company.
In this work, we introduce the Hypermodelling approach and leverage already in enterprises available Data Warehouse technology for software manufacturing activities. Complementary, we provide concrete technical details about an implementation of Hypermodelling for Java. Different usage scenarios show advancements to related research and practice. In special, we present insights how dependencies and other relations of code structure can be revealed with little effort. Additionally, we report about the reasonable effort to create management dashboards for code structure. Furthermore, we test Hypermodelling to plan revision updates and to compute advanced information out of code. First demonstrations to mine developer knowledge, to compute developer-style issue correlations and to report cohesion indicators underline the potential of the approach. Finally, a web and development environment based code search engine and a code clone recommendation scenario show also advancements to related work in the corresponding areas.
Now, Data Warehouse technology can be used for software engineering research and practice. Among others, our current results can be used to carry out the following business cases: Code search, multi-dimensional and statistic code analysis, code structure audits with automated reporting, live code clone detection and recommendation, data integration and revision update progress control.
Future research and implementations can focus on specialized scenarios, Hypermodelling based source code analysis or to the advancement of Hypermodelling itself. Furthermore, plenty of synergies with related work identify additional research trails and future application scenarios.
Outline:
Introduction and Motivation
In Chapter 2, we introduce the general concepts and research that motivates, supports and represents the foundation of our work. We focus on the paradigm of separation of concerns and give examples where object-oriented programming does not meet the demands of the principle.
In Chapter 3, we get into the advancements of multi-dimensional separation of concerns and a view based concern separation technique. Complementing, we present the area of domain specific languages and show a modeling based approach that offers multi-dimensional navigation of a software system. Lastly, we introduce the area of source code analysis.
Chapter 4 leaves the area of software engineering and introduces Data Warehousing. We explain the different used data structures and give application examples how Data Warehouses handle multidimensional data.
Hypermodelling and Implementation
In Chapter 5, we tie the essence of the prior chapters together and derive Hypermodelling. This is the key contribution of our work. First, we give a brief insight how we utilize Data Warehouse technology to express concern associations of software ...
Current research shows new approaches for software development and provides custom source code analysis infrastructures. In order to apply such code analysis infrastructures, new and custom academic tools without professional vendor support would need to be introduced in enterprises. Therefore, a new source code analysis approach is needed that integrates into the current system landscape of a company.
In this work, we introduce the Hypermodelling approach and leverage already in enterprises available Data Warehouse technology for software manufacturing activities. Complementary, we provide concrete technical details about an implementation of Hypermodelling for Java. Different usage scenarios show advancements to related research and practice. In special, we present insights how dependencies and other relations of code structure can be revealed with little effort. Additionally, we report about the reasonable effort to create management dashboards for code structure. Furthermore, we test Hypermodelling to plan revision updates and to compute advanced information out of code. First demonstrations to mine developer knowledge, to compute developer-style issue correlations and to report cohesion indicators underline the potential of the approach. Finally, a web and development environment based code search engine and a code clone recommendation scenario show also advancements to related work in the corresponding areas.
Now, Data Warehouse technology can be used for software engineering research and practice. Among others, our current results can be used to carry out the following business cases: Code search, multi-dimensional and statistic code analysis, code structure audits with automated reporting, live code clone detection and recommendation, data integration and revision update progress control.
Future research and implementations can focus on specialized scenarios, Hypermodelling based source code analysis or to the advancement of Hypermodelling itself. Furthermore, plenty of synergies with related work identify additional research trails and future application scenarios.
Outline:
Introduction and Motivation
In Chapter 2, we introduce the general concepts and research that motivates, supports and represents the foundation of our work. We focus on the paradigm of separation of concerns and give examples where object-oriented programming does not meet the demands of the principle.
In Chapter 3, we get into the advancements of multi-dimensional separation of concerns and a view based concern separation technique. Complementing, we present the area of domain specific languages and show a modeling based approach that offers multi-dimensional navigation of a software system. Lastly, we introduce the area of source code analysis.
Chapter 4 leaves the area of software engineering and introduces Data Warehousing. We explain the different used data structures and give application examples how Data Warehouses handle multidimensional data.
Hypermodelling and Implementation
In Chapter 5, we tie the essence of the prior chapters together and derive Hypermodelling. This is the key contribution of our work. First, we give a brief insight how we utilize Data Warehouse technology to express concern associations of software ...
