TU München - Fakultät für
Es sprechen Studenten über ihre abgeschlossenen Diplomarbeiten und Systementwicklungsprojekte.
Am Dienstag, 02.05.17, ab 14:00 Uhr, im Raum Alonzo Church(01.09.014):
Change effort in unit tests as a result of changes in requirements
Until now, there are little to no insights into the effort changed requirements cause in unit test changes. Estimating the effort inflicted by adding or modifying functionality in a software system is of high importance for several reasons like e.g. project planning, which relies heavily on these estimations in regards to deadlines, costs and resources. In our first study, we provide a classification of changes on unit test and linked requirements documents. For the second study, we use these classifications to classify changes made to an industrial system and their respective documents. This results in a distribution of lines of code changed for change classes and linked documents. Combining the former results with an expert interview, we can provide values describing past effort in changes on unit tests. These results lead to upper and lower bounds for modifications linked to functional change requests, which then can be used to estimate the effort in unit test changes caused by future requirements changes.
Development of an Evidence-based Requirements Engineering Risk Assessment Approach
Requirements Engineering is a critical process in software development. Problems in the requirements engineering process can cause serious consequences and even project failure. This thesis aims to support the identification and assessment of risks, based on experiences of closed software projects. Firstly, we investigated literature for related topics in requirements engineering and risk assessment approaches in software engineering. Based on this data, we built an metamodel that outlines the context of the topic. We analysed two datasets regarding information that can be used for risk assessment and tested tools that are suitable for evidence-based risk assessment. In the end, we developed a prototype of an Excel-based tool for Evidence-Based Requirements Engineering Risk Assessment and formulated a risk management approach that assesses risks in an evidence-based way.
Test Accompanying Calculation of Test Gaps for Java Applications
To avoid bugs in software products a lot of effort is put into testing. Especially code, which has been modified, is likely to induce new bugs and hence should be tested carefully. By analyzing coverage reports test gap analysis assists in finding untested changes. For manual tests these reports are often collected from all testers in nightly jobs. The analysis then returns aggregated results for all tests. To improve efficiency for testers we developed a prototypic tool, which allows testers to get the fine-grained results of their own tests immediately after they tested a feature. The implementation is based on JaCoCo as coverage tool. The evaluation showed we can find relevant test gaps with a high precision and recall.