Science-Induced Retrospective Code Analysis

点击翻译按钮获取中文摘要

Most of temporary scientific research depends essentially on the employment of complex software, although scientists are usually not trained in programming and software engineering. From the perspective of Philosophy of Science (PoS) the increase in scientific software as research tools is of epistemic and analytic interest; and for Software Engineering (SE) the analysis and improvement of the architecture and algorithmic structure of software projects is a key research topic. Thus, both disciplines focus on the analysis of code and share the conjoint interest in the development of newdomain-specific methods for the analysis of scientific (open) software projects.For Philosophy of Science analysing scientific code opens up a new research domain.For Software Engineering, code-analysis is an integral part for the evaluation of the software-designs' feasibility as well as for thetest of the code-quality. However, conceptions and tools for the analysis of code-bases with regard to the scientific content ofprograms are missing. The scientific content of programs and their structure is usually oriented on the scientific models and laws that are embedded with code. Plus, scientists are usually not trained in programming and software engineering. Therefore, with our approach want to enable scholars, foreigners and the scientific developers themselves, who are not trained in programming to understand software, software-structures and their impacts better. Via abstraction from technical details by code separation, we aim to enable a scientifically-driven retrospective code-analyses for a more appropriate way to access big and long-lasting software in science; in particular for collaborative open science software projects.To sum up, the goal of our project is to articulate methods for code separation and to develop algorithms that permit an easy access to code that is of relevance for Science itself, but also for Philosophy of Science used for studying software in science. And it is ofinterest for Software Engineering to enhance the idea of technical / application code separation for the long lasting, long evolvingscientific software projects.