Linking Mathematical Software in Web Archives (1702.01165v1)

Abstract: The Web is our primary source of all kinds of information today. This includes information about software as well as associated materials, like source code, documentation, related publications and change logs. Such data is of particular importance in research in order to conduct, comprehend and reconstruct scientific experiments that involve software. swMATH, a mathematical software directory, attempts to identify software mentions in scientific articles and provides additional information as well as links to the Web. However, just like software itself, the Web is dynamic and most likely the information on the Web has changed since it was referenced in a scientific publication. Therefore, it is crucial to preserve the resources of a software on the Web to capture its states over time. We found that around 40% of the websites in swMATH are already included in an existing Web archive. Out of these, 60% of contain some kind of documentation and around 45% even provide downloads of software artifacts. Hence, already today links can be established based on the publication dates of corresponding articles. The contained data enable enriching existing information with a temporal dimension. In the future, specialized infrastructure will improve the coverage of software resources and allow explicit references in scientific publications.