The long-term preservation of research outputs increasingly depends on software preservation, as data in many disciplines are inseparable from the computational tools used to generate and interpret them. While repositories have established robust practices for preserving data and publications, software remains fragile due to short lifecycles, complex dependencies, and rapidly evolving technological environments. Archiving source code alone is often not enough. This presentation reports on the progress of a repository-based approach to preserving running applications. Building on an earlier proof of concept, the project has evolved into an operational repository service. Recent developments include improved container ingestion workflows, clearer separation between preservation and execution layers, standardized packaging practices, and controlled, sandboxed execution environments that lower technical barriers while addressing security concerns. In parallel, the project has focused on collaboration with researchers, particularly in the digital humanities, to better understand software development practices and to develop guidance, documentation, and support structures for long-term preservation. Rather than proposing a single solution, this work outlines a practical path forward and highlights open questions regarding standards, interoperability, and FAIR principles for research software, contributing to ongoing community discussions on the preservation of executable software.
Raman Ganguly (Tue,) studied this question.