Deep Software Variability for Replicability in Computational Science (talk in Japan)
I talked about deep software variability and reproducibility/replicability in computational science at SES 2023, IPSJ/SIGSE Software Engineering Symposium in Tokyo, Japan. Thanks to Paolo Arcaini and Fuyuki Ishikawa for the invitation! Slides are here: https://docs.google.com/presentation/d/1S2YDDMHw9FJ-ogpiGvUvmeHkYhFOQo4Xbccmjg4FL_Q/edit?usp=drivesdk A few photos: https://twitter.com/acherm/status/1695305466455236925
The exact title of the talk was “Deep Software Variability for Replicability in Computational Science” Abstract:
In the software world, many layers - encompassing the operating system, third-party libraries, software versions, workloads, compile-time options, and flags - are subject to variability. This deep software variability can significantly alter computation results and impact performance. In the talk “Deep Software Variability for Replicability in Computational Science” we report on the complex interactions across these variability layers and their subsequent effects on both the reproduction and replication of scientific studies. While deep software variability often poses challenges to achieving reproducible results, we propose embracing this variability as a pathway to replicate existing studies with more extensive and systematic exploration of hypotheses, methodologies, and analyses. By harnessing deep variability, we can generate more robust, generalizable scientific findings, presenting a fresh perspective for computational science’s future.
It was great to meet the Japanese software engineering community!