Karthik Ram | James Howison
In this captivating episode of Sustain, host Richard welcomes returning guest, Karthik Ram, Senior Research Scientist at the Berkeley Institute for Data Science, and his colleague James Howison, an Associate Professor from the School of Information at the University of Texas, Austin. Today, they delve into their recent research report, “‘Research Software Visibility Infrastructure Priorities,” commissioned for the Australian Research Data Commons. They discuss their eight key recommendations about sustaining open source for the long haul, including ways to recognize software contribution, implement web analytics, and offer low friction ways for researchers to link software. Karthik and James also touch on the future of software citations in academic recognition systems, and the importance of universities valuing diverse academic outputs. Don’t miss this fascinating conversation! Press download now!
[00:01:36] Richard brings up a paper written by Karthik and James. Karthik explains the report titled, “Research Software Visibility Infrastructure Priorities,” produced for the Australian Research Data Commons (ARDC). He describes the process of creating the report and the report’s relevance beyond Australia.
[00:06:24] Richard asks how this is related to open source, and James relates the recommendations, focusing on citing software in publications and creating software bill of materials for research papers.
[00:08:02] James and Karthik discuss recommendations, focusing on citing software in publications and creating software bill of materials for research papers.
[00:12:02] Richard endorses the use of SBOMs for citing all software used in research, aiming to counter the issue of only popular projects getting noticed, but he questions how SBOMs account for the varying importance of different software dependencies. Karthik clarifies the SBOMs are not meant to create equal value citations for all software but to understand the scientific infrastructure that supports research.
[00:15:28] Richard suggests that SBOMs could be useful in industrial contexts for security purposes and infrastructure visibility. Karthik agrees, stating that SBOMs have a broader application and were originally created for security reasons to track vulnerabilities.
[00:17:41] James introduces the third recommendation to create software use infrascope as an observatory based on software mentions publications. He discusses the challenge of identifying software mentions in publications and the work done towards building comprehensive view of software in academia.
[00:22:34] Karthik introduces the fourth recommendation to create detailed use cases for research tools aimed at different skill levels, addressing the challenge researchers face when selecting software tools.
[00:24:29] Richard highlights the necessity of allocating time in research planning for writing documentation and tutorials, which James agrees is crucial for making software tools more accessible to researchers.
[00:26:30] James discusses the fifth recommendation, which is to support existing technology for software archiving, such as Zenodo or Software Heritage, rather than creating new repositories at the institutional level.
[00:28:28] Karthik talks about sixth recommendation and supporting communities of practice like hackathons and other collaborative spaces, which have shown to have a positive impact on research productivity. James describes the need for third spaces that are neither too local nor too public. Where researchers can comfortably ask questions and share insights within a focused community.
[00:31:08] James introduces the seventh recommendation which is about implementing web analytics to gain insights into software usage, as citations alone do not reflect the full impact of research software.
[00:33:48] James acknowledges the need for infrastructure to enhance insight from SBOMs and mentions the necessity of funding to maintain services that provide such data.
[00:35:53] Richard highlights the eighth recommendation, which suggests providing an easy way for researchers to link to software alongside data submissions. James directs listeners to the Softcite GitHub organization and mentions the upcoming blog post about their report on the URSSI Blog.
[00:36:16] Karthik and James tell us where you can find out more about their work and find them on the web.
[00:08:40] “Absolutely every piece of software that you use in your whole stack should be cited, but I’ve had some issues with that in publications.”
[00:09:04] “What we identified was that different fields have different norms for what rises to the level of contribution for actually being mentioned formally in the publication.”
[00:14:40] “The researchers are experts in scientific explanation and they’re going to pick packages to mention that pertain to understanding the research that’s done in the paper, whereas the SBOM is going to give us insight into the software infrastructure that made the research possible.”
- [00:38:26] Richard’s spotlight is iNaturalist.
- [00:39:00] Karthik’s spotlight is Kyle Niemeyer at Oregon State.
- [00:39:34] James’s spotlight is Eva Brown and her Council Data Project.
- SustainOSS Twitter
- SustainOSS Discourse
- SustainOSS Mastodon
- Open Collective-SustainOSS (Contribute)
- Richard Littauer Mastodon
- Karthik Ram Website
- Karthik Ram LinkedIn
- Karthik Ram X/Twitter
- James Howison X/Twitter
- James Howison-University of Texas, Austin
- Research Software Visibility Infrastructure Priorities Report by Dr. Karthik Ram and Dr. James Howison (Zenodo)
- Sustain Podcast-2 Episodes featuring Daniel Stenburg
- Sustain Podcast-Episode 187: Karthik Ram on Research Software Sustainability
- Softcite dataset
- Softcite software mention recognition service
- SoMeSci-Software Mentions in Science
- Mapping the Impact of Research Software in Science
- The Scientific Community Image Forum (Frequently Asked Questions)
- Sustain Podcast-Episode 95: Marko Saric of Plausible Analytics, the most popular Open Source analytics platform
- URSSI Blog
- Incentives and integration in scientific software production by James Howison and James D. Herbsleb
- Python Package Citation Generator-GitHub
- Kyle Niemeyer
- Eva Brown GitHub
- Council Data Project
- Produced by Richard Littauer
- Edited by Paul M. Bahr at Peachtree Sound
- Show notes by DeAnn Bahr Peachtree Sound