Emerging Technology

day one project

Support scientific software infrastructure by requiring SBOMs for federally funded research

02.08.24 | 4 min read | Text by James Howison & Karthik Ram

Federally funded research relies heavily on software. Despite considerable evidence demonstrating software’s crucial role in research, there is no systematic process for researchers to acknowledge its use, and those building software lack recognition for their work. While researchers want to give appropriate acknowledgment for the software they use, many are unsure how to do so effectively. With greater knowledge of what software is used in research underlying publications, federal research funding agencies and researchers themselves will better be able to make efficient funding decisions, enhance the sustainability of software infrastructure, identify vital yet often overlooked digital infrastructure, and inform workforce development.

All agencies that fund research should require that resulting publications include a Software Bill of Materials (SBOM) listing the software used in the research.

Details

Software is a cornerstone in research. Evidence from numerous surveys consistently shows that a majority of researchers rely heavily on software. Without it, their work would likely come to a standstill. However, there is a striking contrast between the crucial role that software plays in modern research and our knowledge of what software is used, as well as the level of recognition it receives. To bridge this gap, we propose policies to properly acknowledge and support the essential software that powers research across disciplines.

Software citation is one way to address these issues, but citation alone is insufficient as a mechanism to generate software infrastructure insights. In recent years, there has been a push for the recognition of software as a crucial component of scholarly publications, leading to the creation of guidelines and specialized journals for software citation. However, software remains under-cited due to several challenges, including friction with journals’ reference list standards, confusion regarding which or when software should be cited, and opacity of the roles and dependencies among cited software. Therefore, we need a new approach to this problem.

A Software Bill of Materials (SBOM) is a list of the software components that were used in an effort, such as building application software. Executive Order 14028 requires that all federal agencies obtain SBOMs when they purchase software. For this reason, many high-quality open-source SBOM tools already exist and can be straightforwardly used to generate descriptions of software used in research.  

SBOM tools can identify and list the stack of software underlying each publication, even when the code itself is not openly shared. If we were able to combine software manifests from many publications together, we would have the insights needed to better advance research. SBOM data can help federal agencies find the right mechanism (funding, in-kind contribution of time) to sustain software critical to their missions. Better knowledge about patterns of software use in research can facilitate better coordination among developers and reduce friction in their development roadmaps. Understanding the software used in research will also promote public trust in government-funded research through improved reproducibility.

Recommendation

We recommend the adoption of Software Bills of Materials (SBOMs) — which are already used by federal agencies for security reasons — to understand the software infrastructure underlying scientific research. Given their mandatory use for software suppliers to the federal government, SBOMs are ideal for highlighting software dependencies and potential security vulnerabilities. The same tools and practices can be used to generate SBOMs for publications. We, therefore, recommend that all agencies that fund research should require resulting publications to include an SBOM listing the software used in the research. Additionally, for research that has already been published with supplementary code materials, SBOMs should be generated retrospectively. This will not only address the issue of software infrastructure sustainability but also enhance the verification of research by clearly documenting the specific software versions used and directing limited funds to software maintenance that most need it.

1. The Office of Science and Technology Policy (OSTP) should coordinate with agencies to undertake feasibility studies of this policy, building confidence that it would work as intended.
   1. Coordination should include funding agencies, federal actors currently applying SBOMs in software procurement, organizations developing SBOM tools and standards,  and scientific stakeholders.
      
2. Based on the results of the study, OSTP should direct funding agencies to design and implement policies requiring that publications resulting from federal funding include an openly accessible, machine-readable SBOM for the software used in the research.

3. OSTP and the Office of Management and Budget should additionally use the Multi-Agency Research and Development Budget Priorities to encourage agencies’ collection, integration, and analysis of SBOM data to inform funding and workforce priorities and to catalyze additional agency resource allocations for software infrastructure assessment in follow-on budget processes.

To learn more about the importance of opening science and to read the rest of the published memos, visit the Open Science Policy sprint landing page.