Open-source models for development of data and metadata standards

Ariel Rokem, Vani Mandava, Nicoleta Cristea, Anshul Tambay, Kristofer E. Bouchard, Carolina Berys-Gonzalez, Andy ConnollyPatterns 2025 · DOI

TL;DR

This paper explores how the principles and practices of open-source software (OSS) development can be leveraged to create more robust, adaptable, and widely adopted data and metadata standards across scientific disciplines. By examining case studies in astronomy, high-energy physics, earth science, and neuroscience, the authors propose recommendations for fostering effective open-source standards development.

What Problem Does This Paper Address?

In today's data-intensive research landscape, the ability to access, integrate, and reuse large, complex datasets is paramount. This is especially true with the rise of machine learning and AI, which thrive on diverse data. However, achieving this interoperability and reusability hinges on the development and adoption of effective data and metadata standards. The challenge lies in creating standards that are not only technically sound but also adaptable to evolving needs, sustainable in the long term, and accessible to a broad community of users and developers. Traditional top-down approaches to standards development can be rigid and slow to adapt, while purely bottom-up approaches can lack structure and sustainability. This paper addresses the need for a more effective model by drawing inspiration from the successful development paradigms of open-source software.

Methods

The authors' approach is primarily a synthesis and analysis of existing knowledge and experiences. They convened a workshop in April 2024 with experts from various scientific fields and sectors to discuss the intersection of OSS development and data/metadata standards. The paper then synthesizes insights from this workshop, along with the authors' own expertise, to analyze case studies from four scientific domains: astronomy, high-energy physics (HEP), earth science, and neuroscience. They examine how OSS tools and social frameworks have influenced the development of standards in these fields, identifying both opportunities and challenges. The paper concludes with a set of recommendations for stakeholders.

Datasets

This paper does not present new datasets or benchmarks. Instead, it analyzes the development and adoption of existing data and metadata standards within various scientific fields. Examples of standards discussed include:

  • FITS (Flexible Image Transport System): For astronomy data.
  • NetCDF (Network Common Data Form): For earth science data.
  • HDF (Hierarchical Data Format): For earth science data.
  • GeoTIFF: For earth science image data.
  • Cloud Optimized GeoTIFF (COG): For cloud-based earth science data.
  • Zarr: For array-based datasets in cloud storage.
  • Neurodata Without Borders (NWB): For neurophysiology data.
  • Brain Imaging Data Structure (BIDS): For neuroimaging data.
  • DICOM (Digital Imaging and Communications in Medicine): For medical imaging.
  • HL7 FHIR (Fast Healthcare Interoperability Resources): For healthcare data.

Key Insights and Ideas

  • OSS as a Model for Standards: The core idea is that the technical tools (e.g., version control systems like Git) and social frameworks (e.g., community governance, collaborative development) that make OSS successful can be adapted to create more effective data and metadata standards.
  • Flexibility vs. Stability Trade-off: Open-source standards offer flexibility to adapt to new needs and technologies, but this can conflict with the stability required for existing datasets and analysis tools. Balancing these is crucial.
  • Community-Driven Development: Bottom-up contributions from users and developers can lead to standards that are more adaptable and address a wider range of needs, including unexpected use cases.
  • Mismatches Between Developers and Users: A significant challenge is the gap in technical expertise and differing motivations between those who develop standards and those who use them, potentially leading to standards that overlook practical user needs.
  • The Importance of Meta-Standards: Developing "standards of standards" or meta-standards can help formalize best practices for standards development, fostering consistency and interoperability across domains.
  • Tandem Development of Standards and Software: Standards should be developed in parallel with associated software tools to ensure practical usability, automated validation, and a virtuous cycle of improvement.
  • Governmental Policy and Funding: Policy directives and funding mechanisms play a critical role in incentivizing and supporting the development and adoption of open-source data standards.

Key Results

This paper does not present quantitative results in the traditional sense. Instead, its key findings are qualitative insights and recommendations derived from the analysis of existing practices:

  • Successful Adoption: Fields like astronomy and earth sciences have a history of successful standards development, often influenced by OSS principles, demonstrating the viability of this approach.
  • Emerging Fields Benefit Most: Neuroscience, being a more recent field in large-scale data collection, has been quicker to adopt OSS-inspired development models for its standards.
  • Identified Challenges: The paper clearly outlines challenges such as balancing flexibility with stability, managing community consensus, overcoming technical barriers to entry (e.g., learning Git), and addressing proprietary data formats.
  • Actionable Recommendations: The paper provides a structured set of recommendations for science and technology communities (e.g., establish OSS-based governance, foster meta-standards) and for policymaking and funding entities (e.g., fund development, invest in data stewards).

Limitations and Future Work

Limitations:

  • The authors acknowledge that their survey may have "gaps and blind spots," particularly in fields where they have less expertise, such as electronic healthcare records.
  • The paper focuses on the adoption of OSS models rather than introducing entirely new OSS methodologies for standards development.
  • The challenges of cultural inertia and the inherent difficulty of changing established practices are implicitly acknowledged but not deeply explored.

Future Work:

  • The authors suggest the development of a "maturity model" for assessing the challenges and opportunities of specific standards, which could lead to the eventual development of a meta-standard.
  • Further exploration of formalizing guidelines for standards creation and versioning, potentially using tools like LinkML, is implied.
  • The paper calls for continued investment in data stewards and the development of training programs to foster expertise in open-source standards development.
  • There is an ongoing need to bridge the gap between technical development and practical user needs, and to ensure that funding and policy align with the goals of open-source standards.

Summary updated

BibTeX
@article{rokem2025open,
  author = {Ariel Rokem and Vani Mandava and Nicoleta Cristea and Anshul Tambay and Kristofer E. Bouchard and Carolina Berys-Gonzalez and Andy Connolly},
  title = {{Open-source models for development of data and metadata standards}},
  journal = {Patterns},
  year = {2025},
  doi = {10.1016/j.patter.2025.101316},
}

Loading related content...

Figure