DDI-L 4.0 BETA - Review Information

Owned by Wendy Thomas
Last updated: Feb 25, 2024
2 min read

Field Level Comparison - spreadsheet showing differences between DDI-L 3.3 and DDI-L 4.0

Background

The vision of DDI has grown over the years from a simple codebook for microdata and aggregate data files, to a model that supports metadata-driven data systems, and then to a suite of products that supports a variety of applications in a broad area of coverage across several related disciplines. This shift in focus reflects changes in the data environment, technology, and user needs over time.

This shift implies a related shift in the technical requirements that will ensure the development and maintenance of a well-integrated suite of products as well as products to support product development and the smooth transfer of metadata between products to support different applications and uses over time. The Technical Committee Roadmap 2023-2027 reflects this shift in focus on technical infrastructure.

The goal is to gradually move all of the DDI products away from hand-crafted production processes and towards increased automation that supports consistency in content, multiple syntaxes, and development processes that allow greater community input and increased iterative testing during development.

Approach for DDI-L 4.0 BETA

  • The content will be the same as version 3.3 but expressed in a technical structure that supports more flexibility and multiple syntaxes

  • Express content in multiple bindings/representations: XML Schema, RDF/OWL, JSON, UML and possibly others

  • Move to an automated production process that will expand our ability to support and test new content and work in a more iterative manner

  • Flatten the model to support full serialization of objects (objects can be published separately) and compiled in sets of objects that describe broader coverage (collections of concepts, study unit information, data collection instruments, etc.)

  • Modify the structure of DDI-L 3.3 to:

    • generalize XML centric content (example - xs:choice)

    • consolidate physical structure descriptions and rename objects as needed to addresses duplication of names within multiple namespaces

    • simplify identification and reference information to allow required content to be expressed in the model and syntax specific content to be added programmatically during the building of the outputs