Repository navigation
DDI Lifecycle Working Group Minutes 2026-09-01 #110
DanSmith
announced in
Announcements
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Minutes 2026-09-01
Attendees
Apologies:
Quick recap
The meeting focused on planning an in-person meeting in Cologne and on discussing recent GitHub issues for DDI Lifecycle. Dan explained the Technical Committee’s plan for a week of face-to-face working group meetings in spring and the group expressed interest, with some concerns about travel funding and duration. The group then discussed adding Population as a new item type in Lifecycle 4, how to handle temporal and geographic coverage, and whether to reuse existing coverage types or move properties, with Guillaume favoring reuse and Romain suggesting deprecation tagging for universe properties. They reviewed several issues raised by Nicolas, including how to represent reusable code and transformations at the variable and dataset level, possible changes to generation instruction, and the need for better support for logical datasets. Romain and Hayley also raised questions about how to describe data capture from non-traditional sources like AI and social media within Lifecycle. Dan explained the process for handling GitHub issues and the use of discussions, and noted that he would continue as interim chair since no one volunteered for the role.
Next steps
Dan
Collaboration
Summary
Cologne In-Person Meeting Discussion
The group discussed a potential in-person meeting in Cologne during spring, organized by the Technical Committee for multiple working groups to participate in face-to-face meetings. The group discussed plans for holding working group meetings at GESIS in Cologne, with potential attendance from the technical committee. Romain, Hayley, and Guillaume expressed interest in attending for 2-3 days, possibly in March or April, with the possibility of travel funding assistance.
Population Type in DDI Lifecycle
The team discussed adding population as a separate item type in DDI Lifecycle 4, as requested by Statistics Canada. They debated whether to reuse existing temporal and geographic coverage wrappers or add these properties directly to the population type, considering the goal to flatten nesting structures in future versions. The group clarified that population would select a subset from a universe, with temporal and geographic coverage being optional properties that could help distinguish between universe and population levels. Romain suggested using deprecation tags to gradually move temporal and geographic properties from universes to populations.
Universe-Population Concept Implementation Discussion
The team discussed implementing a distinction between universe and population concepts in their system. Romain proposed deprecating temporal and geographical attributes from the universe level and moving them to the population level, with proper documentation to guide users on the transition. Guillaume suggested reusing existing types to avoid redundancy, while Romain recommended using a deprecation approach similar to programming languages to gradually phase out the old attributes. The team agreed that implementing a deprecation mechanism would be a good feature for future tooling, Colectica/cogs#365
VTL Code Reuse Solutions
The team discussed issues related to variable transformation documentation and code reuse in VTL. Romain explained that Nicolas raised concerns about not being able to avoid repeating shared code for different variables, particularly when using DTL or UDOs. Dan suggested using URI references to external command files as a potential solution, though Romain expressed concerns about breaking the symmetry of having everything inline in the metadata. The discussion concluded with Dan mentioning that relaxing cardinalities to allow multiple references could be a simpler backward-compatible solution. The team discussed several technical issues related to the Lifecycle model, particularly focusing on generation instructions and how to handle data capture beyond traditional questionnaires and measurements. Romain proposed allowing generation instructions to reference instruments and datasets, while Dan suggested adding a reusable command type to handle shared code.
Lifecycle Model Technical Discussion
The group also discussed a ticket about data capture using AI, with Romain and Hayley clarifying that this should cover data capture processes from various sources including social media and administrative data.
Since no volunteers came forward to serve as chair, Dan will continue as interim chair while noting he is very busy and may have delays in meeting notes and agendas.
All reactions