Add AI meeting summaries (#1948)

Author: benjie

notes/2026/summary-2026-03-05.md

@@ -0,0 +1,56 @@
+# Meeting Summary for Working Group Meeting
+
+**NOTICE**: This summary was auto-generated by Zoom's "AI". AI-generated
+content may be inaccurate or misleading. Always check for accuracy. If in
+doubt, please consult the meeting recording at
+https://youtube.com/@GraphQLFoundation/playlists
+
+- Meeting start: 2026-03-05T18:28:55Z
+- Meeting end: 2026-03-05T20:02:08Z
+- Summary start: 2026-03-05T18:28:55Z
+- Summary end: 2026-03-05T19:57:49Z
+
+The GraphQL specification working group meeting focused on several key topics including the upcoming Working Group Day on May 8th, the Issue Club initiative for reviewing and grooming GraphQL repository issues, and the GAPS (GraphQL Extension Proposal System) project for community specifications. The group discussed the implementation of one-of inhabitability validation rules and the ongoing work on the defer and stream specification changes. A significant portion of the meeting was dedicated to debating the numbering system for GAPS, with participants weighing the benefits of an incremental numbering approach versus a more structured categorization system, ultimately deciding to use pull request numbers for simplicity while maintaining clear delineation between RFCs and other issues.
+
+## Next Steps
+
+- Martin: Flag issues in the GraphQL spec report with "stale question mark" label and encourage owners to comment or join the next Issue Club meeting for review (March 19th).
+- James: Land the "oneof inhabitability" PR into GraphQL.js v17 and then create a PR for v16.
+- Rob: Add links to adapter libraries (e.g., Yaacov's adapter prototype and Rob's Relay demo) in the Defer & Stream PR for reference.
+- Rob: Update PR 1110 to match recent terminology/wording changes from the integration branch and keep it up to date.
+- Mark: Synthesize and list out the proposed numbering systems/options for GAPS in the GAPS repository for further discussion and decision.
+- Benjie: Meet with Linux Foundation representative to clarify legal requirements for merging/spec proposals in the GAPS repository and update the group after the meeting.
+- Mark and Benjie: Reach out to Jordan (Meta) to discuss submitting the cursor connection and global object identification specs as initial GAPS.
+- All interested parties: Continue discussion on GAPS numbering/naming conventions in the GAPS repository issues.
+- All interested parties: Register for Working Group Day (May 8th) by filling out the Google form if planning to attend.
+- Martin: Share the link to the stale issue list for the spec report for community review and comment.
+
+## Summary
+
+### GraphQL Working Group Meeting Updates
+
+The GraphQL specification working group meeting began with introductions from Benjie, James, Martin, Mark, Pascal, and Benoit. Benjie mentioned that Lee would be late and that the agenda had some merge conflicts which he was resolving. The group discussed the use of AI for note-taking and agreed to contribute to live notes for more accurate documentation of technical discussions. They reviewed the agenda, which included topics such as a review of the previous secondary meeting, ready-to-review agenda items, grants, upcoming WG Day, Issue Club, GAP project, and specification changes.
+
+### GraphQL Foundation Grants and Events
+
+The TSC announced a grants program for projects under the GraphQL Foundation repositories, managed by the TSC. Benjie introduced the inaugural Working Group Day scheduled for May 8th, following GraphQLConf, with limited capacity and an open agenda for socializing and technical discussions. Martin discussed the next Issue Club meeting on March 19th, focusing on the GraphQL spec report, and encouraged participants to review and comment on flagged "stale" issues. Martin also inquired about the GAAP process, but the link to the GAAP repo was not provided in the transcript.
+
+### GraphQL Community Specification Project
+
+The meeting discussed the GraphQL community specification project, known as GAPS, which aims to provide a central repository for community specs. Mark explained the purpose of GAPS and mentioned ongoing efforts to upstream specifications from Yelp. Benjie emphasized the importance of having a vendor-neutral home for specifications to ensure longevity and interoperability. He also mentioned that there are legal considerations involved with adding new specifications, and they may need to hold off on merging major changes until these are sorted out.
+
+### GAPS Specification Merge Process
+
+The team discussed the process of merging specifications and opening pull requests, with Benjie noting that pull requests from forks are generally acceptable but may require signatures if needed. Mark encouraged the team to identify potential gaps in their respective companies and proposed a version range system inspired by PEPs for categorizing GAPS directives. Pascal raised questions about collaboration between contributors and suggested the possibility of a GAPS working group to facilitate communication. The team also discussed the potential for deprecating older specifications in favor of newer, more efficient ones, and the possibility of standardizing scalar definitions within the GAPS framework.
+
+### Specification Naming Convention Debate
+
+The team discussed the naming convention for specifications, with Benjie explaining that the numbered system was inspired by Python Enhancement Proposals to avoid issues with name changes and to reduce bias towards named solutions. Michael and Pascal debated the merits of naming versus numbering, with Pascal suggesting that scalars should be prefixed with company names while Michael preferred keeping them in the same namespace as other specifications for better integration and advertising. The group agreed that specifications should be versioned and detectable through structured methods, with Benjie emphasizing the importance of the cursor connection specification as an example of successful renaming without losing its ID. Mark proposed adding front matter to gap specifications with metadata like gap ID, gap name, and version, while Benjie concluded that further exploration of keeping scalars separate from GAPS was worthwhile.
+
+### GraphQL One-of Input Fix
+
+James presented a pull request addressing an issue with one-of input objects in GraphQL, explaining the problem and demonstrating a solution through a small demo. The team agreed that the proposed fix was worth implementing, moving it to RFC1 status. James expressed satisfaction with the current approach, which maintains the one-of type's non-null requirement for fields while allowing for a more flexible implementation. The group decided to implement the fix in both GraphQLJS and Apollo Kotlin, with Lee suggesting that once multiple implementations are complete, the solution would likely be promoted to RFC2.
+
+### GraphQL Specification Progress Review
+
+The meeting focused on the progress and next steps for the Defer and Stream specifications in GraphQL. Rob provided a recap of the current status, highlighting the implementation of the response format in various tools and the ongoing efforts to update the specification. The team discussed the need for further reviews and feedback on the proposed changes, with Benjie and others offering to assist. Michael suggested linking relevant resources in the pull request to aid implementation. The group also debated the numbering system for future specifications, with a consensus forming around using a simple, incremental approach to avoid overcomplicating the process. Benjie emphasized the importance of clear communication and editing processes to streamline the submission and assignment of specification numbers.