Background
In trying to address T335317, we don't want to paint ourselves into a corner and commit to a long-term solution without first validating our approach. We can do this by prototyping against known use cases.
Example: Message boxes (T326587). Can we enable the Web team to convert these to Codex without making the Codex build and development process a nightmare?
Acceptance Criteria
- Identify a use case where Codex adoption is currently not possible or challenging, but where we believe a code-splitting solution would enable adoption.
- Prototype a solution to this use case and validate it with the owners of that product area (and other relevant stakeholders).
- Continue iterating [within a specified time box] until we either solve the adoption hurtle or learn that another approach is needed.
- By EOQ: make a decision on whether it's worth continuing to try to solve this problem or whether to focus on adoption hurtles elsewhere.
Questions to ask:
- Does this create the best experience for Wikimedia end-users?
- Will this solution be easy to use for Codex users?
- How does this affect the scalability and maintainability of Codex?
- Does this make it easier / harder to contribute to Codex?
- Does this require more work in the future? Or would this solution be enough to satisfy the current use case?
- How does the solution map to larger picture of T335785?