You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Decide the model-driven implementation's lifespan and sunset #347
How long does the model-driven implementation under emf/ live, what does the parity contract cost until then, and what is its sunset date?
Options include deleting it right after Task 3, keeping it frozen at a version for the thesis, keeping it in parity until a date, or relaxing parity to a subset of the oracles. #99 holds the deletion work once a date exists. Settle what happens to the specified-ahead pending codes on #324 in each case.
Settled with the maintainer on 2026-10-09, building on Decide what deploy-kit is for after the course: emf/ has no purpose after the grade, so this ticket decides only when and how it ends.
Tag the commit submitted for Task 3, so the graded implementation stays reachable for a repair, a resubmission or an examiner's question.
Delete the coursework implementation at its sunset (Delete the coursework implementation at its sunset #99) lands right after submission. Its trigger changes from "the grade is recorded" to "Task 3 is submitted". The second condition in 0071 stays: the TypeScript suite passes every parity case with the emf job removed from Pipeline Complete.
Rejected: freezing emf/ in the tree until the grade (dead code for weeks, which confuses agents), and 0071 as written (full parity until the grade, the most expensive option).
Before 30 October, a #324 slice that lands changes both implementations. After the deletion, every #324 slice lands in TypeScript only. Its pending codes stay pending on #324 in scripts/lint-codes.ts and leave that list as each TypeScript test lands. That mechanism does not change.
docs/mde/
The task-0 to task-3 report directories stay, frozen, as design history: how the model was argued. This is a stated exception to Decide what deploy-kit is for after the course #344's scale test. The reports catch no mistakes, but they record reasoning that the ADRs only gist.
assignment.md, practicals.md and the course-material submodule hold the course's own text, so they go once the grade is recorded. AGENTS.md's MDE section then shrinks to a pointer to the reports. That work is a separate issue, not part of Delete the coursework implementation at its sunset #99.
Record
No new ADR. 0071 is deleted with emf/, and #99's body now carries the new trigger. The 14 retired emf ADR numbers become part of the register reshape that is still in the map's fog.
Coursework
No effect on Task 2 (16 October) or the 22 October presentation. Task 3 (30 October) gains #338. Deletion waits for the submission.
Activity
ExtraToast commentedon Oct 9, 2026
Resolution
Settled with the maintainer on 2026-10-09, building on Decide what deploy-kit is for after the course:
emf/has no purpose after the grade, so this ticket decides only when and how it ends.Until Task 3 is submitted (30 October)
emf/.The sunset: delete at submission
emfjob removed fromPipeline Complete.emf/in the tree until the grade (dead code for weeks, which confuses agents), and 0071 as written (full parity until the grade, the most expensive option).Pending codes on #324
Before 30 October, a #324 slice that lands changes both implementations. After the deletion, every #324 slice lands in TypeScript only. Its pending codes stay pending on #324 in
scripts/lint-codes.tsand leave that list as each TypeScript test lands. That mechanism does not change.docs/mde/
assignment.md,practicals.mdand thecourse-materialsubmodule hold the course's own text, so they go once the grade is recorded. AGENTS.md's MDE section then shrinks to a pointer to the reports. That work is a separate issue, not part of Delete the coursework implementation at its sunset #99.Record
No new ADR. 0071 is deleted with
emf/, and #99's body now carries the new trigger. The 14 retiredemfADR numbers become part of the register reshape that is still in the map's fog.Coursework
No effect on Task 2 (16 October) or the 22 October presentation. Task 3 (30 October) gains #338. Deletion waits for the submission.