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
Design note carried over from #33 (1/4): a chart edit reaches the top tier only.#47 persists authored per-tier content, but the top tier is still rebuilt from the editor's flat chart, so deleting or editing a note updates the top tier and leaves the same note in every authored lower tier until that tier is itself edited. With no tier-authoring UI in place yet, a lower tier is effectively frozen content.
That is the right model for authored simplifications — a lower tier is supposed to be its own thing — but it has to be designed around rather than discovered later:
Per-phrase tiers 2/4: "simplify this phrase" command to derive a lower tier #34 ("simplify this phrase") — a derived tier is the most likely to go stale against the chart it was derived from. Whatever this sub-issue decides about re-deriving or refreshing a tier is the natural place to make the staleness visible or resolvable.
Per-phrase tiers 3/4: difficulty-tier authoring mode (N variants per phrase) #35 (this issue) — "edits apply to the selected tier only" needs the flip side stated: editing the flat chart is an edit to the top tier, and the UI should say which tier a given edit is landing in, so an author who deletes a note isn't surprised to still see it at lower mastery.
Source: _authored_phrase_levels in routes.py — the docstring, the [Unreleased] CHANGELOG entry, and the architecture note in README.md now all state the boundary explicitly.
Activity
pullfrog commentedon Oct 2, 2026
pullfrog commentedon Oct 2, 2026
Design note carried over from #33 (1/4): a chart edit reaches the top tier only. #47 persists authored per-tier content, but the top tier is still rebuilt from the editor's flat chart, so deleting or editing a note updates the top tier and leaves the same note in every authored lower tier until that tier is itself edited. With no tier-authoring UI in place yet, a lower tier is effectively frozen content.
That is the right model for authored simplifications — a lower tier is supposed to be its own thing — but it has to be designed around rather than discovered later:
Source:
_authored_phrase_levelsinroutes.py— the docstring, the[Unreleased]CHANGELOG entry, and the architecture note inREADME.mdnow all state the boundary explicitly.big-pickle(free) | 𝕏