Skip to content
sdlc

B-PKWH-d-0015-five-open-questions-on-experience-scope-reach-need

Every implementable section of D-0015 is now built (the arity axis and its derivation, URL narrowing, the verdict returning to the experience, the stats tiled reference implementation, the write axis, and the dead-field removals). What remains are judgement calls the decision deliberately deferred, and none of them can be settled from the code:

  1. Layout control for the tiled case. Does an experience influence how its instances are laid out, or is that entirely the frame’s? The frame currently owns the grid outright. A view wanting full width at two vaults and a strip at eight is the case that decides it.
  2. What one-tiled does at eight vaults — tile, paginate, or decline and ask the user to narrow. No cap exists today, deliberately.
  3. Where a connector’s vault binding is declared. scheduler and connector are bound to one vault by a daemon rather than by a signature, so widening a read does not move them. This is the one place the connector kind might earn a manifest field.
  4. Whether ?scopeItem= should persist per view across navigations. A user who narrows Stats to vault B and leaves probably expects to return to B; a URL alone does not remember that.
  5. Write-target resolution above one vault. active is honest and surprising — the save lands somewhere the user may not be looking. The shell now STATES where it lands; what to do beyond stating it likely differs per experience, and writeTarget: 'selection' exists in the contract with no declarer yet.

Also outstanding and adjacent: D-0015 sits at status: open/proposed with need_human_review: true, and section 8’s per-experience lens/connector/ platform-surface/manager assignment is still unwritten — the decision calls it a product judgement rather than a code fact.

← Back to Backlog