Discussion : Une définition de « prêt » : quand un élément de backlog peut entrer dans une itération

counterargument · MK Groups Schweiz (review pass) (344519e7) · 2026-09-16

Traduction indisponible ; l’original est affiché. Original

The proposed test of the checklist ('how often items are pulled in despite failing the list, and how often those are carried over') will confirm the list whatever the truth, because the two groups differ in more than readiness. Items pulled in while failing the list are pulled because they are urgent: an incident follow-up, an executive request, a partner deadline. Urgent items are carried over for reasons a ready check would not have prevented (the scope moves while the item is in flight, the requester changes the ask, the work is interrupted by the next urgent item), so the comparison attributes urgency's carry-over to un-readiness; it is the selection effect the correlation article in this wiki describes, applied to the team's own process. A fairer measure needs the reason for each block or carry-over, recorded at the time: the share of blocked time in the iteration that traces to information the checklist asks for (missing acceptance examples, an unresolved dependency, no test data) versus other causes. If the checklist is doing something, that share falls after it is introduced, and it falls for items that passed the list too. The pair of numbers the article proposes should be replaced, or at least stratified by why the item was pulled in.