Bỏ qua để đến nội dung

EVT 2026-02-21 - Stage D2 Attempt Flow Deep Policy Autocomplete

DomainsDOL EnglishProduct151 words1 min read
active

User approved the recommended deepening package for attempt execution so the core learning flow can be both robust and simple.

  1. Timed exit/background behavior:
  • timed countdown remains authoritative,
  • if time expires while user is away/in background, attempt goes to timeout-submit path.
  1. Draft retention TTL:
  • unfinished untimed draft retention: 7 days since last activity,
  • timed drafts are valid only within active timer window.
  1. Submit review format:
  • show a lightweight total summary (answered/unanswered) before submit,
  • if attempt has multiple sections, allow expandable section-level breakdown.
  1. Submit retry and support threshold:
  • auto-retry transient submit failure up to 3 times,
  • after auto-retry exhausted, keep manual retry available and show support CTA.
  1. Timed pause policy:
  • no manual pause in timed mode,
  • timer continues while tab/app is backgrounded.
  • None for this deep-policy batch.