Stand: 2026-03-16

J01-96 Tagebuch: Quellenregister und Roh-Eingaben verankern
Type
Task
Status
Erledigt
Priority
Medium
Assignee
Dani Y.
Sprint
-
Sprint role
Sprint-Admin
Resolution
Fertig
Updated
2026-03-20T12:36:29.377+0100

Contents key

This page

Canonical public work status for J01-96. This issue adds a small source register for search cases to the tagebuch repo and keeps raw chat inputs as a separate provenance layer for round notes.

Goal

Keep tagebuch as a sense-making and source repo without forcing single notes into normative status. The model separates three layers:

  • notes/raw/ for raw inputs
  • notes/inbox/ for smoothed round notes
  • notes/index/ for a minimal source register for search cases

Implemented now

  • created the new Jira issue J01-96 for the journal/register topic
  • added a tagebuch round note for this decision and implementation line
  • added raw inputs for this topic under notes/raw/2026-03-16.md
  • created notes/index/quellenregister.md as the first active register
  • linked the new register from the tagebuch start index
  • updated notes/meta/system.md: registers are derived views for search cases, not the source
  • updated the project skill: for origin and source questions, check the register first, then search tagebuch, and add new finds back into the register
  • moved Jira issue J01-96 to Done

Working model

  • Individual tagebuch notes stay append-only.
  • The register contains only search case, search notes, and sources.
  • A register entry changes when a search request changes the search state.
  • This happens in exactly two cases:
    • an important relevant source was previously missing
    • a search request produces new relevant sources and new search notes
  • Skills point to the register but do not duplicate topic history or create a meta-register.
  • Jira keeps only one canonical remote link for the same public work doc; the mirror and GitHub Pages still stay fully DE/EN.

Source basis

V1 Rules

  • No fields like current, superseded, or valid inside individual journal notes.
  • No daily mandatory indexes.
  • No general development-line or meta-register.
  • Create registers only where real retrieval failures or new source findings occur.
  • No automatic index generation in V1.
  • No duplicate DE/EN remote links in Jira for the same work doc.