Decide: shared number space for WO # and OPAL # (ABL SIA pattern)

[CST-100]

Sink-or-swim decision for 1.5.0, from ABL Space Systems practice (SIA numbers): parts there carried just a part number + SIA #, and WO/SIA numbers may never have been split counters — a part inscribed "SIA 5577" was either made by WO 5577 or by WO 5576 with 5577 as the item. Parts also showed prior scratched-out numbers (re-identification across transformations, e.g. printed part → assembled part).

Options analyzed (2026-07-06):

1. Same number (WO N creates item N) — rejected-leaning: batch cut (1 WO → N items), multi-output builds, and zero-output WOs (inspection/rework) break the 1:1.
2. Shared counter, separate identities — the ABL property worth keeping: an inscribed number is globally unambiguous; search resolves it without knowing the kind. Forward-only migration is trivial: one shared sequence seeded past max(existing WO, OPAL counters); issued numbers never change (they are inscribed on hardware — renumbering is forbidden by physics, same spirit as Critical Rule 7).
3. Keep separate — status quo; global search already spans both.

Note: the transformation trail (scratched-out numbers) is already modeled — a new inventory record per transformation with production/consumption genealogy linking them.

Decision owner: Lavie. No schema prep needed now; option 2 stays cheap indefinitely.

🤖 Generated with Claude Code