# decision_report_odata_vs_deeper_access

## Executive Verdict

`Not yet sufficient for MVP accounting ontology` under hard 3-check gate.

## Hard-Gate Result (2026-03-22)

Source: `logs/deep_accounting_mvp_gate.json`

1. `document -> posting -> debit/credit account`: `pass`
2. `posting -> subconto[1..3] -> counterparty/contract/item`: `fail`
3. Explain one real saldo by movement set: `pass`

Because check #2 is not fully closed (slot `3` not reproducibly proven), the gate verdict is:

`Not yet sufficient for MVP accounting ontology; deeper access is justified for failed checks.`

Additional slot `3` recon artifact:

- `logs/slot3_recon_report.json`
- Summary doc: `docs/slot3_recon_2026-03-22.md`
- UT compatibility status: `docs/ut_compatibility_blocker_2026-03-22.md`

## What Is Already Strong

1. OData endpoint is stable and metadata is readable.
2. Domain coverage is broad (`1189` entity sets).
3. Core document/register joins are reproducible.
4. Subconto linkage is derivable for practical cases (including counterparty/contract/item dimensions), but not yet complete for full `1..3` slot coverage.

## Required Next Move

1. Keep OData as baseline transport and canonical contract source.
2. Add deeper access only for the missing hard-check segment (`subconto slot 3` closure).
3. Use FoxyLink as primary deeper-access PoC candidate.
4. Treat Universal Tools branch as blocked for current `БП 2.0` test contour (no plug-and-play compatibility).
5. Follow execution plan: `docs/deeper_access_execution_plan_foxylink_ut_2026-03-22.md`.
6. Follow test rollout checklist: `docs/test_contour_runbook_foxylink_ut_2026-03-22.md`.
7. Re-run `scripts/run_probe.ps1` until all three checks pass.