# Synthetic engineering examples

Prepared 13 November 2026 by TMH18 Validator / DATSEMBRA (Pty) Ltd. These invented training files contain no client survey. They do not imply a receiving authority's approval, a completed visual assessment or a real traffic station.

## TMH-14 header count

Source: COTO TMH-14, Committee Draft, Format Version 3.20, issue 4 April 2018, section 8.10, printed pages 50–51.

`tmh14-header20-invalid.rsv` contains `20,60,01,1,3,80`: three declared speed bins but only one boundary (six fields rather than the required seven).

`tmh14-header20-corrected.rsv` contains `20,60,01,1,2,80`: two bins and one boundary. The invented source configuration explicitly specifies two bins separated at 80 km/h. The correction is justified by that source assumption, not simply by a wish to pass.

These are single header-record fragments, not complete mixed RSV files. Select the record-level `TMH-14.HDR-20` template for the exercise. Type 20 after H9 is a different traffic-summary layout.

Reproduced outcome with the application structural validation: the invalid fragment produces `TMH14§9-11/DQ-08`; the corrected fragment has one record and zero structural errors. This does not verify interval completeness, measured traffic, calibration or complete TMH-14 conformity.

## TMH-9 context / TMH-18 flexible exchange

Source context: TMH-9 Part A, Committee Draft Final, May 2016, sections A.1.4/A.2. Exact exchange fields: TMH-18 Committee Draft 5.0, February 2016, section 5.2/Table 13, including SURFACE_FAILURE_DEG and SURFACE_FAILURE_EXT on printed page 36.

`tmh9-flx-invalid.flx` and `tmh9-flx-corrected.flx` preserve all 89 field positions of the selected `TMH-18.flx v5-2016` definition. The only changed field is SURFACE_FAILURE_DEG, from 6 to 2. The invented source note gives degree 2 and extent 2. The invalid 6 exceeds the exchange field’s 0–5 range. The full position list is in `tmh9-flx-field-map.json`.

Reproduced outcome: `RANGE:SURFACE_FAILURE_DEG` for the invalid fixture; zero structural errors for the corrected fixture. Headerless input is interpreted using the selected definition's ordered fields. Optional blank fields are deliberately retained and do not claim a complete assessment. No assessor, network, instrument or authority acceptance is verified by this exercise. WCP is an example permitted authority code, not evidence of an official Western Cape dataset or endorsement.

The automated fixtures are tested in `tests/articles-worked-examples.test.ts`. The public prose describes the narrow result. A zero-error format test does not demonstrate full manual compliance or field accuracy.

## Completed working records

`completed-working-records.json` demonstrates a manifest, reconciliation, edition/profile record and correction note. Its byte count and SHA-256 are calculated from the actual corrected FLX fixture. Its reconciliation is synthetic: 1 source record = 1 loaded + 0 rejected + 0 excluded + 0 unaccounted. No live import was performed. Unknown reviewer, approval and certificate fields remain null.

You may copy and adapt these original synthetic examples for training and project preparation. Retain the scope statement when sharing an unmodified example. Do not use them as real survey evidence.