Inspect · the asset’s medical record
Condition is a trajectory.
A tower block, a bridge, a refinery, a 435-year-old monument — nothing stops existing at handover. An inspection here is not a visit that ends in an adjective. It is an epoch: registered to the same frame as every epoch before it, so a defect can be measured again, given a growth rate, and projected to the date it becomes someone’s decision.
Five things the record does with an asset after it is built: measures its condition, runs its statutory regimes, tracks the trend of every finding, witnesses hold points remotely on a dated epoch, and turns a rapid post-event survey into a quantity by subtracting the epoch before the loss.
- ASSET
- CHARMINAR
- CLASS
- HERITAGE MASONRY
- STANDING SINCE
- 1591 CE · 435 YR
- SURVEYED ENVELOPE
- 4 MINARETS · 56 M · 4 ARCHES
- FRAME
- ONE · VERSIONED r2
- RECORD OPENED
- EPOCH 01 · 2013
- EPOCHS ON RECORD
- 14
- OPEN FINDINGS
- 9 · I:3 II:4 III:2
Vitals · last epoch
- CRACK, WIDEST2.1 MM ±0.2 MMMEASURED · HUMAN-VERIFIED
- MINARET N, TILT41 MM / 56 M ±6 MMCONTROL NET · 8 STATIONS · FIX ±6 MM
- DELAMINATION, TOTAL3.42 M² ±0.11 M²MEASURED + SOUNDED · 2 PATCHES
Coverage ledger · 14 epochs
Heritage · Condition survey · Epoch 14
THE RECORD KEEPS IT STANDING.
Four minarets at 56 m, a public gallery, and nothing on it may be touched — so the survey flies it. Nine findings come back, each with its change against epoch 13. A condition is a photograph. A trajectory is a decision.
The register · Epoch 14
RANKED 09 · S1 3 · S2 4 · S3 2Every row is one object, not one visit. It opened at an epoch, it has been measured on the same registered geometry at every epoch since, and the column that decides anything is the last one before the method: what it did between epoch 13 and epoch 14.
The postcard
Condition · IN-01
Every defect is an object.
An inspection report is a snapshot of a visit: a list of things somebody saw once, described in adjectives, filed, and read by nobody until the next visit produces a different list. A finding here opens once and stays open — measured again in every epoch, on the same registered geometry, by a method that is named on every figure.
- FIRST OBSERVED
- EPOCH 06 · 2018
- CLASS
- CLASS I
- MEASURED, EPOCH 14
- 2.1 MM ±0.2 MM
- RATE
- +0.20 MM/YR ±0.02
Epoch register
| EPOCH | YEAR | MEASURED | ± | METHOD | VERIFIED |
|---|---|---|---|---|---|
| 06 | 2018 | 0.52 MM | ±0.20 | MEASURED · HUMAN-VERIFIED | INSTRUMENT |
| 08 | 2020 | 0.95 MM | ±0.20 | MEASURED · HUMAN-VERIFIED | INSTRUMENT |
| 10 | 2022 | 1.34 MM | ±0.20 | MEASURED · HUMAN-VERIFIED | INSTRUMENT + HUMAN |
| 12 | 2024 | 1.69 MM | ±0.20 | MEASURED · HUMAN-VERIFIED | INSTRUMENT + HUMAN |
| 13 | 2025 | 1.80 MM | ±0.20 | MEASURED · HUMAN-VERIFIED | INSTRUMENT |
| 14 | 2026 | 2.10 MM | ±0.20 | MEASURED · HUMAN-VERIFIED | INSTRUMENT + HUMAN |
Directive chain
- 01RAISED · EPOCH 13
- 02ROUTED · CONSERVATION ENGINEER
- 03ACKNOWLEDGED · 2025-06-04
- 04OPEN · RE-MEASURED AT EPOCH 14, STEP OUTSIDE ERROR BUDGET
Width is measured on the same registered polyline in every epoch. The step from epoch 13 is +0.30 mm against a combined tolerance of ±0.28 mm — outside the error budget, which is the only reason it is reported as growth rather than as noise.
The grammar is fixed and deliberately flat: subject, measured value, tolerance, method, evidence. No verbs of intent, no severity adjectives doing a number’s work, and a directive that closes only when the next capture shows the change — not when somebody ticks it.
Regimes · IN-02
The statute is a capture programme.
Every asset class carries an inspection law: what must be examined, to what standard, how often, by whom, and in what form the result must be filed. Read it as an engineer rather than a lawyer and it is already a mission specification — points, methods, tolerances, recurrence. The record compiles it, flies it, and returns the filing in the format the regulator reads, with the epoch stamped on every field.
Principal inspection · 24-month cycle
A bridge owner must examine every element on a fixed cycle and file a condition rating per element. The cycle is a capture programme with an unusually precise brief: the same soffit, the same bearing shelf, the same waterline, every two years, for the life of the structure.
- SCOPE
- ALL ELEMENTS · DECK, SOFFIT, BEARINGS, PIERS
- CADENCE
- 24 MONTHS · ±60 D WINDOW
- EXAMINER
- QUALIFIED ENGINEER · NAMED ON FILING
- EVIDENCE
- CONDITION RATING PER ELEMENT + IMAGERY
- FILING FORMAT
- ELEMENT REGISTER · OWNER SCHEMA
- POINTS
- 148 · 6 SPANS × 24 + 4 PIER SETS
- METHODS
- UAS SOFFIT PASS · GROUND 360 · CLOSE-RANGE
- TARGET DETAIL
- 3.0 MM DETAIL ON DEFECT SURFACES
- TOLERANCE
- CRACK WIDTH ±0.15 MM · SPALL DEPTH ±3 MM
- RECURRENCE
- EVERY 24 MONTHS · REFLOWN TO THE METRE
- GROUND BEAT
- BEARING SHELVES · HUMAN, ON FOOT
- STRUCTURE REF
- BR-2214
- ASSET REGISTER
- INSPECTION DATE
- 2026-05-14
- EPOCH 14 · SEALED
- DECK · CONDITION STATE
- 2 — MINOR DEFECTS
- EPOCH 14 · INSTRUMENT + HUMAN
- SOFFIT · CRACK, WIDEST
- 0.42 MM ±0.15
- EPOCH 14 · INSTRUMENT
- PIER 3 · SCOUR DEPTH
- 0.61 M ±0.08
- EPOCH 14 · WATERLINE PASS
- CHANGE SINCE LAST
- +0.06 MM ±0.21 — INSIDE ERROR BUDGET
- EPOCH 12 → 14
Regimes are configuration, not code. A new jurisdiction is a new schedule, a new form and a new set of tolerances — added the way a taxonomy is versioned, not the way a product is rewritten. Where seasons write the schedule, the pair of epochs around the season is the cleanest trend measurement an asset ever gets.
REGIME LIBRARY · CONFIGURATION · SPECIMEN FIELDS ILLUSTRATIVETrend · IN-03
Detection finds it. Trajectory dates it.
One survey reads 2.1 mm ±0.2 and calls it moderate. Eight surveys, registered to the same frame across eight years, give the same crack a rate of +0.20 mm per year with an uncertainty of ±0.02 — and a date, with a confidence interval, for when it reaches the limit the conservation engineer set. The first is an observation. The second is a maintenance budget, an outage window and a decision.
Episodic epochs and continuous instruments share the axis, each with its method and its uncertainty carried through: periodic imaging plus vitals, which is what a medical record has always been. The epoch that was never flown stays on the chart as a gap, because a series that hides its holes cannot be used to defend a rate.
The world's first inspection record where a defect carries a growth rate, and the growth rate carries its error budget.
Witness · IN-04
Witnessed remotely. On a dated epoch.
A hold point stops work until somebody with authority accepts the evidence in front of them. Make that evidence a registered epoch — dated where it was taken, and beyond restating afterwards — and the witness no longer has to be standing there for the release to be worth anything later. The platform prepares the release. A named human performs it under delegation of authority, and the release is recorded against the epoch that justified it.
- 01 · REQUESTEDHold point raised against WBS 04.12 — bearing installation, before cover-up. The work stops; the record knows why.
- 02 · CAPTUREDEPOCH 14 · 09:41 UTC · 1,284 frames · 97.4 % of the defined surface at 2.5 mm detail, seen within 35° of square.
- 03 · DATEDDated and placed where it was taken, at the moment it was taken — not when it reached an office, and not when somebody remembered to file it.
- 04 · SEALEDSealed 2026-05-11 10:02 UTC, when the walk closed. Nothing in it can be restated afterwards, and confirming that does not depend on us.
- 05 · WITNESSEDReviewed remotely against tolerance: setting-out deviation 6 mm measured ±2, against a 10 mm design tolerance.
- 06 · RELEASEDReleased under delegation of authority by a named person. The platform prepares the release; a human performs it.
- 07 · APPENDEDRequest, evidence, review and release are appended to the event log. Nothing is edited; agent-mediated steps are marked as such.
The evidence under a release is dated where it was taken and sealed when the walk closed. A witness satisfies themselves of that directly, rather than taking our word for it.
Witnessed remotely
- Setting-out and position against design tolerance, in mm
- Cover, spacing and lap of reinforcement, before a pour
- Surface condition and defect extent, before cover-up
- Coverage and completeness of the evidence itself
- Deviation of an installed element against the model
In person, always
- Destructive tests and sample-taking
- Load tests requiring instrumented control on site
- Confined-space entry certification
- Anything decided by touch — hammer sounding, torque, finish
- Calibration of the site instruments themselves
The right-hand column is published on purpose. A hold point released on the wrong class of evidence is worse than one that waited for a van.
Rapid · IN-05
The loss is a subtraction.
After a storm, a fire, a flood or an impact, everybody photographs the damage. Nobody can say how much of it is new. The cheapest moment to document damage is before it happens: with an epoch already on the record, the post-event survey is not a photo set but the second term of an equation, and the answer is a quantity with a tolerance in the form the people who pay for it already read.
| QUANTITY · METHOD | A · EPOCH 13 · 2026-05-18 | B · EPOCH 15 · 2026-08-02 | Δ = B − A |
|---|---|---|---|
| ROOF MEMBRANE, INTACTSURFACE PATCH · MESH | 4,120 M² ±38 | 3,006 M² ±31 | −1,114 M² ±49 |
| PARAPET, SOUND LENGTHSAFETY POLYLINE · MESH | 268 M ±2 | 214 M ±2 | −54 M ±3 |
| CLADDING PANELS, IN PLACEELEMENT COUNT · REGISTERED | 1,842 EA ±4 | 1,679 EA ±6 | −163 EA ±7 |
| AGGREGATE STOCKPILEVOLUMETRIC · ±2.4% | 18,400 M³ ±442 | 12,950 M³ ±311 | −5,450 M³ ±540 |
| COLUMN B12, PLUMB DEVIATIONREGISTERED MESH · LOCAL | 4 MM ±2 | 23 MM ±2 | +19 MM ±3 |
Δ carries both epochs’ uncertainties combined in quadrature. Subtracting two measurements does not subtract their error budgets.
SPECIMEN ASSET · ILLUSTRATIVE RECORD · 5 OF 34 LINES SHOWN- 01 · EVENT2026-08-01 03:10 LOCAL
- 02 · CAPTURE ORDEREDT+21 H · ACCESS CLEARED
- 03 · CAPTURE FLOWNT+29 H · SAME MISSION AS EPOCH 13
- 04 · EPOCH REGISTEREDT+31 H · DATED, SEALED
- 05 · DIFFERENCE COMPUTEDT+31 H · AGAINST EPOCH 13
Stages with their measured intervals, for one specimen event. Not a turnaround promise — a record that publishes tolerances does not publish a number it has not measured across enough events to defend.
- CONDITION OF RECORD AT INCEPTION
- The pre-loss epoch, dated where it was taken and sealed when the walk closed — made before the event, not reconstructed after it.
- MEASURED LOSS PER ELEMENT
- Quantity, unit, tolerance and method for every line, on the same registered geometry in both epochs.
- EVIDENCE FRAMES
- The frames behind each figure, dated at capture, reachable from the figure itself.
- EXTENTS ON THE SITE PLAN
- Loss extents drawn in the asset's own coordinate frame, so they overlay the drawings already in the file.
- CHRONOLOGY
- Event time, capture time and the interval between them, stated as measured stages rather than a promised turnaround.
The dossier delivers into the instruments the recipients already work in — the adjuster’s estimating platform, the independent engineer’s certificate, the lender’s draw file. Nobody is asked to learn our software to read a loss.
What the record is worth later
Retire, repair, or run.
The decision an owner actually makes is not whether a defect exists. It is whether to retire the asset, repair it now, or run it and watch — and that decision is made on a trajectory, an error budget and a date, not on the last inspector's adjective. Every mark on this line is a moment somebody opens the record and asks it about a date that has already passed.
- CONSTRUCTIONYR 0–2 · EPOCHS 01–24
“Is what we are paying for actually installed?”
Measured quantity in place, epoch by epoch, against the claim on the pay application.
- HANDOVERYR 2 · HANDOVER EPOCH
“What exactly was handed over?”
The handover epoch, including the services captured before the drywall went on.
- RETENTION RELEASEYR 2–4 · DLP END
“Has the defects liability period been served?”
Handover epoch against the DLP-end epoch: every difference listed, each with its tolerance.
- INSURANCE RENEWALANNUAL
“What condition is being insured?”
The last sealed epoch as condition of record at inception — dated before anything happens.
- REFINANCE · VALUATIONYR 5
“What state is this asset really in?”
The condition register with a measured trajectory per finding, and the record's own gaps declared.
- EVENT · CLAIMANY TIME
“How much was lost?”
Post-event epoch minus pre-loss epoch, per element, with the two tolerances carried into the difference.
- DISPUTEYR 3–8
“What was true on 14 March, and what did we know then?”
The record as it stood on that date, carrying both dates: when it happened and when it was known.
- REFURBISHMENTYR 12
“What is behind that wall?”
The epochs captured before cover-up — the only time the inside of the building was ever visible.
- SALE · TRANSFERYR 15
“What is the buyer inheriting?”
The whole record, gaps included. A record that names its holes is the only kind worth pricing.
- RE-TENDERYR 15+
“What does maintaining this actually cost?”
Measured defect rates and closed directives across fifteen years, not the incumbent's account of them.
Every mark is a moment somebody pays — in fees, in exposure, or in a number they have to accept from the other side — to know what the condition was on a date that has already gone. The answer cannot be captured retroactively.
CONSULT LINE · 10 MOMENTS · ONE RECORD · RAIL PANS →Questions of record
The limits, in the headline register.
Prerequisites, thresholds and refusals belong at the top of a specification, not in the small print under it. These are the answers a serious owner asks for on the second call.
- How do I survey the condition of an existing building?
- The asset gets a medical record: one first epoch, measured to a stated tolerance over a defined surface envelope, against which every later change is read. A single visit describes an asset; a first epoch gives every visit after it something to be measured against.
- EPOCH 01 · THE BASELINE EVERYTHING IS READ FROM
- What does the record need before the first survey?
- A defined surface envelope, a coordinate frame with a version on it, access windows, and the tolerance each finding class must be measured to. Without those four, the first flight produces photographs; with them, it produces epoch 01.
- PREREQUISITE · 4 ITEMS · BEFORE EPOCH 01
- What counts as seen well enough on an existing asset?
- For the heritage specimen on this page: ≥ 92 % of the defined surface at 4 mm detail with incidence ≤ 35°, and baseline sufficient for the relief of the fabric. The capture instrument states whether the epoch met that while the operator is still standing there — not a week later.
- COVERAGE ≥92% · 4.0MM DETAIL · INCIDENCE ≤35°
- How many epochs before a growth rate can be published?
- Until the uncertainty on the fit is smaller than the change the fit claims, the record reports measurements and declines the rate. A trajectory that cannot outrun its own error budget is not a trajectory yet, and saying so is cheaper than being wrong in a filing.
- RATE · DECLINED UNTIL THE FIT SUPPORTS IT
- Can this be done with an ordinary phone?
- For accessible surfaces, yes — phone capture with live coverage feedback produces a graded epoch. For 56 m of masonry at 4 mm detail on a protected fabric, no. When the phone cannot meet the error budget the instrument says so and names the capture class that would.
- PHONE · GRADED · BOUNDED BY ERROR BUDGET
- The asset is occupied, and open to the public.
- Most of them are. Capture runs in public hours where it must; the record holds geometry and condition, not people. No facial recognition, and no identification of a worker without cause released under delegation of authority.
- OCCUPIED · PUBLIC HOURS · NO PERSON IDENTIFIED
- What happens when a survey window is missed?
- It is held as a gap in the coverage ledger, drawn and labelled. A missed epoch is a fact about the record, and a record that quietly interpolates over its holes is worth less than one that names them.
- GAP · DRAWN · NEVER A ZERO
- Who can see what?
- Permission is settled before an answer is composed, right down to one zone: a person who is not cleared for a zone does not receive an answer computed from that zone. The restriction is upstream of the answer, not a filter applied to one.
- ACCESS · PER ZONE · SETTLED FIRST