Course · ← 0014 Spatial Networks · capstone of the engine layer
Lesson 0015 · ~16 minutes
Lesson 0001 gave you four classes. The standard has packages, levels, spatial unit groups, boundary faces, an observation model, and — since Edition II — five parts. Here is the rest of it, and which parts you can safely ignore.
LADM is organised into three main packages plus a dedicated sub-package for spatial and surveying representation.1 All classes are prefixed LA_ to distinguish them from the rest of the ISO geographic information series.1
| Package | Core classes | Answers |
|---|---|---|
| Party | LA_Party, LA_GroupParty, LA_PartyMember | Who |
| Administrative | LA_BAUnit, LA_RRR (Right / Restriction / Responsibility), LA_AdministrativeSource | What claim, on what basis |
| Spatial Unit | LA_SpatialUnit, LA_SpatialUnitGroup, LA_Level, LA_LegalSpaceBuildingUnit, LA_LegalSpaceUtilityNetwork | Where |
| Surveying & Representation (sub-package) | LA_Point, LA_BoundaryFaceString, LA_BoundaryFace, LA_SpatialSource | How we know where |
LA_Level — a set of spatial units with a shared kind, dimension and registration type. This is the class that lets one country hold, in one model, a formal parcel level, a customary claims level, an informal settlement level and a utility network level, each with its own rules, over the same ground. Most national systems that "don't fit LADM" fit it perfectly once levels are understood.
LA_SpatialUnitGroup — spatial units aggregate into groups; for example all land parcels in a municipality could form a spatial unit group.1 Groups nest hierarchically, which is how administrative structure — district, block, adjudication area — is expressed without inventing your own parallel tree. Your adjudication_area from lesson 0007 is one of these.
LA_BoundaryFaceString — the boundary itself as a first-class object rather than an implicit polygon ring. The Surveying and Representation sub-package models the bounding elements of spatial units: points (LA_Point), 2D boundaries as linestrings (LA_BoundaryFaceString) and 3D boundaries (LA_BoundaryFace).2
Two parcels share one boundary — this is lesson 0011's topology argument arriving in the domain model. Make the boundary an object and you can attach evidence to it: which survey observed this line, which adjoiners agreed to it, what accuracy it carries, which record fixed it. Bury it in a polygon ring and every one of those facts has nowhere to live, so they end up on the parcel, duplicated and inconsistent between neighbours.
Here is where LADM Edition II directly meets the last four lessons. The surveying and representation sub-package has been refined with types of observation information — distance, level, angular, image, TPS, point cloud, GNSS and GPR.3
Read that list against what you now know. Distance and angular observations are COGO dimensions (lesson 0012). Image is imagery-traced capture (lesson 0009). GNSS is field survey. The standard is explicitly saying: record how each boundary was observed, in a typed way, so heterogeneous evidence can coexist in one register. That is the standards-backed answer to the design question lesson 0013 ended on — a surveyed fabric and fit-for-purpose imagery capture in the same system, each honestly labelled.
Which is also the standards-backed version of the capture_method and accuracy_class columns you sketched in lesson 0007. You reinvented a weaker form of this. Adopting the vocabulary costs nothing and makes your model legible to anyone who knows the standard.
| Part | Covers | Your concern |
|---|---|---|
| Part 1 — Overview & fundamentals ISO 19152-1:2024 | The core model, packages, the four classes | Essential |
| Part 2 — Land registration ISO 19152-2:2025 | Registration specifics, the refined survey model | Essential |
| Part 3 — Marine geographic information | Marine spaces and boundaries | Skip |
| Part 4 — Valuation information | Valuation units, sales, assessments | Later — if valuation is in scope |
| Part 5 — Spatial plan information | Planning zones and constraints | Later — if planning is in scope |
Say "LADM Part 1 and Part 2" in a proposal and you are current. Say "ISO 19152:2012" and you are describing Edition I, which is now the historical version.
LADM's purpose is not to replace existing systems but to provide a formal language for describing them, so that similarities and differences can be understood. That has a practical consequence people misread in both directions.
You are not obliged to name your tables LA_Party. Conformance is about being able to map your model onto the standard's classes without losing meaning — most real systems keep national terminology and publish an LADM profile alongside. And a country's law will always contain constructs the generic model does not name; extending LADM is expected practice, not a failure.
What conformance genuinely buys: your model is reviewable by anyone in the field, your data can be exchanged, and your proposal answers the standards question that donor evaluators now routinely ask. What it does not buy: correctness. An LADM-shaped model of the wrong country's law is still wrong.
"Refining the survey model of the LADM ISO 19152-2: Land registration", Land Use Policy — the Edition II survey and observation model, written by the people doing the refining:
sciencedirect.com/science/article/pii/S0264837724000772
For the package map without a paywall, the ICA wiki entry on ISO 19152 is accurate and short. The purchasable texts are ISO 19152-2:2025 (land registration) and Part 1, listed by CEN as EN ISO 19152-1:2024.