Layer standards are treated as administrative housekeeping until the moment a consultant's file arrives with 340 layers named after the person who drew it. Then they become the difference between a two-hour task and a two-day one.
Firms working across US, UK, and Australian projects face an additional complication: three regions, three standards lineages, and clients who specify whichever one their internal manual was written around. This article compares the major standards structurally and sets out how to run one office system across all of them.
Why structured layer names exist
A layer name is a data field compressed into a string. The purpose of a standard is to make that string machine-readable so that:
- Files can be filtered, isolated, and frozen by discipline or element type without manual selection
- Multiple consultants' files can be overlaid and understood
- Plot styles map predictably to line weights
- Elements can be extracted for quantity or analysis
- New staff and outsourced teams can work without a per-project briefing
Any convention that achieves this is defensible. The specific one matters less than consistency, but interoperability pushes most firms toward a recognized standard.
The AIA / US National CAD Standard structure
The US convention, formalized through the National CAD Standard and derived from the AIA CAD Layer Guidelines, uses a field-based name:
Discipline – Major Group – Minor Group (optional) – Status (optional)
For example, A-WALL-FULL-N:
A— Architectural disciplineWALL— major group, the elementFULL— minor group, full-height wallsN— status, new work
Discipline codes cover the familiar set: A architectural, S structural, M mechanical, P plumbing, E electrical, F fire protection, C civil, L landscape, I interiors, Q equipment, G general.
Status codes are the most under-used and most useful part of the standard, particularly on renovation work: N new, E existing to remain, D demolish, R relocated, X existing to be removed. A renovation set with correct status coding can generate demolition and new-work plans from one file by layer state.
The standard is well suited to drawing production, widely understood in the US market, and has strong tool support.
BS 1192 and ISO 13567
The British Standard lineage, and the international ISO 13567 standard it aligns with, structures names by field with an agent, element, and presentation logic. A typical BS 1192 layer name follows:
Role – Classification – Presentation – Description
For instance, A-25-M-Walls, where the role field identifies the originating discipline, the classification field carries a Uniclass or CI/SfB code for the element, the presentation field distinguishes model geometry from text and dimensions, and the description is human-readable.
The distinguishing feature is classification coding. Where the AIA standard uses a mnemonic (WALL), the BS lineage uses a formal classification code drawn from a maintained taxonomy. This is more rigorous, integrates with cost and specification systems, and is less immediately legible to a drafter who has not memorized the codes.
The presentation field is genuinely valuable and largely absent from the US convention: separating model geometry, annotation, and dimensions onto distinct layers makes it possible to reuse geometry across drawings at different scales without dragging incompatible annotation with it.
Where ISO 19650 fits
ISO 19650 is frequently cited as a CAD layer standard. It is not one. It governs information management across the project lifecycle — container naming, the common data environment, status codes, revision, and approval workflows. It defines how a *file* is named and controlled, not how *layers* inside it are named.
Projects specifying ISO 19650 therefore still need a layer convention alongside it, typically the BS 1192 lineage or a national annex. Conflating the two leads to briefs that specify a file-naming standard and leave layer structure undefined — a gap that shows up in the first coordination cycle.
Australian practice largely follows the ISO 19650 framework for information management, with layer conventions drawn from either the ISO 13567 lineage or client-specific manuals, and a significant proportion of US-influenced practice in the private sector.
Running one office standard across three markets
Maintaining three parallel standards is unmanageable. The practical approach used by firms working internationally has three parts.
Adopt one internal production standard. Choose the one that fits the majority of your work and the templates you already own. Draw everything in it.
Build translation maps. A layer translation table — supported natively in AutoCAD's Layer Translator and equivalent in most platforms — maps your internal names to a client's required standard at issue. Build one map per recurring client standard, test it once, and reuse it.
Enforce at issue, not during production. Translating at the point of issue is a scripted, verifiable operation. Asking drafters to switch conventions per project produces mixed files that no map can fix.
The critical enabler is that your internal standard must be at least as granular as every standard you translate into. Translation can merge layers reliably; it cannot split them. If a client standard separates load-bearing from non-load-bearing partitions and yours does not, no map will recover the distinction.
Beyond layers
A layer standard alone does not make files interoperable. A complete CAD standards manual should also fix:
- Text styles — a small named set, with defined fonts and heights by plot scale
- Dimension styles — named, with defined precision, arrow style, and unit format
- Line types and line weights — assigned by layer, never by object
- Colour-to-plot-style mapping — the CTB or STB file issued with every drawing set
- Blocks — naming, insertion point convention, layer-zero construction, and attribute schedules
- Title block and sheet setup — one authoritative template
- Xref protocol — relative paths, overlay versus attach, and binding rules at issue
- Model and paper space rules — geometry in model space at full scale, annotation and viewports in paper space
Line weight by object rather than by layer is the most persistent bad habit in CAD production, and it defeats every plot standard downstream.
Enforcement
Standards that are documented but unenforced degrade within a quarter. Three mechanisms work:
A locked template. Layers, styles, and title block pre-loaded. Nothing is created from scratch.
A pre-issue audit script. Automated check for non-standard layers, objects on layer zero, overrides, unresolved xrefs, and unpurged content. Run before every issue.
A short standards document. Ten to fifteen pages that a new drafter or an outsourced team can absorb in an hour. Fifty-page manuals are written, filed, and never read.
Mapping model categories to layers on export
Firms producing drawings from a model still need layer discipline, because every DWG export maps model categories to layer names. Left at software defaults, the export produces layer names that match no standard and vary between projects.
The export mapping file should be:
- Authored once for each layer standard the firm issues under
- Held in the office template, not configured per project
- Tested by exporting a representative sheet and auditing the result against the standard
- Version controlled, since a change to it silently affects every subsequent export
The mapping should also handle subcategories, since detail lines, model lines and annotation subcategories are exactly where uncontrolled exports produce dozens of unwanted layers.
A related discipline: decide whether exported drawings are issued with references bound or attached, and whether model space or paper space geometry is delivered. Recipients differ in what they expect, and stating it in the issue transmittal prevents the round of emails that otherwise follows every export.
Receiving files to a standard you do not use
The reverse problem — consultant files arriving in an unfamiliar convention — is handled the same way, in reverse. Build a translation map from their standard to yours, apply it on receipt, and keep the original untouched as the received record.
Two rules make this safe:
Never edit a received file in place. Copy it, translate the copy, and retain the original with its receipt date and revision recorded.
Translate on receipt, not at use. A file translated when it arrives is consistent for everyone who later references it. A file translated by whoever happens to need it produces several divergent copies.
Where a consultant's file is so disorganised that translation is not possible — layers named by author, geometry on layer zero, everything by object override — the correct response is to raise it immediately rather than to absorb it. Absorbing it silently sets the expectation for every subsequent issue.
Frequently asked questions
Do layer standards still matter with BIM? Yes, in two ways. 2D CAD remains in heavy use for details, shop drawings, and civil work; and models exported to CAD map categories to layers, so the export mapping needs the same discipline.
How many layers should a typical set use? An architectural set commonly runs to 40–80 active layers. Files with several hundred are usually the result of consultant merges and unaudited xrefs rather than genuine need.
Which standard should a new firm adopt? The one your primary market expects, applied strictly, with translation maps for the exceptions.
Related reading: PDF to CAD Conversion: What Accuracy Actually Means · DWG File Management for Distributed Teams · The BIM Execution Plan: What to Put In It and What to Leave Out
Vantage CAD Services produces drawings to client CAD standards, including layer translation and pre-issue standards audits. Contact info@vantagecadservices.com or +1 (512) 543-0831.