Site3D and place-authoring tools are often compared as if they were alternatives. They are not. One proves that a site functions and can be built; the other decides what the site should be. The engineering question is "does the road meet the grade and drain?" The design question is "what does it feel like to walk through this courtyard?" A project needs both answers, produced by different object models, reconciled at a seam that nobody owns by default.
TL;DR
| Dimension | Site3D | GenAnywhere |
|---|---|---|
| Role | Site and road engineering | Place authoring |
| Core objects | Alignments, profiles, junctions/roundabouts, drainage networks, surfaces, ponds | Authored place and revisions |
| Questions | Grades, drainage, junctions, earthworks, setting-out | Character, sequence, public realm, assets |
| Revision | Parametric edits to alignments/networks | Edit in place; approved state persists |
| Exchange | IFC for BIM; setting-out tables for field | Design state |
| Platform | Windows desktop | Design seats |
| Best role | Function and constructability | Intent and experience |
Rule zero: engineering proves; design decides. A model that does only one of those jobs is not a site model — it is half a project.
Site3D's Object Model
Site3D is a UK-developed civil design package built around site and road work. Its value is a compact engineering object model:
- Alignments and profiles — horizontal and vertical control for roads and access routes.
- Automated junctions and roundabouts — geometry derived from connecting roads, with curbs, islands, and transitions that follow the change.
- Dynamic drainage — networks and levels that react when the surface or grading moves.
- Surfaces and earthworks — cut and fill, grading, and pond design in the same environment.
- IFC export — the engineering model enters BIM coordination as an exchange object.
- Setting-out tables — construction-ready coordinates for the field.
The model is built for constructability, and it is the right one for that purpose. It is intentionally narrow: no API, limited third-party integrations, and no pretense of authoring the place.
GenAnywhere's Model
GenAnywhere has a different unit: the authored place. Its objects are relationships — how the massing meets the street, how the public realm is enclosed, how a sequence unfolds — and its central operation is revision. Approved elements persist; contested ones move.
This is not a lesser version of a civil model. It answers a question Site3D does not ask: what should the site be?
Revision Semantics Compared
Site3D
- Engineering change → edit the alignment, profile, network, or surface; dependent geometry regenerates.
- Continuity → parametric, but not approval-aware; the model does not know which elements were signed off.
- Iteration cost → proportional to the engineering change, not the design intent.
- Record → drawing sets, versions, and transmittals.
GenAnywhere
- Design change → edit the authored place.
- Continuity → built in; approved elements stay fixed.
- Iteration cost → proportional to the scope of the decision.
- Record → the design state carries the decisions.
The failure pattern is predictable: a design change is pushed into Site3D as a geometry edit, and the engineering model silently drifts from the approved design. Or an engineering constraint never makes it back into the design model, and the place cannot be built as drawn.
IFC: What Actually Travels
Site3D's IFC export is genuinely useful, but it is worth being precise about what crosses the boundary. A coordinated IFC model carries geometry and classification; it generally does not carry the parametric intent of the alignment, the design rationale behind a curb placement, or an approval state. When architects receive the IFC, they receive a realization, not the design conversation that produced it.
That asymmetry is the reason a separate design state matters. The authored place in GenAnywhere records what was decided and why; the IFC records one implementation of that decision. Treating the IFC as the source of design truth is the civil-industry equivalent of treating a screenshot as a spec.
Practical consequences:
- Never round-trip design intent through IFC. Export for coordination, not for editing. Re-authoring from an IFC import loses the revision history that made the design defensible.
- Name the versions. The IFC export and the design state should carry matching dates and revision letters, or the next coordination meeting will be spent reconciling calendars.
- Expect loss, plan for it. Junction edges, drainage connectivity, and surface boundaries are the usual casualties. Document which objects are authoritative in which model.
The Seam: Practical Protocol
- Shared datum first. Agree on coordinate system, units, and origin before any overlay. A single mismatch invalidates every downstream check.
- Name the surfaces. Existing ground, proposed design surface, and finished grade should be distinct, versioned objects with owners.
- Separate intent from realization. The authored place is intent; the Site3D alignment and surface are one realization. Changes to either must be logged against the other.
- Treat geometry conflicts as decisions. When the required curve eats the corner plaza, record the resolution — who approved it, and what changed.
- Exchange in the right format. IFC for coordination; native formats for engineering edits; setting-out tables for construction. Document the export settings.
- Check thresholds and clearances. Turning radii versus drop-off zones, pipe cover versus tree pits, accessible routes versus design slopes.
- Version the pair. The design state and the civil model should travel together, with dates and authors.
Where Each Tool Is Correct
- Site3D: road and estate layout, junctions and roundabouts, drainage networks, earthworks, ponds, setting-out, constructability review, permit sets.
- GenAnywhere: public realm, street character, sequence, the authored place and its assets, and the revision loop that carries it through approval.
QA Checklist for the Handoff
- Coordinates, units, and origin verified in both models?
- Surface names and boundaries documented?
- Design intentions versus engineered geometry reconciled and recorded?
- Junction and driveway geometry checked against the authored place?
- Drainage paths verified against plaza and landscape design?
- IFC export version matched to the approved design state?
- Setting-out tables generated after the final revision?
Verdict
Site3D is the right instrument for the engineering question, and no design tool replaces it. GenAnywhere is the right instrument for the design question, and no civil model answers it. The work is to wire them together with an explicit seam: shared datum, distinct surfaces, recorded conflicts, and versioned pairs.
A site that only drains is not a place. A place that cannot drain is not a project. The handoff is where both become true.


