Interoperability

Interoperability for AEC Computational Design

1. The Interoperability Challenge in AEC

1.1 Why Interoperability Matters

Interoperability -- the ability to exchange data between software tools without loss of meaning, geometry, or relationships -- is the single most critical infrastructure problem in the AEC industry. Every building project involves dozens of software tools, hundreds of files, and thousands of data exchanges. When those exchanges fail, the consequences are measured in millions of dollars and months of delay.

The AEC industry loses an estimated $15.8 billion annually in the United States alone due to inadequate interoperability (NIST GCR 04-867). This figure accounts for redundant data entry, manual format conversion, error correction from data loss, and delayed decision-making caused by information silos.

Unlike the manufacturing or aerospace industries, which converged on STEP/IGES decades ago, the AEC sector remains fragmented across proprietary ecosystems. Autodesk, Bentley, Trimble, Nemetschek, and dozens of smaller vendors each maintain walled gardens with varying degrees of openness. The result is a landscape where a single design decision may need to be re-entered into five or more tools before it reaches a construction site.

1.2 Single-Source-of-Truth vs. Federated Model Approaches

Single-Source-of-Truth (SSOT):

• One authoritative model from which all views and deliverables derive
• Revit-centric workflows often attempt this, with one central model containing architecture, structure, and MEP
• Advantages: no synchronization burden, clear ownership, simpler version control
• Disadvantages: tool lock-in, performance limits at scale, inability to leverage best-of-breed tools
• Practical limit: SSOT breaks down beyond approximately 200-300 MB models or when disciplines require specialized solvers

Federated Model Approach:

• Multiple discipline-specific models linked through coordination mechanisms
• Each discipline uses its optimal tool (Revit for documentation, Rhino for complex geometry, Tekla for steel detailing, ETABS for structural analysis)
• Coordination via shared coordinates, reference planes, and clash detection (Navisworks, Solibri, BIMcollab)
• Advantages: best-of-breed tooling, team autonomy, distributed workload
• Disadvantages: synchronization overhead, version mismatch risk, coordinate alignment complexity
• Industry trend: federated approaches are winning, especially with Speckle and IFC enabling richer exchange

1.3 Open Standards vs. Proprietary Formats

Open Standards:

• IFC (Industry Foundation Classes) -- ISO 16739, the only truly open BIM exchange standard
• gbXML (Green Building XML) -- energy simulation exchange
• CityGML / CityJSON -- urban-scale 3D models
• LandXML -- civil engineering survey and design data
• BCF (BIM Collaboration Format) -- issue tracking tied to model viewpoints
• Governed by buildingSMART International, OGC, and other standards bodies

Proprietary Formats:

• RVT/RFA (Revit), DWG (AutoCAD), 3DM (Rhino), SKP (SketchUp), PLA/PLN (ArchiCAD)
• Offer full fidelity within their ecosystem
• Risk: vendor lock-in, obsolescence, licensing dependencies
• Some are partially documented (DWG via Open Design Alliance) but reverse-engineered support is always incomplete

Pragmatic Reality: Most production workflows use a hybrid. Native formats for authoring, open formats for exchange, and lightweight formats (glTF, PDF) for communication. The goal is not eliminating proprietary formats but creating robust translation layers.

1.4 Data Loss Taxonomy

When data moves between tools, losses occur in four categories:

| Loss Type | Description | Example | Impact | |-----------|-------------|---------|--------| | Geometry Loss | Shape information degraded or missing | NURBS surface exported to STL loses curvature continuity | Visible faceting, dimensional inaccuracy | | Metadata Loss | Properties, parameters, classifications stripped | Revit wall type info lost when exporting to OBJ | Downstream tools lack decision-critical data | | Relationship Loss | Connections, hosting, spatial hierarchy broken | Wall-floor join lost in IFC export | Manual rework to re-establish element logic | | Appearance Loss | Materials, textures, colors not transferred | PBR materials from Blender not mapping to Revit | Re-application of visual properties in target tool |

Additional nuanced losses:

• Precision loss: floating-point truncation at large coordinates
• Semantic loss: a "wall" becomes a generic "extrusion" in target tool
• Topological loss: solid body becomes disjoint surfaces
• Behavioral loss: parametric constraints become fixed geometry
• Unit loss: implicit unit assumptions cause scaling errors (mm vs. ft is a classic)

---

2. File Format Encyclopedia

2.1 Geometry Formats

OBJ (Wavefront Object)

| Property | Value | |----------|-------| | Extension | .obj (geometry), .mtl (materials) | | Version | Originally 1992, no formal versioning | | Geometry | Polygonal mesh, free-form curves/surfaces | | Metadata | Minimal -- group names, material references | | Max Size | No hard limit; practical ~500 MB | | Typical Use | Mesh exchange, visualization, 3D printing prep | | Read/Write | Rhino, Blender, 3ds Max, SketchUp, Unity, Unreal, MeshLab, CloudCompare | | Strengths | Human-readable ASCII, universal support, simple specification | | Limitations | No BIM data, no solid topology, no units, large file sizes for complex models |

STL (Stereolithography)

| Property | Value | |----------|-------| | Extension | .stl | | Version | Original (1987), no updates | | Geometry | Triangulated mesh only | | Metadata | None (only triangle normals and vertices) | | Max Size | No limit; practical ~200 MB for ASCII, larger for binary | | Typical Use | 3D printing, CNC machining, rapid prototyping | | Read/Write | Every CAD tool, every slicer, every mesh editor | | Strengths | Universal 3D printing standard, trivial to parse | | Limitations | No color, no materials, no units, no metadata, triangles only, redundant vertex storage |

3MF (3D Manufacturing Format)

| Property | Value | |----------|-------| | Extension | .3mf | | Version | 1.2.3 (current) | | Geometry | Triangle mesh with manifold validation | | Metadata | Color, materials, print tickets, textures, build platform layout | | Max Size | ZIP-compressed, efficient for large models | | Typical Use | Advanced 3D printing with color/material, digital fabrication | | Read/Write | Rhino, PrusaSlicer, Cura, Windows 3D Viewer, Materialise | | Strengths | Modern replacement for STL, supports multi-material, compact | | Limitations | Not yet universal, limited AEC adoption, no parametric data |

PLY (Polygon File Format / Stanford Triangle Format)

| Property | Value | |----------|-------| | Extension | .ply | | Version | 1.0 (1994) | | Geometry | Point cloud and/or polygonal mesh | | Metadata | Per-vertex color, normals, custom properties | | Max Size | Binary format handles billions of points | | Typical Use | Point cloud storage, 3D scanning output, research | | Read/Write | CloudCompare, MeshLab, Blender, Open3D, PCL | | Strengths | Flexible schema, binary efficiency, extensible vertex properties | | Limitations | No materials/textures in standard spec, no BIM data |

3DM (Rhino 3D Model)

| Property | Value | |----------|-------| | Extension | .3dm | | Version | openNURBS 8.x (Rhino 8) | | Geometry | NURBS surfaces, curves, meshes, SubD, extrusions, points, annotations | | Metadata | Layers, object attributes, user text, render materials, named views | | Max Size | No hard limit; practical ~2 GB | | Typical Use | Rhino native authoring, computational design output | | Read/Write | Rhino, Grasshopper, openNURBS SDK (C++, .NET), Speckle, many viewers | | Strengths | Full NURBS fidelity, open SDK (openNURBS), rich layer/attribute system | | Limitations | No BIM semantics, limited structural metadata, Rhino-centric ecosystem |

DWG / DXF (AutoCAD Drawing / Drawing Exchange Format)

| Property | Value | |----------|-------| | Extension | .dwg, .dxf | | Version | DWG 2018 (R2018), DXF tracks DWG versions | | Geometry | 2D entities (lines, arcs, polylines, hatches), 3D solids (ACIS), meshes, surfaces | | Metadata | Layers, blocks, attributes, extended data (XDATA), object properties | | Max Size | Practical ~500 MB | | Typical Use | 2D drafting, CAD exchange, legacy drawing archives | | Read/Write | AutoCAD, BricsCAD, Rhino, Revit (import), QGIS, LibreCAD, FreeCAD | | Strengths | Industry standard for 2D, massive legacy archive, block/attribute system | | Limitations | DWG is proprietary (ODA reverse-engineers), DXF is verbose, 3D support limited |

SKP (SketchUp)

| Property | Value | |----------|-------| | Extension | .skp | | Version | SKP 2024 | | Geometry | Polygonal mesh, groups, components | | Metadata | Layers (tags), component definitions, material assignments, geolocation | | Max Size | Practical ~300 MB | | Typical Use | Conceptual design, massing studies, early-stage visualization | | Read/Write | SketchUp, Trimble Connect, various importers (Rhino, Blender via plugins) | | Strengths | Intuitive modeling paradigm, large 3D Warehouse library, geolocation | | Limitations | Imprecise geometry, no NURBS, no parametric constraints, limited BIM |

STEP / IGES (Standard for Exchange of Product Data / Initial Graphics Exchange Specification)

| Property | Value | |----------|-------| | Extension | .step, .stp, .iges, .igs | | Version | STEP AP214/AP242 (ISO 10303), IGES 5.3 | | Geometry | NURBS surfaces, B-rep solids, curves, wireframe | | Metadata | Product structure, material (limited), PMI (AP242), assembly hierarchy | | Max Size | Multi-GB for complex assemblies | | Typical Use | Mechanical CAD exchange, manufacturing, CNC toolpath input | | Read/Write | SolidWorks, CATIA, NX, Rhino, FreeCAD, Inventor, Fusion 360 | | Strengths | Neutral CAD exchange, precise B-rep, ISO standard, AP242 adds PMI | | Limitations | Large files, slow parsing, IGES is legacy (use STEP), limited AEC adoption |

SAT (ACIS Save As Text)

| Property | Value | |----------|-------| | Extension | .sat, .sab (binary) | | Version | ACIS R2024 | | Geometry | B-rep solids, NURBS surfaces, curves, sheets | | Metadata | Minimal (body names, attributes) | | Max Size | Practical ~500 MB | | Typical Use | Solid geometry exchange between ACIS-kernel tools, Revit mass import | | Read/Write | AutoCAD, Revit (import), SpaceClaim, Fusion 360, BricsCAD | | Strengths | Exact B-rep, Revit can import as mass/generic model, clean geometry | | Limitations | Proprietary kernel (Spatial Corp), no BIM semantics, limited ecosystem |

2.2 BIM Formats

RVT / RFA (Revit Project / Family)

| Property | Value | |----------|-------| | Extension | .rvt (project), .rfa (family), .rte (template) | | Version | Revit 2025 | | Geometry | Parametric solids, extrusions, sweeps, blends, voids, meshes (limited) | | Metadata | Rich: categories, families, types, instances, parameters, schedules, phases, worksets | | Max Size | Practical ~500 MB (workshared models can exceed) | | Typical Use | BIM authoring, construction documentation, coordination | | Read/Write | Revit only (native), IFC export, various viewers (Navisworks, BIM360/ACC) | | Strengths | Full BIM fidelity, parametric families, scheduling, documentation | | Limitations | Completely proprietary, requires Revit license to edit, large file size |

IFC (Industry Foundation Classes)

| Property | Value | |----------|-------| | Extension | .ifc (STEP), .ifcXML, .ifcZIP, .ifcJSON | | Version | IFC4.3 (ISO 16739-1:2024), IFC4x3 ADD2 | | Geometry | B-rep, CSG, swept solids, tessellated (triangulated), curves, point clouds | | Metadata | Complete BIM: spatial structure, element types, property sets, quantities, materials, classifications, relationships, cost, time | | Max Size | Multi-GB for large projects (IFC4 has improved efficiency) | | Typical Use | OpenBIM exchange, regulatory submissions, archival, coordination | | Read/Write | Revit, ArchiCAD, Tekla, Solibri, BIMcollab, Navisworks, FreeCAD, BlenderBIM, xBIM, IfcOpenShell | | Strengths | Only true open BIM standard, ISO-certified, rich semantic model, vendor-neutral | | Limitations | Inconsistent export quality across tools, complex schema, geometry fidelity varies, round-trip editing unreliable |

NWD / NWC (Navisworks)

| Property | Value | |----------|-------| | Extension | .nwd (full), .nwc (cache), .nwf (reference) | | Version | Navisworks 2025 | | Geometry | Tessellated mesh (view-only, no editable geometry) | | Metadata | Aggregated from source models, clash results, timeliner schedules, viewpoints | | Max Size | Multi-GB (designed for large federated models) | | Typical Use | Clash detection, 4D simulation, model review, coordination | | Read/Write | Navisworks (native), BIM360/ACC viewer, Freedom (free viewer) | | Strengths | Handles massive models, clash detection engine, 4D timeliner | | Limitations | View-only (no editing), proprietary, Autodesk ecosystem only |

gbXML (Green Building XML)

| Property | Value | |----------|-------| | Extension | .xml (with gbXML schema) | | Version | 7.03 | | Geometry | Simplified planar surfaces (walls, floors, roofs as polygons), zones | | Metadata | Thermal properties, construction assemblies, schedules, HVAC zones, location/climate | | Max Size | Typically Export to intermediate format (IFC, DXF, SAT, OBJ, etc.) -> Import into target tool -> Manual cleanup and re-association.

When to use: One-time or infrequent transfers, when live linking is not available, when tools are on different machines/networks, when a frozen snapshot is needed.

Key considerations:

• Always verify export settings (version, units, coordinate system, included categories)
• Document the conversion path for reproducibility
• Validate geometry and metadata in the target tool immediately after import
• Maintain a log of known data losses for your specific tool combination

3.3 Live Linking

Real-time or near-real-time bidirectional connection between tools running simultaneously.

Technologies:

• Rhino.Inside.Revit: Rhino and Grasshopper running inside Revit's process, sharing geometry and data live
• Dynamo ↔ Revit: Dynamo scripting within Revit, direct access to Revit API
• Grasshopper ↔ Tekla Live Link: Real-time structural model exchange
• Revit ↔ Robot Structural Link: Analytical model exchange for structural analysis
• Excel ↔ Revit (Dynamo): Live parameter read/write via Dynamo Excel nodes

When to use: Iterative design exploration requiring immediate BIM feedback, parametric facade design that must update Revit curtain panels, structural optimization with real-time analysis results.

3.4 Data Streaming (Speckle and Similar)

Continuous, version-controlled data flow between tools via a cloud or self-hosted intermediary.

Speckle is the leading open-source platform for AEC data streaming. It provides:

• Object-level versioning (not file-level)
• Connectors for 15+ AEC tools
• GraphQL API for custom integrations
• Web-based 3D viewer for review
• Automation triggers on model changes

When to use: Multi-discipline teams using different tools, continuous integration for design models, when audit trail and version history are required.

3.5 Database-Mediated Excha

…

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: Amanbh997
• Source: Amanbh997/Claude-skills-for-Computational-Designers
• License: MIT

Install and usage instructions live in the source repository linked above.