Modelling plan
The chosen sequence of features and the references each one uses. The single most powerful expression of intent, and the hardest to retrofit.
Definition
Intent is not functionality
Functionality describes what a design does. Intent justifies why it is built the way it is. A hole is functional because a fastener passes through it; the intent is the reason it sits 20 mm from the datum face rather than centred, and the reason it should stay 20 mm from that face when the part grows.
The distinction has real consequences. Two models can be geometrically identical and behave completely differently under edit, because one captured the intent and the other captured only the outcome. In a parametric system, intent is expressed through the choice of references, the order of features, the relations locking each sketch, and the mates holding an assembly together.
Why it is contested
Design intent is not a settled, well-bounded topic. Some practitioners treat it as the set of equations, geometric relations and mating conditions in a model. Others hold that it is the sequence in which a part is created. A third group argues that intent means building the part the same way it will be manufactured, and speaks of manufacturing intent as a distinct idea. All three readings are useful, and this page treats them as complementary rather than competing.
There is a structural gap behind the disagreement. Because design and manufacturing revolve around geometric shape, CAD systems are built on geometric modelling techniques — and geometric models are poor at recording embodiment detail such as rationale, alternatives considered, manufacturing specification and constraint. A CAD model answers what the shape is. It struggles with how is this supposed to work, why was it done this way, what alternatives were rejected, what can be changed and what will be affected.
Mechanisms
Four ways intent is captured in geometry
Sketch relations
Parallel, perpendicular, tangent, concentric, symmetric. These say that a relationship must hold, not merely that it currently does.
Equations
Arithmetic dependencies between dimensions, encoding proportion and derived sizing directly into the model.
Assembly mates
How components connect. Good mates mimic the physical constraint; poor mates merely achieve the right picture.
The test that matters
With sound design intent, changes propagate cleanly through part, assembly and drawing. With poor intent, features update in unintended ways or fail outright. The practical test is to ask, before committing a reference: if this dimension doubled, or halved, or reversed sign, would this feature still be where it belongs?
Decision checklist
The questions a designer should be asking
These are the decision points where intent is either captured or lost.
Geometry and sequence
- Which feature is the base feature?
- In what order should holes, cutouts, chamfers and fillets follow?
- Which tool should create a given feature, when several could?
- Which sketch plane — a default plane, or a face of an existing feature?
Definition and robustness
- How should the feature sketch be dimensioned — radius or diameter?
- Which relations lock the sketch shape?
- Which equations guarantee valid updates?
- Could a drastic dimension change turn the sketch inside out?
Downstream documentation
- How will this feature be dimensioned on the drawing?
- Are the dimensions that matter to manufacture the ones the model exposes?
Assembly and manufacture
- Can the components actually be assembled in reality, and in what order?
- Which constraints reflect the real physical connection?
- Which entities should the constraints reference — a shaft in a bearing should mate on the bearing face, not an arbitrary edge?
- Can the feature be produced, and if so, is it cost effective?
Do not over-engineer intent
There is a real cost to excessive intent capture. Attempting to anticipate every possible future change slows modelling considerably and can suppress creativity during conceptual work. The pragmatic approach is to lock the obvious intent from the outset and discover the rest during modelling. Judgement here comes with experience; no rule set substitutes for it.
Worked comparison
Two plans, one shape, different futures
Consider a rectangular plate with four holes that must remain equally spaced diagonally from the centre of the face, whatever the plate size becomes.
Plan A — cross-section approach
- Sketch1
- Centre rectangle from origin, dimensioned
- Construction rectangle from origin, dimensioned
- Four circles snapped to the construction corners
- Boss extrude One feature produces plate and holes together.
Behaviour: because the construction rectangle is driven from the origin, resizing the plate or respacing the holes keeps the diagonal symmetry automatically. Fast to build.
Limitation: holes cannot be suppressed, patterned or re-toleranced independently, because they are not separate features.
Plan B — features approach
- Plate Centre rectangle, mid-plane extrude.
- Hole
- Sketch on front face, single circle on a construction diagonal
- Mirror / pattern Generates the remaining three from one definition.
Behaviour: slower to build but each hole is a real feature — suppressible, configurable, individually dimensionable on the drawing, and available to a hole callout.
Limitation: more references to manage, so more places for a downstream edit to break something.
Neither plan is correct in the abstract. Plan A wins where the part is a simple detail in a large model and speed dominates. Plan B wins where the holes carry tolerance, appear in configurations, or need to be described on the drawing as manufacturing features. The judgement is the design intent.
Documentation
Making intent legible to other people
Captured intent that nobody can read is only half useful. Modern systems provide several complementary channels.
| Mechanism | Granularity | Best used for |
|---|---|---|
| Feature comments | Single feature | Short notes explaining one decision. Begin with a date and the author's name, keep them succinct, and explain the decision rather than restate the geometry. |
| Design binder / journal | Whole document | Longer rationale: alternatives considered, calculations, screenshots, supplier constraints. An embedded document that travels with the model. |
| Equations | Dimension relationships | Making a proportion explicit and enforceable rather than coincidental. |
| Design tables and configurations | Part or assembly family | Showing which parameters are the size drivers and what the permitted combinations are. |
| Dimension and feature names | Individual entity | Replacing system-generated names with meaningful ones, so equations and drawings read naturally. |
| Feature-tree folders | Group of features | Structuring a long tree into readable stages: base geometry, functional features, cosmetic features. |
Consistency beats completeness
The value of documentation collapses if only some members of a team do it. Agree a minimum standard — renamed features, named driving dimensions, a comment on any non-obvious reference — and apply it everywhere, rather than documenting a few models exhaustively and the rest not at all.
Wider view
Design intent systems
A design intent system is a tool for capturing rationale and making it readily accessible. Its scope extends past the reasons for a decision to include the justification, the alternatives considered, the trade-offs evaluated and the argument that led to the choice. Interest in such systems has grown steadily, though framework support and general engineering adoption both remain limited.
Where they are used, the reported benefits are consistent: better dependency management, better collaboration between designers, higher rates of reuse, cheaper maintenance of legacy designs, faster learning for new team members, and documentation that does not have to be reconstructed from memory years later.
- Immediate Fewer failed rebuilds, because references were chosen deliberately.
- Next revision Changes land as edits rather than rebuilds, and the drawing follows automatically.
- Next project Models are reusable because a second engineer can understand why they are shaped the way they are.
- Long term Institutional knowledge survives staff turnover, which is the real business case.
Key takeaways
- Intent explains why a design is as it is; functionality only describes what it does.
- Geometry captures intent through four channels: modelling plan, sketch relations, equations and assembly mates.
- Choose references by asking what happens if the driving dimension changes drastically.
- Capture obvious intent early and let the rest emerge; over-anticipating change is its own cost.
- Document with comments, a design journal, named features and dimensions, and tree folders — consistently, across the whole team.
- The long-term payoff of intent capture is reuse and staff-turnover resilience, not the current project.
Series
Continue the pathway
The SOLIDWORKS Design Approach series works through computer aided design as an engineering discipline, from first principles to manufacture.
KEVOS®
Precision to Vision
Engineering · Mechanical Engineering
Written by Kevin Jogin
6 min read