Move one window. Does the rest of the model keep up?

Ask a building-design tool to move a window by one foot. The opening slides across the wall and the new rendering looks convincing. But the elevation still shows the old position, and the quantity list has not changed.

You now have three answers to one design decision. A small edit has exposed a larger problem: the outputs are no longer describing the same building.

This guide gives you a practical test for an editable model, whether you are building the software or choosing a tool to use. We will move one window, inspect what changes, and deliberately try an edit that should be rejected. You can repeat the test in the working example.

Give the window one identity

Our window is W01, on the rear wall of the shed used in the planning guide. Its opening is 36 × 36 inches, centered on the wall, with the sill 43½ inches above finished floor.

The identity matters because a window appears in several places: the wall opening, surrounding framing, 3D view, elevation and schedule. Those views need to refer to the same W01. Creating unrelated rectangles with similar dimensions makes later edits difficult to reconcile.

A useful record names the opening, its host wall, position convention, dimensions and vertical reference. Product dimensions and the manufacturer’s required rough opening belong in separate fields when a product is selected. This example contains a modeled opening, not a selected window product.

Decide what the edit is allowed to change

Move W01 12 inches in the positive X direction—the 12 in right control in the example. X measures across this wall, from its center. The opening remains 36 inches wide, so the arithmetic is straightforward:

The opening, window placement and surrounding wall framing must move together. The drawing and inventory must then describe that result. The door, floor and roof stay as they were: this is a horizontal window edit, not a request to redesign the shed.

Writing down what should stay fixed is as valuable as naming what should move. It gives you a way to spot an unexpected change that a plausible rendering might hide.

Follow the consequence into the framing

In the original example, the compiled model contains 136 identified geometric solids. After the move it contains 135. The opening now intersects a different set of stud positions, so the resulting wall inventory changes.

That one-solid difference is specific to this model and this move. It is not a rule that moving a window saves a stud, and it is not a purchasing recommendation. It is evidence that the quantity output responds to the same accepted geometry as the wall and drawing.

Inspect the wall in the model and compare its generated rear elevation. The green selection marks the window’s king studs, jack studs, header and sill. Those members have different roles around an opening; moving the glass alone would leave the assembly behind.

A selected header size or a visibly complete frame still needs its own design basis. Here we are checking whether a software edit stays coordinated. That is a narrower question than whether the resulting assembly is suitable for construction.

Try an edit that should fail

Now select Try corner conflict. It proposes a window center at +48 inches. The window’s right edge would reach +66 inches; including this model’s king-and-jack envelope takes the affected extent to +69 inches. The allowed opening-and-framing envelope ends at +56½ inches, where this rear wall meets the side wall.

The proposal therefore fails the example’s boundary check. The accepted model remains in place.

This is a revealing test. A tool that only demonstrates successful edits has not shown what it does with a bad proposal. Look for an explanation that names the conflict, and confirm that the previous accepted design is still recoverable. Silently squeezing the window smaller or moving another wall would answer a different request.

The demonstration checks this bounded opening-and-corner condition. A production editor also needs the checks appropriate to its supported cases: other openings, wall orientations, products, services and structural arrangements. Each check needs an explicit scope rather than a general “valid” label.

Keep the reference level steady

A horizontal edit leaves W01’s sill at 43½ inches above finished floor. Its head remains at 79½ inches. The floor is zero in this model.

That sounds obvious until a drawing measures from the wall base, a product object measures from its own origin, and a floor change adds an offset elsewhere. “Height” is not enough information without the reference it is measured from.

The same issue appears in open exterior spaces. A porch needs its own boundary and floor reference. If one part of the software treats it as a room, a later update can generate a wall across an opening that was meant to stay open. In the Fieldhouse brief, the proposed porch is kept outside the enclosed shell and area total.

Check the file you actually hand over

Use Download drawing after an accepted move. The SVG contains the plan and rear elevation for the current revision. Use Download record to keep the window dimensions, position, host-wall identity and inventory beside it. In this example, the rightward move is named BL-W01-R12.

Open the exported files and compare the same few values: revision, opening center, sill and member count. A correct screen is not sufficient if the downloaded file belongs to an earlier state. Save the original as well if you want to compare the change later; page edits reset when you reload.

For a project that uses IFC, agree the receiving team’s exchange requirements and test the exported file in the tools they use. IFC represents building information, including objects and relationships, but the information a particular export preserves still needs checking. The browser example here exports JSON and SVG, not IFC. buildingSMART’s introduction to IFC.

Repeat this small test before trusting a large model

You do not need a complex house to learn something useful about a design tool. Start with one wall and one opening:

1. Save the original dimensions and revision.
2. Make one accepted move and compare the model, drawing and inventory.
3. Confirm the dimensions that should remain fixed.
4. Try one known conflict and recover the accepted state.
5. Open the exported files and check their revision and values.

If those answers agree, you have evidence of a coordinated edit within the tested scope. As the project grows, extend the same habit to roof junctions, porches, products and schedules. The next guide explains how to keep each review claim tied to the evidence that supports it.