Most buildings do not have a model. They have a drawing set — a PDF issued at handover, a stack of DXFs from a fit-out, sometimes a photograph of a paper plan taken on a phone in a site office because that was the only copy anybody could find. All three go in through the same door.
Tracing turns those sheets into geometry. Walls come off the drawing and stand up in three dimensions, openings are cut where the drawing says they are, storeys stack in the right order, and what comes back is a real IFC twin rather than a flat picture of one draped over a plane.
The obvious question is whether to believe any of it, and the answer is not a reassurance. It is a check. Every trace ends by setting the modelled area against the figure printed on your own sheet — not against an internal expectation, but against the number the original drafter wrote down.
On the published run that reads 1,396.9 square feet modelled against 1,393.4 printed: 0.25% over, comfortably inside the 1% band the geometry has to agree within, with 16 of 16 rooms enclosed and nothing flagged on the run that produced the verdict.
The band is the point of the whole exercise. A number reported without a tolerance is a claim; a number reported inside a published tolerance is a measurement, and only one of those is worth pricing work against. A check that could not be taken — a sheet with no printed dimension, a room that would not close — costs the run its clean verdict rather than passing quietly and being discovered later.
Rooms are counted rather than assumed, which sounds pedantic until it matters. The count comes back from the model itself, so an unenclosed room shows up as a warning at trace time instead of as a missing area in a takeoff three weeks into pricing.
Where the trace is wrong it is correctable in the scene rather than by starting over. Walls move, openings shift, a mis-read hatch is fixed, and the model re-checks itself against the same printed figure — so a correction improves the verdict instead of invalidating it.
For an owner this is the difference between having drawings and having a building you can measure. Quantities come off the geometry, and a contractor pricing the work is pricing the same object you are looking at rather than their own interpretation of a PDF.
It is also, by a distance, the cheapest way in. No CAD seat, no consultant engagement, no waiting for a measured survey to be scheduled — the sheets already sitting in a folder are enough to produce something openable this week.
The quality of what comes out does depend on what goes in, and it is worth being honest about that. A clean, dimensioned architectural set produces a better twin than a scanned photocopy of a photocopy, and the verdict at the end will tell you which one you had.
That transparency is the feature. You do not have to trust the trace; you have to read the number it reports against your own sheet, and decide whether the building it produced is good enough for the decision you are about to make with it.