In an architectural restoration, there are three ways to create formwork for cast stone facsimiles:

1. Cast an existing artifact in rubber, and pour cast stone into the rubber negative to create multiple copies. This is often the best choice for artifacts which contain complex geometries that would be difficult to reproduce in any other way.
   This method requires that an artifact be provided, and that the artifact be in reasonably good condition, although minor repairs to the artifact can be made using clay or bondo.
   This is a technology that has been in use for centuries. The Victoria and Albert Museum in London, for example, is filled with copies of iconic architectural wonders from around the globe that were deemed irreplaceable and saved for posterity using processes very similar to the ones we are still using today.

2. 3D scan an existing artifact, and machine negatives from the resulting CAD file. This option can be extremely useful, as scans can be taken directly from the structure or from artifacts removed from the structure.
   3D scanning can serve an important role in historic restoration projects, but it can be very time consuming and is often not necessarily the most efficient course of action.

3. Model an artifact from scratch in computer space, using measurements taken from photographs and/or an existing artifact. A negative can then be machined from the CAD model.
   This method is limited by the capabilities of the CAD modeler and the machining hardware.

With simple geometry like these balusters, quick physical measurements of an artifact, removed from the building and shipped to our facility by the contractor, are more than enough to produce replicas. Even an adequate photograph and some accurate overall dimensions would be enough in some cases to produce a convincing copy.

Since the contractors were dismantling the existing structure to replace it anyway, a physical copy of this baluster was provided. We took measurements, created a CAD file, and confirmed the profile by wirecutting a negative template and comparing it to the artifact.

It’s actually quite astonishing the percentage of everyday objects that are faster to model in computer space than they are to scan or make physical impressions of. But of course, works of art are works of art precisely because they are not everyday objects, and the Sophian Plaza contained a number of works of art that I would prefer not to model in computer space.

The technology and the talent exist to model objects like these in computer space, but, at least given current manufacturing capabilities, doing so would be far less cost effective than a number of other processes.