The current generation of building information modeling (BIM) design tools is the result of about five decades of research and development on computer tools for interactive 3D design. In this first blog post I will give a short historical overview of what happened during these past decades and how that resulted in the tools for sale now.
Since the 1960s, modeling of 3D geometry has been an important research area. It was at that time recognized that 3D representations had many potential uses, including movies, architectural and engineering design and games. Early experimenting with polyhedral forms for viewing took place in the late 1960s. This led to the first computer-graphics film: Tron (1982), which was regarded as revolutionary and continues to be celebrated as a milestone in the industry.
The initial polyhedral forms could not be modified easily. Independent research efforts therefore came up with two approaches for modeling 3D solid, volume-enclosing shapes more easily: Boundary representation (B-rep) and Constructive Solid Geometry (CSG). The first approach represents an object as a collection of connected surfaces, while the second uses Boolean operations to combine primitive shapes with each other. Manipulation of the B-rep objects was more complex and less user-friendly than CSG objects, but the first approach was preferable for direct interaction, computing mass properties, rendering, and checking spatial conflicts.
While these approaches initially competed for supremacy, it was soon acknowledged that both should be combined. Since then, all parametric modeling tools have used both approaches. The CSG object representation is used for editing functions and B-rep for non-editing uses such as visualization, measuring and clash detection.
Designers then began to recognize the need to associate materials and other properties with the 3D building objects. Conceptually, that became problematic as combinations of objects with different associated materials did not have a clear, intuitive meaning. While Subtractions are intuitive (for example: windows in a wall), Intersections and Unions of shapes with different materials are not. This problem was solved by embedding ‘features’ into a primary shape, such as connections in precast pieces and holes in walls. Features are placed relative to the object they belong to and can be modified independently at a later stage.
These early modeling systems were functionally powerful, but still very expensive, unable to generate drawings and reports and they overwhelmed the available computing power. Despite this, the manufacturing and aerospace industries saw huge potential benefits in these systems, such as error reduction and a movement to factory automation. The architectural industry followed later and instead adopted drawing editors such as AutoCAD, Microstation and MiniCAD. Essentially, the industry tried to redo the same operations that they did before with paper-based systems – but then in a more efficient way.
The last major step involves the parameters that define one shape to be linked to the parameters of another shape. Floor planes, wall and ceiling surfaces together define the boundaries of a wall, for example. When a designer attempts to move this wall, the other related objects need to be updated as well. First, designers could reevaluate and rebuild new shapes on demand, but later a resolver function was developed to analyze changes and to choose the most efficient order to update them.
Such automatic updates are currently the state-of-the-art in BIM design tools. These tools represent objects by parameters and rules that determine the geometry as well as some other building properties. The BIM design tools for sale now all have different predefined sets of object classes and association behaviors. Research and development efforts will continue to push the potential of these tools in the near future.