Much Vexation About Representation
Jerry Laiserin

Whether designing roads, refineries or residences, many engineers and architects still struggle with the three axes of representation:
> from paper to digital
> from 2D drawings to 3D models
> from geometry to building information
The trick is understanding that these axes are independent. For example, keeping 3D models and building information in digital form from end-to-end throughout the project lifecycle doesn't mean that humans must forever cease checking geometry in 2D paper drawings. Put another way, recognizing the continuing benefits of paper drawings is no reason to reject the emerging advantages of digital information.

Way back in IssueFour's LaiseBoyPhilosophy (Part 3—"D-ness" Envy), we presented and analyzed two of these axes—from to 2D to 3D and from paper to digital—which yield the consultant's classic 2X2 matrix with the quadrants labeled "Two-D Paper Drawings (A), Two-D Digital Drawings (B), Three-D Digital Models (C), and Three-D Physical Models (D)" (reading clockwise from the back left quadrant—with the actual constructed asset at the center/origin or 0,0 point). One of our conclusions in that earlier analysis was that direct representation of the constructed asset in a 3D digital model, while offering the greatest benefit in terms of streamlining overall processes, was not yet achievable in practice—other than a few isolated examples, such as the "3D+D" engineering work of CH2M (covered in IssueFourteen's LLemma), the DESTINI project of design-builders the Beck Group, or the architectural work of Foster and Partners.

In the real world, most projects that involve 3D digital models still require "extracting" or "reporting" 2D digital "drawing files" and then plotting those to 2D paper drawings for quotidian uses such as construction permit applications, plan review, bidding and contracts, and portable reference by tradesfolk in the field. However, the continued existence and necessity of 2D paper-based viewing/reviewing media does not eradicate the benefits of 3D digital models any more than does viewing/reviewing 3D on a 2D computer monitor display. There are many other valid reasons to continue using paper as a view/review medium during the course of an otherwise end-to-end digital process—check plots and archived progress point approval sets are just two examples that come to mind. Still, checking project progress via a paper plot should be seen as a kind of window or "paper viewport" into the digital process rather than a disruption of it.

Thus, there is no real conflict between keeping project geometry in 3D digital form from start to finish versus viewing/reviewing the geometry in 2D paper representations whenever convenient. Similarly, if we add the third axis of presentation—from geometry to data—we find that there is no real conflict between keeping project and component attributes linked to geometry in a digital building information model from start to finish versus viewing/reviewing that information in paper-based or screen-based non-graphical reports, schedules, estimates and the like.

This third axis of representation—from geometry to building information—has been with us at least as long as the axis of 2D-3D geometry. Ancient builders may have used quaint units of measure, such as cubits, but the notion of descriptive building component attributes and properties to supplement and complement geometric description dates back thousands of years. What seems novel and perhaps uncomfortable to us today about building information modeling is the explicit linkage of component attributes and properties to the geometric descriptions of and relationships among those components in a computerized representation.

Think of these three axes of representation as defining a box, or to indulge in jargon, a "solution space." In the far left, rear, bottom corner is 2D, paper, and geometry. Moving "forward" brings us towards 3D; moving to the "right" brings us towards digital; moving "up" leads from geometry towards building information. Before computers, we had 2D and 3D geometry and we had non-geometric building information, but all the coordination among and linkages between these elements had to be managed inside the minds of the architects and engineers who designed each project, and then put to paper and transmitted to those who would build and operate the facility (and who, in turn, would have to construct their own mental models derived from the paper representations). The first waves of computerization extended the edges of the box or solution space: CAD as computer-aided drafting gave us 2D digital drawings, while computerized specification writing, cost estimating, energy analysis and the like gave us textual and tabular representations of non-geometric attributes that could be manipulated digitally and viewed on-screen or on paper.

Computer modeling of 3D geometry "turned the corner," so to speak, by giving us representations that could support visualization, animation and the like (all appearance—albeit in 3D—but no data). Dianne Davis, of AEC Infosystems, Inc. in Baltimore, makes an extremely useful distinction "Between 'visual' models and 'computable' models—computable models being those that work for simulation, information generation, and not just visualization." A dozen or so years ago Dianne was one of the first to figure out the "front loaded process" of making better design decisions earlier through digital modeling that combined 3D geometry and attribute data, and she has since been instrumental in creating "Probably more government building models in ArchiCAD than any other software at this time—including major undertakings for US Coast Guard (USCG), US Army Corps of Engineers, and several other Federal agencies."

So we move from that left, rear, bottom corner of 2D geometry on paper to the front, right, top corner of 3D geometry and building information in digital form—not in one direct diagonal thrust across the solution space, but in incremental step-wise movements along first one axis then another. Paper doesn't go away, any more than mental processes of analysis and visualization go away. But the new modes of representation allow us humans to off-load to computers some of the tedious mental housekeeping of coordinating dimensionality and data. In effect, this is what the "aided" in "computer-aided" was always supposed to be: freeing human creativity from representational drudgery.

< back