Over the top drawings

I am creating a Building Information Model (BIM) of a nearby K-8 public school for a capital improvement project report. I will use the model to analyze rooms, life safety features, and other facilities issues. I am basing my model on drawings from 1937, 1950 and 2000. This modeling effort has in turn become a case study in how drawings have changed over the past century. Problems with the 2000 set are partly due to changes in the construction industry, but are also due to the bumpy roll-out of CAD production in architecture offices during the 1990s. BIM is susceptible to the same problems, but also offers an opportunity for significant quality improvements going forward.

A particularly annoying feature of the 2000 set is the paper size: 36 inches by 48 inches. This excessively large paper size is the reason for this post.

Briefly, the 1937 set is 9 sheets, sized 21 inches by 36 inches. By today's conventions, 9 sheets is absurdly small for a 10,000-square foot building. The set seems to describe the majority of the building, but not everything. There are, for instance, very few details, of anything. There are no flashing or waterproofing details. There are no roof details. There are some building sections, but nowhere near what we would expect today. Then, drawings were a shorthand for the building. Much was inferred. It was inferred that tradesmen knew their trades.

A nice feature of the 1937 drawings is the use of what I refer to as an elevation set: an elevation, plan and section of a particular area of the building, laid out next to each other so that one could see the area in elevation, plan and section at the same time. Elevation sets are a useful way to describe exterior wall assemblies. Much of the built environment is constructed orthogonally. These orthogonal constructions can be efficiently described using an elevation set, and designers from a century ago knew this.

Back in 1937, architects were OK with leaving out an exterior elevation if that drawing would really not add value to the set. If the west and east elevations of a building were mirror images of each other, only one needs to get drawn. A note indicated the west was "typical" to the east.

Interestingly, the 1937 set includes a count of brick courses on the outside of the building. There are 22 courses on the first floor, and 25 courses from the second floor to the parapet. I do not see this on modern sets, and I think it is because in 1937, a draftsman actually drew the 22 lines on the first floor and the 25 lines on the second floor when he created the elevations. Today, brick is drawn as an electronically generated hatch pattern.

The 2000 set is an AutoCAD set, meaning the paper is white and the markings are black, the result of a digital set being printed on a printer. A particularly annoying feature of the 2000 set is the paper size: 36 inches by 48 inches. This excessively large paper size is the reason for this post. These drawings are over the top.

The 2000 set has drawings segregated by type. Plans are in front, followed by elevations, then details. The 1937 set would place details on the perimeter of the plan sheet, thus saving paper. The 2000 set is not interested in saving paper. The plan sheets have abundant white space.

Drawing segregation by type is also codified by industry standard committees. National CAD standards committees were interested in creating drawing systems that would make it easy to find a particular type of drawing. This kept draftsmen from hiding a casework detail among the window details. Thus, plans are placed on A2.x drawings, and exterior details on A8.x drawings. Some confusion could remain. What if one wanted to show the relationship between a window and a cabinet? The answer, more often than not, was that those details stopped being drawn.

Segregation of drawing types is also the product of digital technology. CAD programs favor the development of typical detail sheets, and these are easy to organize by detail type. Thus, architects create efficiency in drawing by populating their template files with sheets of similar details. Drawing segregation also makes it easy for the casework person to edit the casework detail sheet, and the roof person to edit the roof detail sheet. In the case of the window and cabinet situation, segregated drawings also allowed a designer to believe he was being thorough be having cabinet details and window details, without actually thinking about the cabinet's relationship to the window.

The 36-inch by 48-inch paper size of the 2000 set is the current bane of my existence. The drawings are incredibly unwieldy. I keep thinking that on the construction site, these drawings would have been very difficult to use. Unfolded, the drawings take up 3 feet by 8 feet. My office has standardized on the 24 by 36 (D Size) sheet size, although occasionally we use 18 by 24 (C Size), and have used 30 by 42 (E Size) once or twice in the past ten years. Very large buildings are easier to depict on larger sheets. It used to be that the sheet size was based on the smallest size that could fit the entire building at 1/8th scale on a single sheet. We no longer concern ourselves with that requirement. Smaller pieces of paper are much more convenient than larger pieces of paper, and our goal is to make it easy to use our drawings.

The 2000 set could have used smaller pieces of paper. It appears the architect had at some point decided to standardize their detail sheets on 36 by 48, and so all projects were drawn on that size sheet. What this indicates is that the product coming out of this office was optimized to be produced cost effectively, with little thought for the needs of the consumer, whether it was the owner or the contractor. Those decisions were made when production technology went from hand drawing to electronic drawing. The new technology was seen as a tool for firm profitability, not as a tool for producing a better product.

The 1950 set is about as slim as the 1937 set, although it is a smaller project. The set is more fleshed out, meaning most parts of the building are documented in both elevation and plan. I get the sense that the 1950 set is interested in exploring new and innovative design details, such as a massive glass block wall on the south side. The 1937 set was not interested in innovation. It was interested in noting how this building shares in the conventions of a public school project. By 1950, innovation becomes a feature, and innovation requires documentation in order to be understood.

By 2000, the goal of construction documentation became the complete and unambiguous description of the project. There no longer were conventions and assumptions that could remain unsaid. Construction assemblies were fine-tuned for their intended goal. While an 8-inch masonry wall had fire, acoustical and impact resistance qualities all at the same time, a particular situation might only require fire resistance. There was a cost effective metal stud wall that would be perfect in this situation, since it had the fire rating but not the acoustic or impact qualities. The number of different wall types ballooned from three or four to 16 or 20.

The 2000 set has stair details. The earlier sets simply indicate where the stairs are.

The 2000 set also has redundant details, and this is a result of a peculiar aspect of electronic production. It is very easy to create a template drawing file with standard drawings already existing within the file. As architects realized they were using the same wall type details over and over again, they realized it is easier to delete a detail from a template set than it is to create a detail. It is also easier to create a new detail by copying and modifying an existing detail than it is to create a detail from scratch. Years ago, a set might be missing a wall type. Today, a set might contain two identical wall types, with different type marks.

Pemetic school addition original drawings
Where are we now? We are nearing the point where BIM (Building Information Modeling) technology is mature enough for us to think about producing better design documents from the consumer's point of view. My firm is well past the point where we realize BIM increases profitability. I want our consumers to value our products more than our competitor's products. Part of BIM's allure, and benefit to consumers, is in the benefit of well-modeled output. Windows and cabinets both get modeled. Better yet, their relationships get modeled. The model offers the ability to show not only the building components but how they relate to one another. The best product I offer right now is our digital model. Unfortunately, owners and contractors do not know how to access our models. This is not simply because they do not understand how to use BIM technology. There is a hardware problem. Contractors on construction sites are hard pressed to maintain digital tools amidst the dust and debris of a typical construction site. It is much easier to walk around with a piece of paper than it is to walk around with a flat screen, even if that screen is an iPad. The very nature of construction sites would need to change in order to make effective use of digital technology. Those contractors and construction managers who realize this may be able to benefit from rich digital media in ways their competitors cannot. We find ourselves in the position to provide rich digital media, and are eager to find contractor partners who know how to benefit from them.
Follow Mike Sealander, Maine Licensed Architect:


Principal at Sealander Architects, Ellsworth Maine. Revit guru. Married with 3 children. Avid gardener. Lived in San Francisco for nine years. Master in Architecture from Columbia University Bachelor of arts in religious studies, Wesleyan University. Graduated Staples High School, Westport CT. Hope to spend some time in Hokkaido before all is said and done.