djec.org

CAD Layer Guidelines

Commentary on Revisions to the AIA CAD Layer Guidelines

This commentary is a reply to the invitation for "reviewer response" that accompanied the initial publication of the AIA (American Institute of Architects) CAD Layer Guidelines, 2nd edition, Draft 6/11/96. This draft was subsequently removed from the AIA web site, and is made available here for reference purposes.

Commentary on the AIA CAD Layer Guidelines, 2nd edition

The concept of expanding the organization of the AIA CAD Layer Guidelines to include File Names is of great interest. Not only is this timely in terms of the publication of the proposed ISO standard for CAD layer names, but the ISO 9000 quality management standard as well. This scope change potentially broadens the AIA standard to encompass document management, establishing a uniform structure for the naming of drawings and other documents that comprise a project, as well as the layers within individual drawings.

The observations contained herein are made strictly on a basis of the WWW page Draft dated June 11, 1996. No assumptions have been made as to the completeness (or incompleteness) of the proposed new guidelines as represented by the draft itself.

The single most significant difference between the new Draft and the 1990 Guidelines (in terms of layer- naming) appears to be the elimination of the distinction between building and drawing information layers. As defined in the AIA's "blue book" (p. 15), "Building information generally represents the physical form of the site, building, or objects in the building...." while "Drawing information adds annotation, dimensions, [etc.]...." This is a fundamental and important distinction.

Within the drawing, the differentiation between sets of building and drawing layers exactly parallels the distinction between model Files and sheet Files in the Draft. The new guideline considers this important enough to state that "using numerical descriptors for sheet file names and alphabetical descriptors for model file names allows a person to distinguish between the two types of files simply by inspecting their file names."

Terminology

The term model is more abstract than building, and is preferable since its scope includes physical site elements as well as furniture and other non-building objects. Sheet and drawing are relatively interchangeable, as they both relate to annotations describing the model/building (the architecture) and the physical medium on which all the information is ultimately disseminated (the plot).

The draft asserts that "in some cases models and sheets coexist in the same CAD file [because] for many small projects, it is not worth the effort to separate a model and sheet into two separate CAD files." This is entirely correct. In fact, if this logic were carried to its theoretical extreme, a single drawing file could contain the model of an entire project, with all the sheets/drawings/plots obtained by using different layer setups and views.

The committee makes a database analogy in which "the models are the data and the sheets are the reports of the data." More precisely, the "models" (plural) are like individual database tables representing two-dimensional views (i.e., plans, elevations, sections) of the data. Queries into the database are made in order to work on the model (singular) from various points-of-view, and to set up different sheets to plot. The database is comprised of a hierarchy of files referencing other files (blocks, reference files). By setting the visibility status and other characteristics of the layers within the individual files, "reports" are produced.

There needs to be a means of classifying drawing information within each CAD file that is separate and distinct from the model information. The mechanism of using an Annotation field appended to model layers does not offer sufficient flexibility, as the drawing information must be tied to a specific major or major/minor building element. Dimensions, for example, do not just relate to floor plans, but to the type of floor plan: small-scale, large-scale, furniture, outlet, etc., as described in the following case study.

Case Study

The construction drawings for a school were prepared with each of the overall base plans in its own (model) file. A separate (drawing) file was created for plotting the 1/8-inch scale plans for each floor, from which the base plan drawing was referenced twice, once for the floor plan and once for the reflected ceiling plan. Both plans, including separate sets of dimensions, were positioned respectively on the upper and lower halves of the plotting sheet. The layer setup of the plotting sheet controlled the visibility status of each instance of the same base drawing.

In addition, several areas of each plan (such as kindergarten, primary, and special education classrooms, administration, faculty center, and mechanical support areas, etc.) were also referenced at 1/4-inch scale onto individual plotting sheets (drawings), which also contained the interior elevations (models) for each type of space. Again, the dimensions, notes, room tags were separated using the drawing layers for large-scale floor plans. Thus a hierarchy of sheet files was created for each floor, all referencing the same model.

The new Draft Guidelines very adequately treat the various plan types in its File Naming Conventions. The same amount of flexibility needs to be maintained in the Layer names. Significantly, a close comparison of the file name categories suggested for the new system with the drawing layer names from the 1990 guidelines reveals that the classifications are virtually identical (see spreadsheet [tables referenced above]).

Construction Type

In the new draft's brief discussion of the ISO format, mention is made of certain optional fields, such as phase, view, and scale. Construction type (existing, new work, and demolition) should be added to the list of potential qualifiers which may be included in the suggested Annotation field. (The 1990 Guidelines provided a User- Defined field for this purpose.)

It is important that a classification such as construction type be given a unique identifier, not necessarily its own field but at least its own column within a field. This is so the class may be used as a basis for selecting elements of the model (or drawing), in software.

It is useful to be able to identify a group of building (model) elements (walls, doors, windows, etc.) by construction type. This can facilitate the actual editing of a drawing file by enabling the selection of a class of objects identified as existing, then changing them all to demolition while in the process of adding new construction in the same physical area.

If Interior Design is to be given its own "Agent Responsible" code, one might suggest that there should also be a fourth construction category for base building. Interiors professionals often maintain libraries of "core drawings" which are used repetitively for many projects and seldom, if ever, change. Editing capability is enhanced if a project, containing existing, new, and demolition elements, can be separated from its base building context and moved about.

Comparing the 1990 and Draft Guidelines

The [tables above] rigorously compare the old and new layer guidelines. [Their] overall organization is based upon that of the 1990 AIA "blue book" as follows:

• The first section (ll. 1-152.) corresponds to Chapter Three: Summary Layer List without Modifiers.
• The second section (ll. 153-519.) corresponds to the Building Information Layers portion of Chapter Four: Master Layer List with Modifiers.
• The third section (ll. 520-646.) corresponds to the Drawing Information Layers portion of Chapter Four: Master Layer List with Modifiers.

The column headings use the Agent-Major-Minor terminology of the new Guidelines, and both long and short layer names are included. The right-hand column is titled Status, and indicates new layers added or old ones deleted. In a few instances, question marks in the status column indicate that it was unclear from the web page whether the layer was deleted or unintentionally omitted.

Lower-case letters indicate where a (partial) layer name has been added. The most significant recommended change is the use of the NULL Minor field where a Major category does not break down. This would seem to be preferable to losing column alignment in layer listings on the one hand, or having strings of hyphens as filler on the other. The word null was chosen because it is considered in SQL to be a non-value, and because it conveniently has four characters. (Possibly NU might be used for the short form.)

Ellipses (or partial ellipses) are otherwise used to indicate missing parts of the short format names, where they conjoin with other partial names:

• Electrical Building Information Layers (ll. 86-104. with ll. 420-429.)
• All Drawing Information Layers (ll. 520-630 with ll. 631-646.)

The "Designations applying to all disciplines," which the draft spells out under File Naming Conventions, have been added as a separate category to the Drawing Information Layers (ll. 619-629.) with the exception of those relating to construction type. Old partial names are shown as deleted where they are redundant.

One minor point: If a GRID designation is added under floor (where it usually is drawn, l. 188.) the need for a Major Category (l. 2.) is eliminated.

It should be noted that the ELEV, SECT, and DETL Major Categories have been eliminated, except for the Architect Agent.

The Short and Long Formats

Finally, in response to the question raised at the end of the web page, it is unnecessary to eliminate the long form in order to conform to the proposed ISO standard. As long as 100% compatible definitions of both long and short layer names are defined, it is a trivial matter to perform two-way translation through software.

Within an office or a project team, the long form could thus be used for understandability, and converted to the short form for distribution. To the extent that layers might be selectively consolidated in the translation process, a certain amount of security could be gained by limiting the ease with which CAD files may be modified after distribution.

Jeffrey E. Clark, July 28, 1996

Top of Page