IN DEPTH — Developing Digital Visual Literacy of Complex Digital Objects: Deconstructing CAD
- saavmsblog
- 3 days ago
- 8 min read
This is the first of a three-part series by Aliza Leventhal and Julia Larson on enhancing archivists’ visual and technical literacy in handling complex digital objects.
Deconstructing CAD: Layers, Xrefs, and You.
Physical materials in a design archive are inherently visual and present a defined point of view that is well documented and defined for future reference and research. Born-digital design records require a more complex merging of visual and technical literacy, best described as digital visual literacy, to navigate, identify, and understand the compounding and nuanced information present in individual files predominantly made with proprietary software programs. A design archive from a creator working since the mid-1980s can contain a plethora of born-digital file types ranging from mainstream office software and services (Word docs, Excel spreadsheets, PDFs, JPGs, TIFs, email), and more specialized drafting and design files (e.g., DWG, INDD, RVT). The latter, being complex digital objects, are more problematic to steward within an archive due to the extensive amounts of data they can hold and varying intended interpretations of their renderable outputs. Complex digital objects consist of multiple components: copies of graphic elements, textual and numeric data, files from external sources, visual elements with descriptors, and symbols representing a plethora of physical manifestations. All these pieces can be combined, manipulated, copied, resized, and restructured to create multiple drawing and data files.
In this series of articles, we will deconstruct complex CAD files into their constituent parts and explore the different elements and components that make up these proprietary files. Much of the literature surrounding CAD files is written from a practitioner perspective by and for architects, designers, and technologists who create files using these platforms, and often don’t account for the basic structure of a CAD drawing. The elements covered in this series are more prevalently used and can contain valuable data about the visual representation. Much of this information becomes tacit to creators who build up their digital visual literacies to almost muscle memory, as several vendors have normalized the software interface. This interface is not always intuitive for archivists to explore the rendered output and determine whether there is additional information hidden or externally linked. There is some correlation between some of the visual elements in born-digital files and analog drawings, but the vocabulary and idiosyncrasies of the software programs do not always match traditional paper processes.
Our discussion will focus mainly on components of CAD files created in the program AutoCAD, the most dominant software program (with its proprietary file format DWG), which defined an extensive specific vocabulary for components and properties that has been widely adopted. For archivists working with legacy data and older files, the second most popular software program, MicroStation by Bentley (and their proprietary file format DGN), will also be discussed to provide additional context and visual examples. The images used to illustrate these programs are all two-dimensional files from the late 1990s up to the mid-2010s, however, the programs used to open and view the files are current versions of AutoCAD and MicroStation. This series seeks to lay a foundation of shared understanding of components of two-dimensional CAD files before exploring three-dimensional design models.
Layers
Layers are the foundation of a CAD file; designers create drawings using layers as the building blocks of the organizational structure for the various facets of a drawing. For example, a base layer might contain lines delineating the walls of a building, another layer shows door or window symbols, and additional layers contain information pertaining to electrical or mechanical fixtures. Layers are structured and named based on established guidelines, both within individual firms and as a community through the American Institute of Architects and the National Institute of Building Sciences. The consistency of naming conventions is essential for collaboration and sharing of files across disciplines, as well as internal file structure for discoverability and reuse of extant templates. Layers are analogous to sheets of transparencies used in analog drafting, where a base drawing would have other sheets physically overlaid to form a complete plan; the see-through nature of the drafting paper created a complex drawing from multiple simple sheets. There is no limit to the number of layers a CAD file can contain, but the larger the quantity - which often means a higher level of granularity or detail in the drawing - may weigh the file down and require additional time to fully load/render. In Image A, the example shows 233 separate layers, including multiple roof and wall layers, distinguished by different line colors.
The user can decide which layers are visible or editable by adjusting a layer’s settings to on/off, frozen/unfrozen, locked/unlocked. When plotting to print a CAD file, various layers can be selected to print different versions of the file. See Image A, which shows the list of layer names, a lightbulb symbol (on/off), a snowflake (frozen/unfrozen), lock symbols, as well as line colors, line weights, line types, and whether or not the layer can be printed (click here for a full list of AutoCAD’s layer property settings). It can be important to note layers that are off/frozen/locked because those layers might be hidden and the information contained within those layers may be valuable to researchers.
Archivists should make note of the indicators of a layer’s properties as well as how layers are labeled to identify which were intended for specific views or purposes. In addition to layers reflecting data produced in the software, layers can also link to externally referenced (x-ref or xref) files, which contain information such as font styles, colors, textures, other drawings, etc. Archivists will need to familiarize themselves with how layers present in common software programs, and request additional information from donors or search project documentation to understand how specific architects/firms utilized naming conventions for layers and other procedural documentation.
Adding to the complexity, CAD software programs each maintain proprietary vocabularies for different parts of a CAD file. For example, MicroStation refers to 'layers' as 'levels,' with a somewhat similar interface and slightly different structure. See Image B for a screenshot of the MicroStation Levels Manager, which shows all of the files with their associated levels, and specific color, line weights, line types, etc. (click here for a full list and definition of MicroStation’s level property settings). Pre-defined levels can be copied from a shared DGN Library and inserted into a new drawing to facilitate efficient use of extant elements and designs.
Viewing older files on current software can/will create error messages if links have been lost or moved. For example, in Image B, opening the file INTELEV.dgn triggers a warning that the file is a V7 format design file and gives the option of upgrading to V8 or viewing the file in read-only mode as a V7. As concerning as a warning message may be, it is advisable to select the "view in read-only mode" option, when possible, to determine if any information can be rendered from an older file type. Additionally, the interface will provide a visual warning, such as in Image B, where the file names listed in red are external reference files that were not transferred with the base file, and their levels cannot be accessed. The ability of current programs to render legacy files as intended is not always consistent due to these types of issues. The question of 'good enough' versus 'fully rendered complete file' is a determination that will need to be made based on operational capacity, researcher interest, and archivist comfortability.
Xrefs
External Reference files, commonly referred to as x-refs or xrefs, are separate files linked to or inserted into a born-digital drawing without increasing the drawing file’s size. Xrefs can be a variety of file types that contain basic or complex information and contribute to a CAD file visually, parametrically, or contextually. Xrefs are used to add a range of content from templated information (such as title block) or a visually rendered texture of construction materials, to complete drawings (such as mechanical or electrical plans) that are linked to the born-digital drawing file with the goal of bringing various information together from a range of design and engineering disciplines into a single file.
By linking xrefs to a central file, teams of designers and/or engineers can simultaneously work on a variety of aspects of a project and remain aware of adjustments to the design or structural options through the hub CAD file. This capacity to facilitate collaboration is a tenuous strength of xrefs, as the external links can be easily broken or lost if the file structure is changed or a file is renamed. The external connection is further complicated by the way the software vendors structure the relationships between files, which presumably reflects the ways creators desire to connect various pieces of technical and visual information together. For instance, AutoDesk allows for xrefs to be either overlaid or attached, while Bentley MicroStation offers linking or embedding xref files. For Bentley MicroStation, the linked object that can be viewed from its container application, such as a DGN file, but it actually exists in its source application; or, the xref file can be embedded, where it actually becomes part of and can be dynamically changed in the container application. For AutoDesk’s AutoCAD, the attached xref is akin to the MicroStation embedded xref that brings all components into the central CAD file, while the overlay option only pulls the top-level information from a potentially more complex xref.
An outcome of the way xrefs can be treated by MicroStation or AutoCAD, a common vocabulary of linking (e.g., overlay or link) or binding (e.g., attached or embed) to a central CAD file through a relative or absolute connection The former uses the host drawing’s location as a starting point, which provides more flexibility for moving project files, while the absolute path is a complete and fixed path that starts with the drive that will result in broken links if the files are moved. Most CAD software will present a pop-up error message or a color delineation (see Image C for AutoCAD and Image D for MicroStation) if an xref file cannot be located or accessed. Designers may also bind or package the central CAD file with its xrefs to keep all relative information together. This most often occurs when the file is being shared externally or going to a printer. Significant amounts of data or drawing components can be lost if xref files are not linked, bound, or packaged to their central file.
Archivists need to be aware of how xrefs are used by a creator, whether xrefs are part of internal workflows or broader collaboration, and the amount of information shared through xrefs. As part of the initial project file overview, be sure to inquire about folder structure, the type of information xref files were used for, the storage location of xref files, and how central files were bound or packaged.
Conclusion
CAD files are particularly complex digital objects that are not only visually dynamic but also host numerous additional components that host additional complex information that is described in a nuanced and jargon-riddled vocabulary. Their compounding complexity has proven a significant challenge for archivists navigating hundreds or thousands of these files within a single archival collection. The goal of this series is to break down these component parts into accessible pieces to empower archival professionals to explore their collections. This article addressed two heavily utilized facets; the next article will discuss blocks, libraries, and attributes as commonly used resources leveraged by creators to increase efficiency and consistency in their drawings.
We recognize this is an incredibly challenging record type to work with and welcome questions, feedback, and suggestions for components you'd like to see explored in future articles!
References
[1] United States National CAD Standard V5: https://www.nationalcadstandard.org/ncs5/pdfs/ncs5_clg_lnf.pdf
[2] AutoCAD by AutoDesk Standards User Guide: https://help.autodesk.com/view/ACD/2024/ENU/?guid=GUID-D64F8076-4978-44B7-B056-D921C77FEA88
[3] MicroStation by Bentley User Guide: https://docs.bentley.com/LiveContent/web/MicroStation%20Help-v22/en/GUID-63AE1540-29CE-BCB4-41CC-B722814267F5.html
Aliza Leventhal, Library Director, Fleet Library - Rhode Island School of Design
Julia D Larson, Collections and Digital Archivist, Environmental Design Archives - University of California, Berkeley
Comments