Appendix A. Using Computer-Aided Design Programs for Urban and Architectural Reconstructions: The Case of Early Islamic Jerusalem

Al-Asad, Mohammad

Save Cite Share

Free

Appendix A. Using Computer-Aided Design Programs for Urban and Architectural Reconstructions: The Case of Early Islamic Jerusalem

Mohammad Al-Asad

The Shape of the Holy: Early Islamic Jerusalem

This appendix addresses the subject of architectural representation. More specifically, it deals with the use of computer-aided design (CAD) programs for representing historical works of architecture and urbanism, taking early Islamic Jerusalem and a number of its important buildings as examples.The participants...

Author

Mohammad Al-Asad

PublisherPrinceton University Press

Stable URL: https://aaeportal.com/?id=-100045

Appendix A. Using Computer-Aided Design Programs for Urban and Architectural Reconstructions: The Case of Early Islamic Jerusalem

MOHAMMAD AL-ASAD

This appendix addresses the subject of architectural representation. More specifically, it deals with the use of computer-aided design (CAD) programs for representing historical works of architecture and urbanism, taking early Islamic Jerusalem and a number of its important buildings as examples.1The participants in this project would like to thank Kirk Alexander, Manager, Interactive Computer Graphics Laboratory (ICGL) at Princeton University, and ICGL personnel, Carlo Balestri and Kevin Perry, for the invaluable help they offered us concerning the use of computer-aided design technologies for the reconstruction of early Islamic Jerusalem. We are also indebted to Harrison Eiteljorg II, Director, Center for the Study of Architecture, for the valuable advice he gave us during different phases of the project. Parts of this essay previously appeared in Mohammad al-Asad, “Computer-Aided Design Programs and the Study of Architectural History: The Case of Early Islamic Jerusalem,” Newsletter of the Center for the Study of Architecture 6 (August 1993): pp. 5–10.

Authors on architecture, whether writing as historians, critics, or theoreticians of architecture, usually need to supplement their writings with images. These images consist of two-dimensional drawings like plans, longitudinal sections, and elevations or three-dimensional drawings such as perspectives and axonometrics. Other media which represent a work of architecture include photographs and three-dimensional models. Through these various media one can represent works of architecture that exist in their entirety or survive only partially as well as buildings which were completely destroyed or were never executed, but for which information exists to allow for reconstructions.

These methods of representation make reference to external physical realities, but differ in the degree to which they interpret these realities. They can attempt to mimic an external physical reality, as with a photograph, or to abstract it, as with a plan drawing. A photograph of a completely or partially surviving work of architecture provides a record of how that work appeared at a certain point in time defined by various elements such as specific lighting conditions. A 50 mm lens approximates the angle of vision of the human eye; a wide angle or zoom lens registers different angles of vision. Regardless of how faithfully a photograph may attempt to mimic the manner in which the human eye perceives an object, photography not only involves an interpretation, but also a degree of abstraction of that object since it transforms three-dimensional information by transferring it to a two-dimensional plane.

A variation on the photograph for representing a work of architecture is the motion picture, which adds the element of movement. It has not, however, achieved the ubiquity of the photograph for the purpose of architectural representation, in part because it is more expensive to produce and more cumbersome to exhibit.

Paintings and drawings, whether free-hand or drafted, provide versatile methods for representing works of architecture. Certain drafted two- and three-dimensional drawings, such as plans, sections, elevations, and axonometrics, do not aim at reproducing an external physical reality as much as providing information about it. Such information can include the dimensions of a work of architecture, the relationship between its different components, and between the work and its surroundings. Although these methods do not always attempt to register the manner in which the object appears to the human eye, they at least maintain the integrity of its dimensions, a characteristic which the effects of perspective sacrifice.

Since a perspective drawing relies on the diminution in size of objects at a distance, it does not preserve the actual dimensions of a work of architecture. However, it approximates the manner in which a work of architecture appears to the human eye. Perspective drawings can vary in the degree to which they imitate an external physical reality. They therefore differ in their representation of features such as texture, light, shade, and intricate details. They can include the sketchy and schematic, which rely on evocation rather than mimesis, and the highly detailed which aim at producing high levels of realism. Like photography, perspective drawings inherently include a degree of abstraction since they transfer three-dimensional information to a two-dimensional plane, but because of the arbitrary quality of the drawing medium, the abstraction is twice removed from the original object.

Three-dimensional physical models are only occasionally used for the purposes of architectural representation. Often, they are used by architects to explain unbuilt designs to clients. Models can provide highly realistic representations of a work of architecture. In contrast to drawings and photographs, they maintain its three-dimensional character. They usually differ from the original object in scale and in the materials used for construction. Like perspective drawings, they can range from the schematic—such as the massing model—to the highly detailed that attempts faithfully to emulate textures, colors, and intricate architectural details.

Since models are not always easy to move, they are directly accessible only to a small number of viewers. In contrast, the reproduction of drawings and photographs is relatively easy. If a model is to reach a wider audience, it usually needs to be photographed. In such a case, the three-dimensional character of the model is lost and the result is a representation of a representation, or two levels of abstraction from the original object.

CAD (computer-aided design) programs provide the latest addition to the available methods for architectural representation. These programs have the potential to influence significantly the manner in which architectural historians deal with past works of architecture. CAD programs are computer software that can be used to create digital objects in two and three dimensions. The results can be presented as wireframes (using separate lines resembling strands of wire, see fig. 3), as shaded surfaces, or as solid objects (fig. 4).2Definitions of the computer terms used in this essay are available in computer dictionaries such as Microsoft Press Computer Dictionary: The Comprehensive Standard for Business, School, Library, and Home (Redmond, Washington, 1991). For an introduction to the principles and applications of computer-aided design technologies, see Malcolm McCullough and William J. Mitchell, Digital Design Media: A Handbook for Architects and Design Professionals (New York: Van Nostrand Reinhold, 1991).

~

Description: Aqsa Mosque, Umayyad period, bird's eye view of the southwest, wire frame

Figure 3. Aqsa Mosque, Umayyad period, bird’s-eye view from the southwest, wire frame version.

~

Description: Aqsa Mosque, Umayyad period, bird's eye view from the northeast, reconstruction

Figure 4. Aqsa Mosque, Umayyad period, bird’s-eye view from the northeast, shaded version of the same drawing as in figure 3.

This method of representation combines the qualities of the traditional methods discussed above and, in some cases, offers advantages over them. It is a very powerful tool for creating two- and three-dimensional drawings such as plans, longitudinal sections, elevations, axonometrics, and perspectives. A major strength of these programs lies in their versatility. With traditional drafting methods, obtaining a new view of an object requires the construction of a new drawing. With CAD programs, one can create a single comprehensive three-dimensional geometric model of an object, and subsequently obtain an infinite number of two- and three-dimensional views of the object from that single model.

Consequently, a CAD three-dimensional geometric modeling system can provide highly comprehensive documentation of an architectural object. Since a comprehensive three-dimensional CAD geometric model should enable the computer user to view an object from an infinite number of points, the construction of an object using CAD programs requires the input of a relatively complete set of geometric information about the object. In contrast, no reasonable set of traditionally drafted two- and three-dimensional drawings can provide an equivalent number of views of an object. Traditionally drafted drawings will, at best, only illustrate parts of the object.

Although the computer screen is a two-dimensional surface, CAD programs have the ability to preserve the three-dimensional character of a given object. In order to construct a three-dimensional CAD geometric model of the object, one needs to input information defining the x, y, and z coordinates of its various entities. This allows CAD programs to combine the qualities of two- and three-dimensional drawings with those of a physical three-dimensional model.

CAD programs allow for high levels of accuracy in the representation of a work of architecture or urbanism in comparison to drafted drawings or physical three-dimensional models. Dimensions and coordinates can be specified with extreme precision. Also, since information is fed into the computer through the intermediacy of devices such as the keyboard and mouse, the relation between the human hand and the resulting computer-generated model is indirect. Consequently, human errors resulting from factors such as a trembling hand—which are not uncommon in hand-drafted drawings or hand-made physical models—are eliminated. In addition, since all views of an object constructed using a CAD program belong to a single geometric model, accidental discrepancies between the dimensions of similar elements in different views are avoided.

Changes are easy to make with CAD-generated models. Similarly, CAD programs allow for ease of replication and for the three-dimensional projection of objects. They also allow for distributing the elements of the model among layers each of which can be viewed or worked on separately. In addition to enabling the computer user to view a model through a variety of two- and three-dimensional representational drawings such as plans, sections, elevations, axonometrics, and perspectives, CAD programs allow for the simultaneous display of different views. Moreover, changes made to one view are automatically updated in the database to affect all others.

Numerous CAD programs have shading and rendering capabilities which transform line drawings into highly realistic images containing features such as textures and shadows. Photographs can also be incorporated into a CAD-generated model to create highly realistic effects. A number of CAD programs have animation capabilities which approximate the experience of walking—or flying—through or around an object.

All in all, CAD programs are flexible systems for the visual documentation of architectural objects and provide high levels of interaction between the object and the viewer. They allow the user to create comprehensive two- and three-dimensional objects in relatively little time. This last remark needs further elaboration since constructing a complete CAD model of a complex architectural object can be time consuming, and if one is only interested in a single view of that object, the construction of a complete model may be inefficient. Under such circumstances, a CAD program can use a two-dimensional Cartesian coordinate system in which only x and y (but not z) values are represented in a single two-dimensional plan. If more than one view is required, the construction of a complete three-dimensional digital model may be well worth the time.

CAD programs are more than merely advanced drafting tools. They have the potential of transforming the manner in which architectural historians address historical works of architecture and urbanism, especially for the purpose of reconstruction. On a basic level, these transformations show similarities to the changes which word processing programs have had on the act of writing. Before the advent of word processing, typing the final draft of a manuscript was separate from the act of writing it. Now, boundaries between creating and typing the final statement of a written work have become amorphous.

In a similar manner, the advent of CAD programs has meant that one can integrate the research needed for reconstructing a work of architecture with the process of representing the final results. Because of this integration, finished drawings can easily be made available at various stages of the research process. As research progresses, the model from which drawings are created can be easily modified and updated. When using traditional methods of architectural representation, initiation of the presentation phase needs to wait until the research is completed.

The changes introduced by CAD programs can function as a double edged sword. On the one hand, these programs provide the historian with additional time to carry out research since the presentation phase is no longer a separate phase which cannot be initiated until the research phase is completed. The negative consequence is an example of the rule of diminishing returns. Since changes can be made up to the last moment, one can spend far too much time making what only amount to minor changes.

CAD programs can also function as valuable tools for teaching architectural history. The interactive nature of these programs allows the user to explore thoroughly a past work of architecture or urbanism through still and moving images. These programs can be customized to allow the user who has little or no training in the use of CAD programs to view an object with relative ease. Through these programs, the student can explore a past work of architecture or urbanism with a degree of thoroughness not possible through the images of traditionally printed texts. Because of their ability to simulate internal and external views from above ground level, these programs allow the user to experience a work of architecture or urbanism in a way not readily available even to visitors to that work.

An integral part of this study of early Islamic Jerusalem has been to create a series of two- and three-dimensional drawings which explain the evolution of the city during the period under consideration. The drawings illustrate the city and a number of its major buildings at three moments in time. The first is during the second quarter of the seventh century, just before 637, when control of the city transferred from Christian to Muslim hands (fig. 11); the second belongs to the second quarter of the eighth century, during the Umayyad period (fig. 63); and the third belongs to the middle of the eleventh century, during the Fatimid period (fig. 69).

The drawings include site plans showing the city at the three points in time, plans of the city’s important buildings, and axonometric and perspective views of the city and its buildings. A number of the perspective views are taken at ground level, and are intended to explore the evolving relationships between the city’s major monuments, primarily the Dome of the Rock and the Church of the Holy Sepulchre (fig. 58).

Computer-aided design technology has proven to be ideal for this project. We used this technology to construct a digitized three-dimensional model representing the city of Jerusalem as it may have existed during the late Byzantine period (fig. 12), and modified the model through additions and deletions to represent the city at the two other points in time under consideration. From the final model we were able to obtain easily a large variety of two- and three-dimensional views (figs. 56–59). Constructing this variety of views using traditional methods would have been unrealistically time consuming. Additionally, since CAD programs allow the user to make changes and additions with little difficulty, we were able to begin constructing the model of the city as soon as the research project was initiated, and to incorporate new results as they became available.

The digitized model of the city contains a number of architectural and urban elements most of which underwent changes during the period under consideration. These include the topography of the city, its urban components (city walls and gates, streets, residential neighborhoods, and the Haram platform) and individual monuments, the most important of which are the Church of the Holy Sepulchre, the Dome of the Rock, and the Aqsa Mosque.

We began inputting data for the digitized model of the city using AutoCad release 11, and converted to release 12 after it became available. We primarily used an IBM Personal System/2 model 80 386 which processes data at 20 Mhz for this task. The machine was initially equipped with 8 MB of Random Access Memory (RAM) and later expanded to 14 MB. Its hard-disk drive has a capacity of 320 MB. In order to run AutoCad, the machine had to be equipped with an 80387 math coprocessor chip. For the purposes of presentations and animations, we relied on the much faster UNIX-based Silicon Graphics, Inc. IRIS workstations operated by Princeton University’s Interactive Computer Graphics Laboratory. Large AutoCad drawings which take over three minutes of regeneration time on the IBM 80 386 only need about thirty seconds on the Silicon Graphics workstations.

We were able to create high quality line drawings of the AutoCad drawing files using a 600 dots per inch (dpi) laser printer. For the purposes of rendering the AutoCad line drawings, we relied on the Silicon Graphics, Inc. Inventor program. Most of the computer images used in this book are reproductions of these rendered images.

The Inventor program also has animation capabilities. Consequently, we created a series of animations which simulate the experiences of flying around and walking through the city and a number of its structures during the various periods under consideration. These animation sequences were converted into a short documentary video (about ten minutes) dealing with this project.3See Jerusalem: 600–1100. Produced by Intermedia Communications for the Institute for Advanced Study, Princeton, New Jersey. (Authored by Mohammad al-Asad, Abeer Audeh, and Oleg Grabar. Produced in conjunction with Kirk Alexander, Manager, Interactive Computers Graphics Laboratory, Princeton University, 1993).

Abeer Audeh and I worked on inputting information for this project over a period of about two years, both of us working on the project on a half-time basis. Both of us were trained as architects and I had previous training in the use of the CAD programs. However, neither of us was trained as a computer programmer; we worked on AutoCad in the capacity of software users. When transferring the information to the Silicon Graphics, Inc. Inventor program, which was still in its earlier versions, we worked in collaboration with Princeton University’s Interactive Computer Graphics Laboratory personnel who had specialized technical knowledge in computer graphics programs.

The digitized model of the city consists of over 25 separate drawing files connected to a master file using external file referencing. (External referencing permits files to be combined for viewing or plotting, but also to retain their specific identities for individual editing, viewing or plotting.) Each of the files contains one of the city’s buildings or part of its urban composition such as the topography, walls, or streets. The largest of these drawing files occupies about 0.8 MB of memory. When combined, the files take up over 7.0 MB of memory, which is an extremely large amount of memory for AutoCad to process. Limiting the size of the drawing files using CAD technologies presented a major challenge for this project.

As the size of a file grows, CAD programs function at a slower speed. The maximum CAD file size which can be processed at an acceptable speed (or even processed at all) depends on the computer hardware and CAD software used. Still, the data contained in this project proved to be very large for the software and hardware technologies currently available to most users. In order to control the size of these CAD files we had to keep representations of architectural details such as surface decorative features to a minimum since such details occupy large amounts of memory. For example, a digitized three-dimensional CAD model of an Ionic capital can take up more memory than a simple building.

As a result of this restriction, we had to concentrate more on forms and spaces than on surface decorative elements or the “skin” of buildings. This approach has provided our CAD-generated objects with a degree of minimalism and abstraction. Since surface decoration plays an integral and very important role in most pre-modern architectural traditions, a study of such traditions usually needs to include representations of their decorative characteristics. When dealing with large CAD files, one way of representing decorative details is to supplement these files, which need to concentrate on forms and spaces, with traditionally produced drawings and with photographs of architectural details. These drawings and photographs can even be copied into the computer through scanner or ray tracing and incorporated into the CAD model. One can also create additional separate CAD files concentrating on drawings of architectural details. For example, we produced two versions of the Aqsa mosque (figs. 3–4) for this project. The first version is connected to the master file linking the various files of the whole city through external referencing and represents the exterior of the structure. This version occupies about 0.3 MB of memory. The other version is more detailed and consists of two files which are independent of the master file but are connected to each other through external referencing. One of these two files represents the exterior of the Aqsa mosque, the other its interior. Together, the two files occupy about 1.9 MB of memory.

Of course, one should keep in mind that new generations of more powerful computers that can process large amounts of information at higher speed are continuously becoming available. This means that the difficulties presented by large CAD file sizes will diminish in the future, and a single file or a group of connected files will be able to include considerably more architectural details.

This resulting simplification and abstraction of the CAD models of the city raises another issue related to the use of CAD technologies for the purposes of architectural and urban reconstructions. CAD programs have the capability of creating highly realistic images. In the case of architectural reconstructions such realism can be misleading since computer generated images can be sufficiently effective to convince the viewer that what is presented is a faithful representation of a preexisting physical reality, even though the reconstruction may be of a conjectural nature. In the case of the Jerusalem project, we primarily present possibilities about what might have existed, and do not claim to replicate a preexisting reality. Consequently, a degree of simplification and abstraction of the CAD-generated images is desirable since it serves to remind the viewer that these images are an approximation of what might have existed and not realistic representations of preexisting urban and architectural compositions.

As mentioned earlier, in order to construct a three-dimensional CAD geometric model of an object, one needs to provide a set of comprehensive and precise geometric information about it. We therefore had to input specific data about the city’s architectural, urban, and topographic elements, even when accurate data was not available. This necessity of inputting a comprehensive and precise set of geometric information can be avoided when using traditional drafting methods. With traditional drafting methods, one can deal with the unavailability of geometric information for a part of an object by simply not representing that part. Therefore, if the information available for the facade of a building is insufficient, one would not draw that facade.

As a result of the precision which CAD programs require for generating geometric models, these programs do not allow for some of the “fudging” that can take place using traditional drafting methods. For example, with a traditionally-produced perspective drawing, the draftsperson can easily modify dimensions or manipulate relationships between components of the object, but a three-dimensional CAD-generated geometric model uses one set of dimensions for all views, and calculates perspective drawings using specific mathematical formulas that maintain a consistent relationship between the various components of the object.

Because we had to provide a complete and precise set of geometric information for the geographic, urban, and architectural elements of the city—even when sufficient information did not exist—we needed to differentiate these elements according to the degree of certainty of our reconstruction. We considered relying on a visual system of differentiation, but since we had already used color in the CAD geometric model of the city for the purposes of chronological differentiation, we decided that adding another visual system of differentiation would only result in confusion. Consequently, we decided that the differentiation of elements according to the degree of certainty of their reconstruction would be done through the text rather than the images.

The representations of the city’s topography provide an example of the challenges faced in using CAD programs for a project of this nature (fig. 5). The city is located on a hilly, and extremely uneven terrain. Constructing a three-dimensional model of its contours and connecting the contours to the city’s architectural and urban components is in itself a challenging task. An additional challenge arises from the fluidity and continuously changing character of the topography of an urban setting which result from changes in factors such as habitation patterns and geological conditions. Our knowledge of the city’s topography during the period under consideration is scarce. We decided to deal with this lack of adequate information by beginning with what is known. We therefore used the topography of the bedrock underneath the city as the point of departure since the bedrock reflects the overall characteristics of the contour above it. Taking the topography of the bedrock as a guide, we represented the city’s topography by constructing a three-dimensional mesh on a 50 by 50 meter grid and fitted the mesh with the city’s urban and architectural features which we reconstructed for the late Byzantine period. We modified the topography to accommodate the architectural and urban developments which took place in the city during the two later periods under consideration (figs. 5–6).

~

Description: Jerusalem, basic structure grid of city plan from the south, schematic...

Figure 5. Jerusalem, basic structure grid of city plan from the south. The Haram is the partly empty space in the middle right, while the better lit section is the “living” area, reconstructed as a model.

~

Figure 6. Jerusalem in wire frame. This picture shows all the information available in the computer, from whatever period. The city is seen from the southeast with the Haram in front and the Holy Sepulchre in the back.

Had our goal been to provide an archaeological reconstruction of the city, and to register only that of which we are certain, this approach would not have been acceptable because it involves an approximation of the city’s topography. However, since our aim is to present visual impressions of how the city might have existed during the periods under consideration, this approach towards representing the topography of the city proves quite suitable.

Finally, had this study of Jerusalem’s urban and architectural development extended into the present, we would have followed a reverse chronological order for its reconstruction. This means that we would have begun by constructing a three-dimensional geometric model of the city as it exists today, and from there, we would have moved back chronologically. In other words, we would start with the known and move towards the lesser known.

These remarks have dealt only with the production of images. A final word on the display of these images. One can, of course, display CAD-produced images using a computer screen. CAD images can also be transformed into more traditional formats, such as slides and photographs, which can be viewed directly or reproduced in printed works like books, journals, or brochures. It is also becoming increasingly possible to combine digital and traditional display methods. Both still and moving images can be projected directly from a computer to a large screen, and currently available printing technologies can now directly convert electronically stored images into high quality printed ones which can be viewed separately or as part of a printed work.

This is not to say that computers should displace traditionally printed works for the display of images, but that they should be used in conjunction with them. Under present circumstances, the book or journal remains the indispensable physical object for the display of images. In most cases, print media are cheaper to obtain and easier to handle and transport than a computer. If nothing else, looking at a printed page remains more comfortable than staring at a computer screen.

1 The participants in this project would like to thank Kirk Alexander, Manager, Interactive Computer Graphics Laboratory (ICGL) at Princeton University, and ICGL personnel, Carlo Balestri and Kevin Perry, for the invaluable help they offered us concerning the use of computer-aided design technologies for the reconstruction of early Islamic Jerusalem. We are also indebted to Harrison Eiteljorg II, Director, Center for the Study of Architecture, for the valuable advice he gave us during different phases of the project. Parts of this essay previously appeared in Mohammad al-Asad, “Computer-Aided Design Programs and the Study of Architectural History: The Case of Early Islamic Jerusalem,” Newsletter of the Center for the Study of Architecture 6 (August 1993): pp. 5–10. »

2 Definitions of the computer terms used in this essay are available in computer dictionaries such as Microsoft Press Computer Dictionary: The Comprehensive Standard for Business, School, Library, and Home (Redmond, Washington, 1991). For an introduction to the principles and applications of computer-aided design technologies, see Malcolm McCullough and William J. Mitchell, Digital Design Media: A Handbook for Architects and Design Professionals (New York: Van Nostrand Reinhold, 1991). »

3 See Jerusalem: 600–1100. Produced by Intermedia Communications for the Institute for Advanced Study, Princeton, New Jersey. (Authored by Mohammad al-Asad, Abeer Audeh, and Oleg Grabar. Produced in conjunction with Kirk Alexander, Manager, Interactive Computers Graphics Laboratory, Princeton University, 1993). »

Next chapter

Base text size -

Chapter images ()