The design development of the provides an understanding of the limitations of the metaphor of physical architecture in the virtual realm. The discussions which took place about the hypotheses, both in face-to-face meetings and on-line, have been recorded on a World Wide Web site (see Appendix B). The conclusions that been reached from all of these discussions are broadly categorized into the topics of orientation, navigation, transition, enclosure, and scale.
The ability of a participant to orient oneself in architecture, whether physical or virtual, is critical to the success of the design. The project designed as part of this study indicates that orientation can be established by using a clear parti, landmarks in a grid, and the enhancement of direction and depth perception, especially when physical constraints are removed.
Architectural parti, the organization of and relationship between elements and information in space, is vital to the understanding of that space by the participant. Partis may be linear, radial, clustered, or fall into many other categories of geometric arrangement. In all cases, however, it is used to provide the participant with a broad understanding of how the architecture is organized.
By developing a design with enough variety of forms and spaces, yet keeping them organized clearly, landmarks in the grid are established. The "grid" is the conceptual and regular placement of forms and spaces in a rhythm or regular pattern. "Landmarks" are the variations in the pattern which serve as unique elements or points of reference by which to orient oneself in architecture (Lynch). They are critical to orientation in both the physical and the virtual realms.
The ability to establish a participantís depth perception and direction in space and is also valuable for orientation (Henry). These orienting devices are accomplished in both the physical and the virtual realms by use of color and texture, lighting, and atmospheric and linear perspectives. Lighting as an orienting device is harder to establish in the virtual realm because of the computational expense it requires from the simulation technology (Nimeroff). Solid colors require less computational overhead than textures in virtual environments as well. The visibility of atmosphere, like mist, can establish some depth cueing in the physical world, but too much of it, like fog or smoke, can disorient participants. The same is true of the virtual realm, although heavier mist or fog can actually aid oneís orientation in the virtual world by exaggerating depth cues and eliminating distant, distracting elements.
A participantís understanding of oneís sense of direction is just as important in the virtual realm as it is in the physical (Satalich). However, without the context and constraints of horizons and gravity, there is no inherent sense of up and down. The project in this study indicates that although each space or room has proportions which imply an up and down, these "local ups" did not need to correspond to one another from room to room. Thus, without a "global up," oneís sense of direction and orientation is with the body in relationship to an immediate space, even though the orientation of other adjacent spaces may differ.
The success of physical and virtual architecture depends not only on how easily one can orient oneself in the design, but also how well one can move, or navigate, through the environment (Bednar). Participants navigate using cognitive maps, varying degrees of freedom, and by a combination of environmental cues and tools.
Successful navigation depends on oneís ability to develop a cognitive map (Lang). For a participant to establish a cognitive map in an environment, whether physical or virtual, that person must be able to move and explore freely in the environment. Software functionality like pre-defined paths of movement and pre-defined viewpoints of key spaces diminish a participantís ability to construct an accurate cognitive map of the space. These tools, however, prove to be efficient ways to navigate through a world after a participant has already built a cognitive map of oneís surrounding environment.
In the physical world, one typically navigates through oneís environment in three degrees of freedom: movement along the earthís surface as a plane (X and Y), and rotation by turning oneís body around (yaw around the Z axis). In the virtual realm, there can be up to six degrees of freedom to navigate through Cartesian space: movement along three axes (X, Y, and Z), and rotation around these axes (roll, pitch, and yaw). Movement in all six degrees of freedom evokes the sensation of oneself as a free-floating body in space. With six degrees of freedom, and without a "global up" or gravity, the architecture need not be expressive of an inherent up or down for one to navigate.
Whether limiting movement to three degrees of freedom with a "global up" in the physical world, or by allowing six degrees of freedom with only "local ups" to orient participants virtually, the architecture needs to offer several other cues as aids to navigate in space. The physical world offers signage, symbols, and text, in addition to architectural elements (Beck), all of which are often culturally specific aids to navigation (Rapoport). There are also tools such as compasses and maps by which one navigates in the physical environment.
In the on-line virtual realm, the architecture is accessible to participants of all cultures and languages, so text is a limitation. Although the on-line culture has its own set of symbols and iconography, there are few well-established ones for navigation in three-dimensional space at this time. While software tools like compasses and maps are also available in the virtual realm, these are not as intuitive or as accessible as using the environment itself to guide in navigation. A viable solution, then, is to use abstract architectural elements and environmental cues which transcend cultural interpretations to aid in the navigation of virtual architecture.
The metaphor of physical architecture can be used for orientation and navigation through virtual architecture, with minor limitations. This metaphor also applies to transitions in the virtual realm, but with stricter limitations. Issues related to transition that are considered include the establishment of a context and acknowledgment and expression of hyperlinks.
In both the physical and the virtual realm, architecture is best understood as approached in a context. Physically, a sense of arrival is established in geographic, Cartesian space as participants must approach a building from the outside in order to enter it. Thus, a context is often perceived and understood from the exterior. In the virtual realm, however, there is no inherent starting point from the exterior. Still, it is helpful for the participant to view the design from an "exterior" or "godís eye" omniscient viewpoint before moving into the virtual architecture. This heightens the sense of arrival, aiding in the establishment of place, and also helps to orient the participant. Thus, a sense of approach is still achieved in the lack of an exterior context.
In the physical world, participants move through the time and space relative to one another under the laws of physics. Distance and time are integral and inseparable, as one moves through space over a period of time. In the virtual world, however, hyperlink technology does not require that movement through space and time continuously. Links allow immediate access to other data, and participants (conceptually) move through space without moving through time.
However, not all links actually allow "immediate" access to sets of data, and not all data carries equal weight in relevance to the environment from which one accesses it. Thus, links which move the participant between sets of spatial data expressed as virtual worlds can be categorized into different categories, and these can yield different architectural expressions. There are two metrics by which the character of the link is determined: whether or not the data on each side of the link is geometrically adjacent, or contiguous, and how closely the data in each virtual world is related. The combinations of these factors yield four possible kinds of links, three of which are relevant.
One combination of these metrics yields a kind of hyperlink which connects spatially adjacent rooms of a specific design of virtual architecture, and in which there is a strong informational relationship between the worlds. These links, developed as "level of detail" switches, can be given the transparent expression of a door if the computer running the simulation has pre-loaded the adjacent rooms into memory. If this is not the case, these links require architectural expression as airlocks or vestibular spaces between rooms to allow the participant to move through the space freely, even if the room one is moving into is not fully loaded and rendered.
A second category of hyperlinks is that which connects two spaces (or rooms) which are not spatially adjacent but do contain related information. These links deserve special architectural articulation as something other than a doorway into an adjacent room, like a texture-mapped "window." The metaphor of moving through space is appropriate to the time it takes for this hyperlinked world to download and be rendered.
A third kind of link is that in which the data sets on either side of it are geographically adjacent, but describe unrelated data. Mathematically, the possibility of such a link exists, but the probability that the unrelated worlds have links to each other in the same Cartesian space is infinitesimally small, so this is considered a null category.
Finally, there are links to places which are neither spatially adjacent nor of directly related information. These links, too, require articulation as elements in the design which are not ordinary doors. Again, the metaphor of a spatial transition is appropriate while the world is downloading. However, this transition to unrelated data, and the links expressing it, should be different from links leading to related data in order to inform the participant of that fact.
Whether developed as physical architecture or virtual architecture, designs use enclosure as a device to separate and define spaces and rooms. Site limitations, separation of function, the creation of interior space, and the movement between solid and voids are all factors which influence the creation of enclosure.
In the physical realm, designs inherently exist on a physical site. That site has geographic boundaries as defined in a social, cultural, or political context. The limits of a site or property serve as strong indicators for the presence of surfaces of enclosure. Virtual architecture, on the other hand, inhabits unlimited and infinite expandable space. Without limits to the site, there is little indicator of where to place surfaces of enclosure.
In both the physical and the virtual realms, surfaces of enclosure exist around spaces to signify separation of function. In the physical world, this is also done for issues of privacy, either visual or auditory. Issues of privacy do not exist as such in the virtual realm, as the representation of other participants and the representation of oneself to other participants is often under oneís own control, aided by software functionality.
Functions are also enclosed in the physical world to define interior spaces which shelter these functions from unfavorable climatic conditions. There are no such conditions for virtual architecture, and therefore surfaces of enclosure are not justified by climate. Rather, virtual spaces are given surfaces of enclosure to reduce the amount of information a participant is exposed to at any one time. While information overload does present a risk to participants in the physical realm as well, it is more likely in the information-rich virtual realm (Wurman).
In the physical world, movement between spaces takes place in the voids which penetrate solid surfaces of enclosure. To walk from one room to another, one passes through a doorway (void), between the door jambs, above the threshold or sill, and below the head (all solids). Thus, solid elements are not passed through, but between. In the virtual realm, collision detection with hyperlinks are often necessary to achieve movement between spaces. This means that a participant must pass through a polygon (conceptually, a solid) to achieve a transition to the next space. At the same time, passage through polygons rendered as solid walls simply puts the participant on the other side of those polygons without activating a link to another world. Conceptually, the rendering of surfaces of enclosure is inverted, in terms of solids and voids, from their function in the virtual realm.
Issues of scale apply to architecture of both the physical and the virtual realms, although in different ways. Scale issues were studied in terms of "human" scale, the detail of oneís environment, and the development of connections.
The achievement of "human" scale is necessary in both the realms of physical and virtual architecture, although they are accomplished through different means. In the physical world, scale is established by the size of elements relative to the participant. Virtually, the participant has no inherent size, so scale is only indicated through oneís velocity. That is, the scale of oneís environment is indicated by the rate at which one moves through it. Whether scale is achieved by size (physical) or movement (virtual), however, both are affected by the size of openings. A participant gains a sense of scale when passing through a relatively small opening whether by virtue of oneís tight fit or by having to reduce oneís velocity to ensure passage through the opening.
The detail of oneís environment also influences oneís perception of scale. The richness or articulation of surfaces in the physical world can affect oneís sense of scale in much the same way that the size of elements does. Similarly, the detail and number of polygons rendered within a participantís field of view affects oneís perception of scale virtually. By introducing a range of complexity of detail to the project in this study, it has been found that a participant tends to move through a complexly detailed environment more slowly than a simpler one.
The development and expression of connections can also affect oneís sense of scale in architectural spaces. In the physical world, the detailing of a connection is accomplished as a hinge, a reveal, or any number of other ways to join elements; often, this joint introduces a smaller element at the connection. The presence of these smaller connecting elements introduces a richer scale to the physical environment. In the virtual realm, connections as understood in traditional terms are not necessary. Without physical forces of nature, virtual elements simply exist adjacent to one another as abstract geometry without the need for connecting elements, or the expression thereof. Connections, then, do not serve to enrich the scale of virtual architecture.