Computational support for sketching and image sorting during the early phase of architectural design | | Posted on:2003-02-02 | Degree:Ph.D | Type:Dissertation | | University:University of California, Berkeley | Candidate:Elliott, Ame Marie | Full Text:PDF | | GTID:1462390011984116 | Subject:Architecture | | Abstract/Summary: | PDF Full Text Request | | This dissertation reports on a study of architects' work practices during the early phase of design, and describes the prototyping and evaluation of interfaces to support these practices. Using ethnographic studies of how architects work, this research reveals two tasks that are central to the early phase of design: sketching and image browsing. Two related projects examine ways of supporting these tasks. The first project investigates the affordances of display size and pen-input device on a digital desk for sketching and image sorting. The second project, the Flamenco Image Browser, is a prototype image search interface that uses textual metadata to support browsing a collection of architectural images.; Architects do not use computers in the early phase of design because pens, paper, and books are more successful at meeting their sketching and image browsing needs. This claim is supported by an analysis of architects' layering and collage making practices during the early phase of design. These results are also used to inform two separate computer tools, which are then evaluated in terms of their ability to support sketching and image browsing.; The first tool, a large digital desk display with a stylus input device, was compared with a small tablet display with a stylus input device. The evaluation indicates that (1) a general purpose stylus is not appropriate for specialized design tasks, (2) the relatively low resolution of the desk's large rear-projected display is not a problem for image browsing, (3) parallax errors limit the desk's usefulness, and (4) ergonomic aspects of the desk and input device make it an undesirable combination for image sorting.; The second tool, the Flamenco Image Browser, is a prototype interface to a collection of 36,000 architectural images. The interface uses textual metadata about the images to do two things: (1) generate query previews to give users hints about where to go next, and (2) help users expand or refine their queries to avoid the problem of having too few or too many search results. An evaluation with professional architects using the Flamenco Image Browser indicates that textual metadata describing a visual collection is helpful for exploratory image browsing tasks, and an interface making greater use of text is desirable.; Despite the limitations of the tools evaluated, the enthusiasm of the architects evaluating the tools indicates potential for computational support to be helpful during the early phase of design. The dissertation concludes with recommendations for how devices and image search interfaces could be designed to support architects' work practices. | | Keywords/Search Tags: | Early phase, Image, Support, Architects, Practices, Work, Architectural, Device | PDF Full Text Request | Related items |
| |
|