Resource type
Thesis type
(Thesis) Ph.D.
Date created
2010
Authors/Contributors
Author: Botta, David
Abstract
The increasing availability of web based collaboration tools fuels design conversation between heterogeneous stakeholders across organizational boundaries, underscoring the need for new designers to get to the heart of conversations that might include huge numbers of entries. The goal of this work is to show that linkography is a viable candidate to help make that kind of discovery possible. A linkograph links design moves with prior moves, resulting in a model of the design episode. The research methods were mixed, though primarily qualitative. The primary data comprised records of a series of eleven two to three hour design meetings over a six month duration, with five participants. A model for predicting the location of topic shifts was developed on the first two exploratory meetings, and tested on the remaining nine design meetings. The model used a finer-than-topic-shift granularity linkograph of the nine meetings to predict topic shifts. It combined a measure of both backward and forward links, plus a threshold, in order to segment the design discourse on topic shifts. An additional threshold comprising a number of segments was used to filter transitive links to retrieve contextualizing information from the discourse. The test included quantitative comparison of model segmentation with human segmentation, and qualitative evaluation of relevant contextual information drawn (using the model, the segmentation, and the linkograph) from previous design conversations. The results suggest that employment of linkography is a viable and pragmatic addition to design rationale.
Document
Copyright statement
Copyright is held by the author.
Scholarly level
Language
English
Member of collection
Download file | Size |
---|---|
etd5832.pdf | 4.01 MB |