Design rationale representation and testing
Communication is one of the most important aspects in any project. This is especially true in the case of projects in the Architecture-Engineering-Construction industry. The traditional forms of communication such as drawings and specifications are not adequate in achieving the required levels of understanding among project personnel. The concept underlying the project, i.e. the collection of arguments and decisions and the thought underlying them are not communicated. This may lead to various project personnel having differing views about the requirements of the project.
This research focuses on an attempt to communicate the thought process underlying the project to the various project personnel so that they have an unambiguous and clear understanding of the essentials of the project and to test the success of this attempt using a field example. The thought process or concept underlying a project is defined as the design rationale for the project.
There have been many approaches to the communication of the concept that is the basis of a project. After a literature survey of the various approaches, the design rationale approach, which attempt to explicitly communicate the rationale was chosen. Considering the various definitions and models of design rationale, a model based on the Lee and Lai representation was chosen as the basis for this work.
Making certain adaptations to the Lee and Lai representation based on the requirements of the industry, a representation structure consisting of rationale networks was developed. Based on the rationale representation, DRARS, an object oriented rationale representation tool was developed. The rationale was represented in DRARS using object networks. DRARS was then tested on an actual project involving the renovation of a school building. Based on the collection and representation of the rationale, DRARS was evaluated for its abilities to easily author and represent design rationale. DRARS performed well in achieving both the objectives of ease of authoring and the richness of representation.
DRARS, however stores rationale that is not computer processable and presents the rationale for human interpretation. DRARS also uses multimedia to support the richness of the representation.