Interacting with Design Alternatives

Many designers tend to work with multiple, simultaneously-available design alternatives. The evidence for this fact can be found in many different domains of design, such as architectural, product, industrial and mechanical design, but also in drawing, painting, sculpting, and fine arts in general. Numerous empirical studies, likewise, report instances of designers developing, and then simultaneously working with multiple design solutions in parallel. The need for alternatives-enabled work can be further confirmed by instances of expert opinion, as well as explained by theoretical accounts based on first perceptual principles, such as those governing the human visual system, or human perception in general. Yet, despite the available evidence confirming the importance of multi-state work, most contemporary computational tools, including computer-aided design (CAD) tools, their important sub-variant (parametric CAD, or pCAD tools), as well as other types of computational tools, are steeply immersed in the single-state paradigm, whereby the user can work with, and modify, just a single computational model at a time. Consequently, since the overwhelming majority of users nowadays still employ single-state tools, they necessarily have to recur to various ad hoc methods when engaging in multi-state work, in order to circumvent the limitations inherent to single-state tools. However, such workaround methods are inefficient, and error-prone. The aim of this work is to address the aforementioned shortcoming of single-state computational tools, while focusing on pCAD tools. The thesis of this work, accordingly, is that ``alternatives-enabled (i.e. multi-state) pCAD tools, designed in concordance with the interaction design guidelines and principles presented in this dissertation, support designing''. I break down my thesis statement further into the following three main research questions, each addressing one separate aspect of the statement: (1) What are the essential system features of novel multi-state pCAD tools that can support design work? (2) What are the design guidelines and principles for building and evaluating multi-state pCAD tools? and (3) What methodological approach can be followed that can help system designers build multi-state pCAD tools? In order to support my thesis, I first conceptualize (or design, develop) the class of multi-state pCAD tools, by first employing both (1) various pre-design methods (such as literature review, and probing of the ``design space'' of feasible interaction designs), as well as (2) by employing various research-through-design methods characterized by the development of numerous models, interactive research prototypes, and design artifacts in general. All these activities provide evidence about the effectiveness of specific, individual features of multi-state pCAD tools. The two main research methods that I use include (a) an expert study whereby six experts voice their opinion on the CAMBRIA 1 prototype and its features, and (b) a cognitive analysis which provides scores of the CAMBRIA 2 prototype evaluated along a number of ``cognitive dimensions''. The experience that I acquire by performing pre-design and research-through-design activities allows me to synthesize high-level bodies of knowledge such as design guidelines and principles, as well as to describe the method that I followed in order to conceptualize multi-state pCAD tools. These, in turn, answer postulated research questions, and thus support the thesis.