Abstract
Computers actually support, almost automatically, routine tasks such as those related to the optimization in design. Besides, the scientific community shows a growing interest in developing computer systems to aid non-routine tasks as a key to enhance individuals' creativity and innovation potential. In such a context, several attempts have been made to create tools based on the TRIZ logic to support inventive problem solving; some of them have been commercialized since decades, but still there is no established paradigm and all of them suffer from several limitations. So far the analysis of those limitations has been focused on the structure and on the nominal features of the software tools, while no in-depth and systematic investigation has been made to identify the reasons behind the partial failure of the existing systems. This paper proposes a set of general criteria to perform the evaluation of computerized tools supporting inventive design and reports an exemplary application, through protocol analysis, to the dialogue-based computerized algorithm for problem analysis, published by the authors in the past. © Springer-Verlag London 2013.