Grant: Discovery Grants Program - Individual

The objective of our research is to introduce a set of tools and techniques for building and configuring highly-customizable software. Software that is highly-customizable allows users with limited or no technical skills and knowledge to easily configure it so that it perfectly fits their unique characteristics and needs. The current software development practice underestimates the need for intuitive and accessible customization facilities, mainly due to the absence of systematic methods for constructing them. This is evident in several application domains, such as common desktop applications (word processors or e-mail clients) and their "bloated" configuration screens or packaged information systems and the overwhelming costs that organizations pay for customizing them to fit their individual business needs. To address these issues, we propose a problem-oriented approach to software customization. We suggest that, instead of referring to the technicalities of the desired configuration, users can use specially designed languages for describing the characteristics of the problem at hand, such as their particular goals and preferences and the state of the environment they are in. Then, software tools can translate the description of the problem into the actual software configuration that best solves it. This would allow common users to obtain the customization that best fits their needs without any reference to low-level technical details. Our investigation involves studying language designs that are appropriate for this purpose, exploring methods for designing translators of problem descriptions into software configurations of existing systems, as well as understanding the process of designing software that is amenable to such customization methods and whose configurability aspect correctly and completely reflects the variability of the problem. Such a problem-oriented approach would both make common personal software more accessible to less technically skilled users, and, in the case of organizations, dramatically reduce the cost of customizing and administering their complex software infrastructures.