What do information reuse and automated processing require in engineering design? Semantic process

Abstract

Purpose: The purpose of this study is to characterize, analyze, and demonstrate machine-understandable semantic process for validating, integrating, and processing technical design information. This establishes both a vision and tools for information reuse and semi-automatic processing in engineering design projects, including virtual machine laboratory applications with generated components. Design/methodology/approach: The process model has been developed iteratively in terms of action research, constrained by the existing technical design practices and assumptions (design documents, expert feedback), available technologies (pre-studies and experiments with scripting and pipeline tools), benchmarking with other process models and methods (notably the RUP and DITA), and formal requirements (computability and the critical information paths for the generated applications). In practice, the work includes both quantitative and qualitative components. Findings: Technical design processes may be greatly enhanced in terms of semantic process thinking, by enriching design information, and automating information validation and transformation tasks. Contemporary design information, however, is mainly intended for human consumption, and needs to be explicitly enriched with the currently missing data and interfaces. In practice, this may require acknowledging the role of technical information or knowledge engineer, to lead the development of the semantic design information process in a design organization. There is also a trade-off between machine-readability and system complexity that needs to be studied further, both empirically and in theory. Research limitations/implications: The conceptualization of the semantic process is essentially an abstraction based on the idea of progressive design. While this effectively allows implementing semantic processes with, e.g., pipeline technologies, the abstraction is valid only when technical design is organized into reasonably distinct tasks. Practical implications: Our work points out a best practice for technical information management in progressive design that can be applied on different levels. Social implications: Current design processes may be somewhat impaired by legacy practices that do not promote information reuse and collaboration beyond conventional task domains. Our work provides a reference model to analyze and develop design activities as formalized work-flows. This work should lead into improved industry design process models and novel CAD/CAM/PDM applications, thereby strengthening industry design processes. Originality/value: While extensively studied, semantic modeling in technical design has been largely dominated by the idea of capturing design artifacts without a clear rationale why this is done and what level of detail should be favored in models. In the semantic process presented in this article, the utility and the chief quality criteria of semantic models (of technical information and artifacts) are explicitly established by the semantic processing pipeline(s). This constructively explains the significance of semantic models as communication and information requirement interfaces, with concrete use cases.