Abstract

“All models are wrong but some are useful”. To set the context of this paper, we may introduce “modeling” as a methodological approach or an activity seeking to provide a usable, effective, efficient and satisfactory representation of real world objects for end-users wanting to leverage those models to analyze or make critical decisions. And today, as Digitalization and Model-Based Processes are rapidly growing in the Systems and Product engineering domains, the location, the number, the heterogeneity as well as the abstraction level of those models are exploding. Thus, not only we may question their “rightness and validity" domain, but we face as well the challenges of traceability, interoperability and efficiency of model-based collaboration between disciplines in a concurrent and integrated engineering approach. PLM-centric platforms are offering a unified and interconnected framework and engineering DMU-oriented modules to support digital continuity and model traceability. However, third party or in-house applications, processes and models authored outside those platforms can still be difficult – but necessary - to handle and the transition to a definitive “single source of truth” will not be achieved in the near future without important efforts. By bridging the gap between applications, disciplines and models with an agnostic perspective, KARREN software provides a very straightforward answer with out-of-the-box capabilities and features to leverage heterogeneous processes and models content in the context of a collaborative and concurrent process dealing with fine-grained artefacts like parameters or general attributes: the ID-card of the system or product. This presentation will show how KARREN can supplement a PLM platform such as the 3D EXPERIENCE to share key product information and metadata from and to other application in order to embrace the global complexity and heterogeneity of the engineering ecosytem landscape. The principles of an agnostic and fine-grained data management system supplementing PLM-oriented software will then be introduced and we will insist on why and how models could be “more useful” and reliable as soon as we put more control on the data that directly feed them.