Many Cyber-Physical Systems (CPSs) nowadays are highly adaptive systems that consist of many variable components. Experts from different domains (e.g., mechanical, electrical/electronic, and software engineering) use heterogeneous modelling languages to describe the system, leading to overlapping system information and inconsistencies between different development artifacts when evolved over time. Product Line (PL) development techniques can be used to model variability, and view-based modelling helps in managing heterogeneity and making consistency relations explicit. Managing variability and heterogeneity at the same time leads, however, to an increased potential for inconsistencies, which can be attributed to the co-evolution of problem- and solution-space, as well as the co-evolution of variants and the PL. In this paper, we present an initial approach and a vision for extending the view-based Vitruvius approach to achieve consistent-variability modelling. We describe how variability-aware View Types and Consistency Preservation Rules (CPRs) address inconsistencies between the Variability Model (VM) and solution space artefacts of different domains. For that, we are extending NeoJoin, the declarative view definition language developed for view-based model-driven software development, to support transformation-time configuration and the Reactions language with variability information to enable product- and platform-oriented development.