The design of Cyber-Physical systems (CPS) demands to combine discrete models of piecesof software (cyber) components with continuous models of physical components. Such hetero-geneous systems rely on numerous domains with competencies and expertise that go far beyondtraditional software engineering: systems engineering. In this paper, we explore a model-basedapproach to systems engineering that advocates the composition of several heterogeneous arti-facts (called views) into a sound and consistent system model. Rather than trying to build theuniversal language able to capture all aspects of systems, we rather propose to bring togethersmall subsets of languages to focus on specific analysis capabilities while keeping a global con-sistency of all these small pieces of languages. We take as an example, an industrial processbased on Capella, which provides (among others) a large support for functional analysis fromthe requirements to the deployment of components. Even though, Capella is already quite ex-pressive, it does not provide a direct support for schedulability analysis. However, AADL is anlanguage also dedicated to system analysis. It focuses on schedulability analysis, but that doesnot provide direct support for functional analysis. Rather than trying to extend either Capellaor AADL into always more expressive languages to add the missing features we rather extract apertinent subset of both languages to build a view adequate for conducting schedulability anal-ysis of Capella functional models. Our language is generic enough to extract pertinent subsetsof languages and combine them to build views for different experts. It also maintains a globalconsistency between the different views.