A model, to fulfil a given purpose, is constructed by choosing a
representation paradigm in accordance to its fitness for that purpose.
Some properties of the model are intended to be consistent with the
requirements and these are used to keep the model accurate and
relevant, especially when the model evolves through modification. This
implies that some properties are preserved when models are combined or
modified. Explicitly or implicitly, such properties are imposed on a
model when first constructed and are expected to be consistent through
out its lifecycle. Therefore, identifying and validating them is an
important issue in requirements engineering and in modelling.This talk
aims to introduce an investigation of this issue from the viewpoint of
a logical analysis, to provide a formal description and some automation
to assist in the validation of this type of requirement. Three features
of this problem are examined. On the theoretical side, we describe a
way of checking consistency of a class of properties. This involves the
identification of information about each property, the proof of a
consistency property and the possible assumptions required for its
establishment. On the application side, we give an automatic way of
performing a limited form of validation of consistency when within a
modelling lifecycle. This involves identifying the bridging mechanism
and the necessary information provided manually as the premise for the
validation of the derived model. On the practical engineering side, we
show how this can be embedded into a component based synthesis process.