Compared with frame-based systems, description logics have the advantage of well-defined semantics and powerful inferences. In order to exploit these advantages in technical domains, the ability to use concrete domains is needed, e.g. systems of (in)equalities over (non)linear polynomials to handle physical laws. Existing systems can only cope with comparisons between attributes. We present an approach that considerably improves the expressiveness of the concrete domains. CTL (Configurable (or Constraint-based) Terminological Logic) is based on the ideas presented in [Baader, F. & Hanschke, P. 1991] and [Hanschke, P. 1993]. Concrete domains are realised through a well-defined interface to external algorithms. Constraint Logic Programming (CLP) systems allow us to easily realise a whole range of concrete domains, e.g. over sets of symbols and numbers. In particular, we are able to handle systems of arbitrary linear polynomials. They also enable us to automatically participate in recent and future improvements in the areas of CLP and computer algebra, e.g. systems capable of handling arbitrary non-linear polynomials.