The goal of research in programming languages should be to develop languages that integrates the best of concepts and constructs from the various programming paradigms. We do not argue for a multi-paradigm language, where the programmer alternates between the different paradigms/styles. Instead, we find that the languages of the future should integrate the best available concepts and constructs in such a way that the programmer does not think of multiple paradigms when using a given language. In this paper, we describe to what extent the BETA language has been successful in obtaining a unified style and where more research is needed. In addition to traditional paradigms such as object-oriented-, imperative- functional- and logic programming, we also discuss concurrent programming and prototype-based programming. We discuss language features such as the BETA pattern construct, virtual procedures and classes, higher order classes, methods and functions, part objects, blockstructure, and ...