Many languages for software architectures have been proposed, each dealing with different stakeholder concerns, operating at different levels of abstraction and with different degrees of formality. It is known that a universal architectural language cannot exist since the various concerns, requirements, and domains may change. Moreover, stakeholder concerns and needs are various and ever evolving even while designing a single system. Model-driven techniques may be used to answer the need for supporting the creation of extensible, customizable and stakeholder-oriented architectural languages (i.e., next generation architectural languages). Part of this approach is developed in a framework called BYADL. In this paper I present the big picture behind the approach, the research aspects considered in order to get BYADL closer to an ideal architectural framework and future research issues. Categories and Subject Descriptors D.2.11 [Software Engineering]: Software Architectures—Domainspeci...