B. Paech et al. (Eds.): REFSQ 2008, LNCS 5025, pp. 123 – 128, 2008. © Springer-Verlag Berlin Heidelberg 2008 Can We Beat the Complexity of Very Large-Scale Requirements Engineering? Björn Regnell 1,2 , Richard Berntsson Svensson 2 , and Krzysztof Wnuk 2 1 Sony Ericsson, Lund, Sweden http://www.sonyericsson.com 2 Lund University, Sweden bjorn.regnell@cs.lth.se, http://www.cs.lth.se Abstract. Competitive development of complex embedded systems such as mobile phones requires management of massive amounts of complex require- ments. This paper defines and discusses orders of magnitudes in RE and impli- cations of the highest order of magnitude that we have experienced in industrial settings. Based on experiences from the mobile phone domain we propose re- search areas that, if addressed successfully, may help beating the complexity of Very Large-Scale Requirements Engineering. 1 Introduction The complexity and size of software-intensive systems continues to increase, which in turn gives increasingly large and complex sets of requirements. How many require- ments can an industrial system development organisation manage with available Re- quirements Engineering (RE) processes, methodology, techniques and tools? This is hard to know as RE research often falls short in characterizing the scalability of pro- posed methods. How large and complex sets of requirements do we need to consider when researching new RE technology? We have no complete picture of current indus- trial practice in terms of complexity of sets of requirements, but we have experiences from industrial cases with enormous complexity where current RE technology have useful but partial effect [4,5,6]. Our objective with this paper is to share some impor- tant research opportunities that we have found in our observation of what we call Very Large-Scale Requirements Engineering (VLSRE). The paper is organized as follows. Section 2 proposes a definition of VLSRE based on the size of a requirement set as a proxy for its complexity. Section 3 provides a case description of the mobile phone domain that illustrates an instance of VLSRE. Section 4 highlights some research opportunities that we through our own industrial experience have found relevant to VLSRE. Section 5 concludes the paper. 2 Orders of Magnitude in Requirements Engineering Table 1 defines four orders of magnitude in RE based on the size of the set of re- quirements that are managed by an organisation that develops software-intensive