اتخاذ درسهای کلیدی از تولید رایانه برای توسعه چابک محصول نرم افزار - مطالعه تطبیقی

Many industrial new product development (NPD) software projects apply nowadays agile methodologies. These methodologies, such as Scrum, eXtreme Programming (XP), and Feature-Driven Development (FDD) date back to 1990s, and the Agile Manifesto was declared in 2001. However, already before that the concept of agile manufacturing (AM) was discovered to describe a corporate ability for quick adaptation to changing requirements. There is surprising amount in common between these two fields. This raises a question of whether NPD software development companies could take even more overall advantage of those different agile approaches. This interdisciplinary paper explores the commonalities between the key concepts of AM and some of the most popular agile software methods, and consequently suggests potential new areas for software process improvement (SPI) in large-scale NPD organizations. An industrial case example illustrates how agility in embedded software product development can be enhanced by following typical NPD principles. We conclude that there is potential for further improvements in software product development industry in general by seeing agility as a wider, organization-oriented business concept following the AM/NPD learning. Current agile software process models cover only a subset of this space.

Many new product development (NPD) companies operate nowadays in uncertain and dynamic competitive environments. There are many sources of turbulence, stemming from such factors as intensified global competition, reduction in lead-time and life expectancy of products, diversification of demand, and new technologies (Ismail et al., 2006). The companies must nevertheless be able to compete in sustainable ways. In the early 1990s, the concept of agile manufacturing (AM) was devised to address those considerations (Goldman et al., 1995; Preiss, 2005). The key is to cope with irregular and unpredictable demand, unlike in traditional mass production. Software production faces currently many similar problems and challenges. In software product development, agile principles were addressed mostly independently starting from the late 1990s. The Agile Manifesto was declared in 2001 (Agile Manifesto, 2001). Many lightweight software development process models and methodologies emerged (e.g., XP), in particular in response to the problems with the traditional software development models (Waterfall) faced in the new competitive environments. Many of the AM principles are by nature general-purpose and technology-independent. Although software product development and software production have certain unique properties and profound differences compared to typical manufacturing operations, it is by and large not obvious how exactly the two fields are related, and how much cross-domain knowledge could be shared. That line of reasoning leads to the following specific questions to study in this paper: 1. How do key AM concepts and current agile software development models compare? 2. What new manufacturing concepts could potentially be adopted to software production? There is no one straightforward way to tackle such broad questions. This is further complicated by the fact that many of the concepts and terms lack rigorous and generally agreed definitions. This paper therefore approaches the questions by composing an extensive comparison table with multiple different viewpoints. Such an approach produces a composite answer showing that there are indeed many essential similarities, but also that the AM discipline addresses many larger-scale business and organizational areas much more extensively than the current agile software development models. The rest of the paper is organized as follows. Section 2 explores the background and related work of AM research and agile software development. Section 3 develops the comparison matrix, and Section 4 analyzes it with inferences and insights (Question 1). Section 5 further illustrates the comparison points with an industrial product development case example. This is followed by concluding discussion and implications in Section 6 (Question 2). Finally, Section 7 summarizes the work with pointers to further research.

This comparative investigation has illuminated that current agile software development models address only a subset of this space, and a wider holistic view is thus necessary to achieve sustainable agility of the entire company. Many AM principles and concepts can be applied for this. In AM, the concept of agility has evolved first from mass-production to FMS and then to wider, business-oriented organizational approaches (Christopher, 2000; Narasimhan et al., 2006). In current agile software development models, there is clearly room for similar evolution. The key is to realize that agility is really a NPD company-wide strategic system. Having stated that, this paper leaves room for further research: • Developing a NPD software development agility improvement framework following the principles of manufacturing agility frameworks (Gunasekaran, 1998; Ismail et al., 2006; Sharifi and Zhang, 1999; Vázquez-Bustelo and Avella, 2006). We have tentatively addressed this elsewhere (Kettunen and Laanti, 2006; Kettunen, 2007). Cost factors could be incorporated (Laanti and Kettunen, 2006). Performance measures of agility are subject to further investigation also in manufacturing (Adeleye and Yusuf, 2006). In general, value-orientation is one of the profound principles (Ojala, 2006). • Collecting more empirical evidence about the propositions in Table 4: Under what circumstances is agility a critical success factor in new software product development (Gunasekaran and Yusuf, 2002)? • What are the future development trends both in AM and agile software development models? Are they converging? Will such factors as sustainability and resilience be future competitive edges? • Comparing the concepts of current agile software development models with learnings from other potential reference disciplines—in particular IS. Apparently not many large industrial NPD organizations have yet fully achieved this kind of enterprise-wide agile software product development in practice. There is clearly a need to expand the view in software engineering, adopting applicable key learnings from other disciplines. This paper has attempted to bridge such gaps.