Abstract Agile software development methods have been increasingly adopted by many organizations at different organizational levels. Whether named agile adoption, agile transition, agile transformation, digital transformation or new ways of working, the success of embracing this change process mostly remains uncertain. This is primarily because there are many ways of evaluating success. Based on the existing agile assessment models, we developed a model of principles with associated practice clusters that serves as a core for a new agile assessment model that is capable of assessing agile organizations at different scale. Towards our ultimate goal to establish a lightweight, context-sensitive agile maturity model, we validated our initial findings in an expert interview study to identify improvement points, and ensure the at hand model’s applicability, coherence and relevance. The results of the interview study show that the structure as well as the content of our assessment model fits with the experts’ expectations and experience.

Agile software development is considered as a game changer by some [1]. There are others who approach the domain more skeptical [2]. Clearly, there is a gap in terms of how Agile is perceived. This gap could be reduced, if proper measures would have been timely employed. Agile maturity assessment models have been helpful in providing means and guidance for reducing this gap. Yet, a detailed look into existing models, comparative studies and literature reviews in the domain suggests that the agile maturity assessment models themselves are far from being mature. Further, the gap between what the proposed models offer and what industry really needs is frequently discussed by those comparative studies. In this study, we reflect on the existing agile maturity assessment models, compare them against a comprehensive set of criteria derived from the literature and from our background in assessing the capabilities of software development organizations. We conclude that none of the analyzed agile maturity assessment models are sound enough to be used in a practical context. Nevertheless, some of the models have interesting elements that can be reused for the development of a new agile assessment model.

Learning interpretable representations in an unsupervised setting is an important yet a challenging task. Existing unsupervised interpretable methods focus on extracting independent salient features from data. However they miss out the fact that the entanglement of salient features may also be informative. Acknowledging these entanglements can improve the interpretability, resulting in extraction of higher quality and a wider variety of salient features. In this paper, we propose a new method to enable Generative Adversarial Networks (GANs) to discover salient features that may be entangled in an informative manner, instead of extracting only disentangled features. Specifically, we propose a regularizer to punish the disagreement between the extracted feature interactions and a given dependency structure while training. We model these interactions using a Bayesian network, estimate the maximum likelihood parameters and calculate a negative likelihood score to measure the disagreement. Upon qualitatively and quantitatively evaluating the proposed method using both synthetic and real-world datasets, we show that our proposed regularizer guides GANs to learn representations with disentanglement scores competing with the state-of-the-art, while extracting a wider variety of salient features.