Controlled online experiments in which envisaged changes to a web site are first tested live with a small subset of site visitors have proven to predict the effects of these changes quite accurately. However, these experiments often require expensive infrastructure and are costly in terms of development effort. This paper advocates a systematic approach to the design and implementation of such experiments in order to overcome the aforementioned drawbacks by making use of Aspect-Oriented Software Development and Software Product Lines.
展开▼
