Quantitative assessment of mobile web guidelines conformance

Conformance metrics for the mobile web can play a crucial role as far as engineering mobile websites are concerned, especially if they are automatically obtained. In this way, developers can have an idea in numeric terms of how suitable their developments are for mobile devices. However, there are a plethora of devices with their own particular features (screen size, formats support, etc.) that restrict a unified automatic assessment process. This paper proposes a tool-supported method for device-tailored assessment in terms of conformance with Mobile Web Best Practices 1.0, including the definition of five quantitative metrics for automatically measuring mobile web conformance: Navigability, Page layout, Page definition, User input and Overall score. The behaviour of these metrics was analysed for different devices and different web paradigms, both mobile web pages and their equivalent desktop pages. As expected, the results show that mobile web pages on more capable devices score higher. In addition, 20 users took part in an experiment aimed at discovering how conformance-based scores relate to usability dimensions. The results demonstrate that automatic scoring approaches strongly correlate with usability scores obtained by direct observation, such as task completion time and user satisfaction. This correlation is even stronger for the device-tailored assessment than the one that assumes a general profile for all devices. For instance, results show a strong negative correlation between Overall score and task completion time: ρ (9) = −0.81, (p < 0.05) for the generalist approach and ρ (9) = −0.88 for the device-tailored one, entailing that mobile web guidelines and the metrics based on their conformance capture usability aspects. This result challenges the widely accepted belief that conformance to guidelines does not imply more usable web pages, at least for web accessibility conformance.