Continuous web personalization using selector-template pairs

Due to the fast growth of the Web and to the interactivity brought by Web 2.0, government, entertainment, and education services are more and more available through the Internet. However, this highly interactive Web still excludes people by not considering capabilities of all users. Although there are techniques and approaches towards Web Accessibility, issues related to different user needs may emerge in unforeseen contexts of use. This paper proposes a personalization approach that considers detailed observational data from usage and generates user interface adjustments (pieces of JavaScript code) from a set of selector-template pairs. Selectors are regular expressions used to identify patterns in event streams and templates are adjustments skeletons, referring user interface elements identified by the selectors. The approach supports continuous personalization of websites aiming to reduce the number of accessibility barriers and usability problems. This work is part of a long-term project and it reports the study conducted "in the wild" with 408 users, during a period of 30 months. The results point that the approach is a scalable solution towards Web personalization by inserting JavaScript pieces of code to tailor websites according to how they are used, on the fly.