Engineering Web Augmentation software: A development method for enabling end-user maintenance

Nowadays, end-users are able to adapt Web applications when some of their requirements have not been taken into account by developers. One possible way to do adaptations is by using Web Augmentation techniques. Web Augmentation allows end-users to modify the Web sites' user interfaces once these are loaded on the clientside, i.e., in the browser. They achieve these adaptations by developing and/or installing Web browser plugins (augmenters) that modify the user interface with new functionalities. This particular kind of software artifacts requires special attention regarding maintenance as-in most cases-they depend on third-party resources, such as HTML pages. When these resources are upgraded, unexpected results during the augmentation process may occur. Many communities have arisen around Web Augmentation, and today there are large repositories where developers share their augmenters; end-users may give feedback about existing augmentations and even ask for new ones. Maintenance is a key phase in the augmenters' life-cycle, and currently, this task falls (as usual) on the developers. In this paper, we present a participatory approach for allowing end-users without programming skills to participate in the augmenters' maintenance phase. In order to allow this, we also provide support for the development phase to bootstrap a first version of the augmenter and to reduce the load on developers in both phases, development and maintenance. We present an analysis of more than eight thousand augmenters, which helped us devise the approach. Finally, we present an experiment with 48 participants to validate our approach.

Automated summary

Web Augmentation allows end-users to modify web interfaces with browser plugins after loading, which requires maintenance attention due to dependency on third-party resources. A participatory approach for end-users without programming skills to participate in maintenance is presented. Analysis of over eight thousand augmenters and an experiment with 48 participants were conducted to validate the approach.