Personal cloud services such as Dropbox are used increasingly to support collaborative work, even though they typically have poor support for tracking files and users' activities and collaborators often rely on other communication channels to be notified of changes. We present a meta-cloud storage service, MUBox, that, independent of a particular cloud storage service, provides improved support for collaboration. First, users can switch to activity views that list user activities rather than files, which is an example of an increasingly available feature in popular cloud storage clients. Second, MUBox introduces multi-user aware folder views that embed information on the last changes performed by collaborators. These folder views are enhanced based on a new concept of shadow files which act as placeholders for files that have been moved or renamed. A user study (N=16) with realistic folder exploration tasks shows that activity views have a significant effect on the accuracy and confidence of users in workspace awareness tasks, while shadow files significantly improve the speed, accuracy and confidence of users in traceability tasks. We describe how existing services could implement these features as well as a new concept for voting on changes to shared folders that could improve asynchronous collaboration.

While there are many established methodologies for information systems development, designing by example has not been formally explored and applied previously. Our work is also motivated by the desire to explore interface-driven development techniques that could complement existing approaches such as model-driven engineering with the goal of reducing the need for modelling and reengineering of existing applications and interfaces, while still supporting the development task. We explore the example-based technique for rapid development of powerful and flexible information management tools based on the example of Adobe Photoshop Lightroom, a system that was originally designed to support the workflow of digital photographers in a flexible way. We analyse experiments in which two new systems---one for managing collections of research papers and another for software project management---were developed based on the Lightroom paradigm. We derive a conceptual framework for engineering by example and assess the method by comparing it to traditional model-driven engineering.

Although web sites have started to embed semantic metadata within their documents, it remains a challenge for non-technical end-users to exploit that markup to extract and store information of interest. To address this challenge, we show how tools can be developed that allow users to identify extractable information while browsing and then control how that information should be extracted and stored in a personal library. The proposed approach is based on an extensible framework capable of using different kinds of markup to aid the extraction process and a unique fusion of several well-established techniques from areas such as the semantic web, data warehousing, web scraping and web feeds. We present the Sift tool which is a proof-of-concept implementation of the approach.