DireWolf - Distributing and Migrating User Interfaces for Widget-Based Web Applications

Abstract

Abstract. Web applications have overcome traditional desktop applications especially in collaborative settings. However, the bulk of Web applications still follow the "single user on a single device" computing model. Therefore, we created the DireWolf framework for rich Web applications with distributed user interfaces (DUIs) over a federation of heterogeneous commodity devices supporting modern Web browsers such as laptops, smart phones and tablet computers. The DUIs are based on widget technology coupled with cross-platform inter-widget communication and seamless session mobility. Inter-widget communication technologies connect the widgets and enable real-time collaborative applications as well as runtime migration in our framework. We show that the DireWolf framework facilitates the use case of collaborative semantic video annotation. For a single user it provides more flexible control over different parts of an application by enabling the simultaneous use of smart phones, tablets and computers. The work presented opens the way for creating distributed Web applications which can access device specific functionalities such as multi-touch, text input, etc. in a federated and usable manner.

Lessons Learned from the Development of the ROLE PLE Framework

“…Additionally, Go-Lab will also use Graasp and the OpenSocial Spaces specification to enable inquiry-based learning spaces for STEM education at school. On the other hand, researchers of the Learning Layers project 26 are using and extending the ROLE SDK as their learning platform (Kovachev et al 2013). As mentioned, the ROLE SDK is mainly meant for developers to extend their existing learning environments 23 Strophe.js, http://strophe.im/strophejs/ 24 WebSocket, http://www.websocket.org/ 25 Go-Lab, http://www.go-lab-project.eu/ 26 Learning Layers, http://learning-layers.eu/ or extend the ROLE SDK itself to support their requirements.…”

SeViAnno 2.0: Web-enabled collaborative semantic video annotation beyond the obvious

“…Since Web widgets are used for displaying and/or updating information, they can be successfully used as service interfaces in different environments and to build complex and composed visualizations of data. Due to their nature, complex Web applications can adapt to different devices, are customizable and extendable (e.g., more widgets can be added for extra functionality) or they can even be distributed across devices [10]. Important prerequisites for distributing individual elements of complete Web applications are a clear separation into conceptual and functional units, a context for managing separation, and cross-device communication between these units.…”

“…The different functionality needed to perform semantic annotation operations on videos has been split in multiple widgets, according to the underlying service structure. The communication between the widgets is realized using an interoperable communication mechanism between widgets, referred to as inter-widget communication (IWC) [10]. This is due to the fact that by default these components cannot share a common state and behave independently.…”

“…We used the widget-based approach, e.g. for distributing the user interface over different devices [10], for exploring near real-time collaboration in semantic video tagging, for mobile video tagging, or for different application contexts like storytelling. Even if SeViAnno is taking an outsider position in the semantic video annotation research community and is often ignored in overview papers [3] [11], it has been successfully used in research projects.…”

DireWolf Goes Pack Hunting: A Peer-to-Peer Approach for Secure Low Latency Widget Distribution Using WebRTC

“…To this end in earlier work we have presented the DireWolf framework [2]. DireWolf is a widget based Web application framework that allows distributing widgets over multiple devices while keeping the application state.…”