W3C Ubiquitous Web Domain

Report: W3C Workshop on the Ubiquitous Web

Executive Summary

On March 9 and 10, 2006, W3C held a workshop at the Mita Campus of Keio University to discuss the vision of the Ubiquitous Web, with an examination of enabling technologies and consideration of what remains to be done to fulfill the vision. The Workshop provided an opportunity to share use cases, research results, and implementation experience.

The Call for Participation required participants to submit position papers. 32 position papers were received, and workshop participants came from a wide range of backgrounds. The workshop program was structured into position papers, breakouts, and discussion periods, with keynotes from Professor Hideyuki Tokuda from the Faculty of Environmental Information at Keio University, and Professor Larry Rudolph, from the Computer Science and Artificial Intelligence Laboratory, MIT.

The increasing pervasiveness of networked computing devices of widely varying capabilities poses challenges for how to create distributed applications that dynamically adapt to the context in which they operate, for instance, user preferences, device capabilities and environmental conditions. The Ubiquitous Web is an amalgam of the Web and pervasive computing, drawing upon concepts like markup, scripting, events, device coordination, declarative models, and the Semantic Web, to enable a new generation of context aware Web applications.

The W3C Staff are now following up with further study as to how the Ubiquitous Web vision can be realized through existing and future standards work at W3C. A follow on workshop is being considered for early 2007 on Declarative Web Application Models. A wiki and public email list have been set up to help with these goals.


Workshop Discussions

Use Cases, Security, Privacy and Identity

Don Retallack told us about the importance of device to device communication in factory applications, as well as the challenges posed by long latencies (the Web above 10,000 meters), and rapidly changing networks (e.g. military applications) where there are hundreds or thousands of devices that connect and drop out on a very rapid basis.

Shinichi Matsui described Panasonic's goals for a Web of home appliances, including the car, and opportunities to use the electrical power distribution cables for broadband networking. For instance the ability to use a wireless phone to view video of who is at the front door and to open the lock from wherever you are in the house.

Tahar Cherif described SHARP's perspective on the Ubiquitous Web. They see a huge challenge for Web accessibility. One possibility is to use a personal mobile terminal (e.g. a mobile phone or PDA) to assist with access to applications running on shared terminals (e.g. kiosks). The mobile device can help with book marking, authentication, and personal information.

Nobuo Saito introduced us to the PUCC (personalized ubiquitous computing consortium) in which academic and industrial partners are working together on a common peer to peer protocol layer that sits above a wide range of existing networking technologies, with a focus on home appliances.

Franklin Reynolds told us about Nokia's vision for smart homes, with wireless communication used to link appliances to gracefully meet the needs of human users. The Digital Living Network Alliance (DLNA) is focusing on an interoperable framework for appliances such as digital media servers and players. Challenges include the longevity of products and the resulting mix of standards and capabilities. Home users can't be expected to become system admistrators, and where possible systems should self organize. Service discovery needs to be ubiquitous and cater for dynamic discovery of devices, services and content.

Device Coordination

Kapil Sachdeva discussed the opportunities for using smart cards (e.g. credit cards) to enable stronger authentication of users for Web applications. Users have lost faith in the security of ActiveX controls, and Axalto have been looking at using Web Services as an alternative means to access smart cards. Ideally, smart cards would be exposed to Web browsers in a standardized way, along with the means for discover and bind such devices.

Gerrie Shults noted that device to device interactions suffer from implementations that are often complex and/or proprietary. This needs to be addressed by ubiquitous open standards. He presented a number of user cases, including the means for digital cameras to securely deliver photos to the home server and home photo printer. Business travellers may want to find nearby printing services (e.g. a local Kinko's branch) and arrange for the output to be delivered to a convenient location.

Don Wright discussed the issues involved in ubiquitous printing. The problems to be solved include discovery, delivery, capabilities and security. He described the approaches that have been developed by the PWG (Printer Working Group) and the UPnP Forum, and concluded by making some recommendations for the path forward.

Context Awareness, Meta-Data and the Semantic Web

Vincent Quint gave us an overview of the challenges involved in adaptation. An on-going effort is underway at W3C, but further research is needed on transformation, and there is a lack of standardization effort for negotiation protocols.

Reto Krummenacher talked about the role of the Semantic Web (triple spaces) in supporting the Ubiquitous Web. In particular, this enables adaptivity, scalability and global coverage.

Juan Ignacio Vazquez describe the potential for both the Semantic Web and informal approaches like folksonomies to support context adaptation. He presented examples of how these could be used to represent adaptation profiles, capabilities, and context information.

Johan Hjelm carried on the theme of context awareness as explored in the MobiLife project. He suggested three aspects of usefulness: relevance, filtering and availability. Challenges include discover, interaction and where and how to reach services. He then described work on trust and privacy. He sees a need for standarization on privacy management through policies, group management, user profiles, ontologies, and ambient network interfaces.

Distributed Applications

Satoru Takagi focused on the opportunities for involving small devices (ultra tiny computers) for example RFID based environmental sensors. He then went on to talk about work on combining SVG with map-based annotations, and the requirements for a lower case ubiquitous web platform.

Toshihiko Yamakami discussed several use cases focusing on different aspects. These included easy network configuration for digital televisions, interoperability requirements on devices, e.g. for memory and other capabilities, and mobile access to Web applications. He concluded that further standarization is needed, and that this will have to deal with the different business models in consumer electronics.

Charles McCathieNevile and Hallvord Steen gave us a joint presentation on the idea of a ubiquitous user agent for access to Web applications. The independence from the operating system enables scripting-based applications on any device through adherence to Web standards.

Kangchan Lee raised the question of how to coordinate applications across a wide range of environments, e.g. cars, mobile, home, office, shops and outdoors. He suggested that one way to do this was via events, and introduced WS-ECA, a rule-based Web services composition mechanism, involving events, and associated conditions and actions.

Summing Up

Further work is need to collect use cases and to clarify what standards work is merited. W3C is already working on a number of areas relating to the Ubiquitous Web. The number of working groups involved can cause resourcing problems for small companies. Good problem definitions can help a lot.

W3C should focus on mobile friendly technology rather than on mobile specific working groups. Perhaps too much time is spent on mobile rather than on non-traditional devices. There is also a concern about not leaving developing areas of the World behind. There was general agreement that there is a real opportunity to use Web technologies to make it easier to create context aware distributed Web applications, and that further work is needed, e.g. on scalable discovery, trust and identity, and the role of the Semantic Web.

Report: Dave Raggett