Vodafone Position Paper

W3C Workshop on Web Applications and Compound Documents

Daniel Appelquist
Vodafone W3C Advisory Committee Representative

25 May 2004

Reviewed by:
Jason Cline
Bill Howard
Timur Mehrvarz
Nick Roberts
Kevin Smith

Mixing SVG and XHTML

Towards the end of this year, devices which support combined markup of XHTML-MP and SVG will be available. The mechanism by which the SVG is embedded in the XHTML will be through the <object> element. Vodafone will definitely be taking advantage of these features to power its next-generation user interfaces for data services products. However, we feel a more comprehensive approach to creating compound XML documents put together from different XML languages would benefit the industry.

The value of delivering compound documents to handsets in a mobile context is to be able to deliver rich user experiences that are also consume less bandwidth and have lower latency.

As with SVGT support on mobile handsets, Vodafone would like to see compression technologies (specifically GZIP encoded content) be a requirement for composite documents

New Approaches vs. Existing Approaches

Any new approach that emerges should clearly differentiate itself from currently used techniques (such as multi-part MIME encoding). No new technologies should be developed unless there is a clear differentiation and advantage. Rather, existing efforts and/or available standards should be supported and referenced/profiled. It is therefore important that the workshop cover these existing approaches.

Intermediary Pattern

Vodafone would like to see any work on compound documents support an intermediary pattern, where downstream intermediating entities can participate in creation of compound documents from multiple content sources.

Intermediaries could provide "rendering" functionality (for example, detecting the user agent and reformatting output based on that agent's capabilities). Intermediaries can also provide more complex content "mediation" functions (for example, creating a composite "page" based on multiple content sources or content metadata sources.

In this model, multiple intermediary functions can be chained to produce a final output. For example, a mediation function can create a composite document from various sources which can then be passed on to a rendering function for device adaptation. In this way, content sources can participate in creation of personalized content applications, or in applications which adapt the display of content based on contextual information. One example of this could be displaying different information based on time of day or location. Another example could be filtering out or blocking adult content when the user is a minor.

It is important in any work on compound documents that this concept of intermediaries is made part of the requirements discussion. Requirements should not make the assumption that the only participants are "content providers" and "content consumers."