Satama Interactive has experience in developing solutions for interactive services across a range of digital media, in particular focusing on mobile and web-based delivery. The following document provides an outline of some of the key challenges we see in the field of cross-platform content services, in a multi-device and multi-network environment.

The widespread adoption of digital content management has enabled content creators and service providers to dramatically increase the value of their content assets by delivering content in a targeted way to an increased number of channels. Content has become a more flexible asset that can be presented adaptively with regard to the delivery platform (device), the user (profile and preferences) and usage context (time and position etc.). Efficiency and manageability is increased by the ability to create content once and then easily adapt it to various platforms.

XML and associated technologies provide a robust and highly flexible infrastructure for modelling, storing and adaptively presenting web-based content. Several challenges still remain, however. Many of these concern the relationship and division of responsibilities between standards and the products that implement or are affected by them. A question requiring specific attention is how to provide user-friendly support for high-level metadata and ontologies in content management solutions.

Providing cross-platform content is challenging in particular because content creation and content adaption are not clearly separated concerns in current content management solutions, even though the requirements and basic principles are well known. Many content management tools are currently too closely tied to individual delivery methods and technologies to truly provide a user-friendly, high-level environment for creating and adapting content. Often content management platforms also put the responsibility for ensuring separation between content, presentation and logic on the shoulders of the developer or system integrator. Standard means for handling these issues are required to reduce the amount of redundant implementation work. Providing support for initiatives such as XForms, which encapsulate elements of interaction logic, will be important in realising these concepts.

One of the disputed issues in handling separation of content and presentation relates to the choice between template-driven and content-driven presentation generation. In the former, a template (based HTML, WML, or other destination format) is marked up with references to XML documents, in a manner similar to embedded database queries, and content is retrieved from documents only as required by the template. In the latter approach, an XML document is transformed by a stylesheet (XSL) to determine the rendering of each document element. Both approaches have their strengths and supporters, and often they are used side-by-side in a single system (for example by using templates to describe overall page layout and stylesheets to render the separate page elements).

Raising the abstraction level of content management tools to encapsulate the underlying issues of separating content and presentation is a relevant challenge. How can usability be improved while retaining flexibility? What kind of tool support is required at each stage of the workflows? Where is automation most relevant?

The metadata requirements for cross-platform content are manifold. The most immediate concern is to provide information useful in automatically adapting content to a specific platform or device and user. Metadata may include elements such as:

• Locale (languages, countries and cultures content applies to)
• Device (devices content is suited for)
• Technical attributes (formats, size, etc.)
• Delivered keywords (e.g. HTML META-tags etc.)
• Indexing and keywords for content management and selection purposes
• Audience (high-level grouping of target audiences)
• Interests (more specific targeting based on interests of the user)

It is important to reach concensus on the metadata mechanisms used to adapt content (through controlled vocabularies and efforts such as CC/PP), as well as the actual means of attaching metadata to content (annotation), in order to enable automatic adaption based on machine-to-machine negotiation. Applying RDF (Resource Description Framework) to specific vertical domains, perhaps using OIL (Ontology Interchange Layer), is a potential path to getting this type of more intelligent metadata into real-world applications.

In addition to metadata focused on individual content elements and documents, it is becoming increasingly important to describe collections of content items - what the semantics of the collection are and what role each element plays. The emergence of complex composite documents with a temporal dimension, such as SMIL, further emphasizes this requirement. Protocols for content exchange also address part of this area, for example with emphasis on business rules for syndication (ICE).

The increasing multitude of devices available for the deployment of services and content raises some questions regarding the role and scope of standardisation in general.

• How quickly do and should standards keep up with the latest devices?
• How should the standards be structured and segmented in order to cover all device types, current and future?
• How is the boundary defined between what is covered by standards and where proprietary practises should be used?