Integrating Interactive TV with the Web using XSL

Integrating Interactive TV with the Web using XSL

Vedhagiri Valliappan
Computing Laboratory
University of Kent at Canterbury
Kent CT2 7NF
United Kingdom
David H. Shrimpton
Computing Laboratory
University of Kent at Canterbury
Kent CT2 7NF
United Kingdom
Chris Dobbyn
Department of Computing
The Open University
Walton Hall
Milton Keynes MK7 6AA
United Kingdom


Interactive Television provides a novel environment for the support of multimedia content. We present some ideas and initial results from an investigation, where XSL is used to automate the transformation of content developed for the web environment into a form suitable for display on an Interactive TV environment.


Convergence, multimedia, digital television, MHEG, ITV, XSL


The recent introduction of set top boxes with Interactive Television (ITV) capabilities provides a platform for applications involving interaction with multimedia content. There is already a number of competing ITV application platforms on the market, using a variety of content representations. A second generation of Set Top Boxes (STB) that aims to provide standardised platforms is being developed, with XML-based languages as a notation for content representation [1]. One advantage of adopting XML [2] is its potential to enable content sharing between heterogeneous devices.

Figure 1 illustrates a scenario where a STB shares content with a broadcaster, the Web, and a remote hand held device such as a PDA. An example of an application that might exploit such an environment would be a TV advertisement, where the broadcaster displays a form designed for the Web on the viewers television and invites them to order goods directly. The user could be given an option to transfer the form to a hand held device where a number of fields could be filled in automatically (e.g. name, address, credit card no. etc.). The form could be completed by the user and subsequently returned to the broadcaster or the original content provider. The hand held device might have the option of returning the form via the STB or an alternative communication channel.

Although there are transport mechanisms that enable XML-based content to be exchanged between virtually any of these devices, there is currently no guarantee that the content developed for one device can be satisfactorily rendered on another.

Use of Extensible Style sheet Language (XSL) [3] enables a document defined in one XML-defined mark-up language to be transformed into another (via XSLT or XSL Transformations [4]) and to be formatted according to different display requirements using the XSL Formatting Objects. We are investigating the use of XSL to automate the transformation of XML based content between ITV and web environments.

Figure 1 - Integrated scenario


The transformation of Internet based content into ITV content requires more than the simple process of replacing XML markup tags from one domain with their equivalents in the other. The following issues have to be considered in order to produce an effective model of the transformation process.

Figure 2 - Typical device characteristics

Presentation and Layout

The display constraints of both environments have to be considered. Examples of such constraints include the availability of scrolling mechanisms, positioning method, window resizing and the screen resolution. The lack of a scrolling mechanism reduces the display area to the physical screen resolution, whereas the use of absolute positioning fixes the layout of text/images on the screen.

Navigation Models and Input Mechanisms

The different physical input devices used, directly influence the navigation model of the applications. Environments that incorporate mouse and keyboard permit applications based on sophisticated navigation models to be incorporated in the application; whereas typical television remote controls for example provide minimal support for navigation and selection. Navigation using such basic controls may require the user to move between several levels of context.

Text, Colours and Images

These features are directly dependant on the screen capabilities of the device. The range of permissible font styles and sizes often vary between platforms. Different devices use different colour schemes/palettes and correct rendering of colours has to be ensured by some means of colour mapping. Images may have to be resized or recoloured and image format conversions may be needed.


We are currently implementing a transformation of XML documents into formats suitable for display in ITV environments based on MHEG-8 [5]. MHEG-8 defines an XML notation for the series of MHEG standards currently used in a number of set-top-boxes.

The transformation involves converting content designed for a scrollable display environment with relative positioning to one that is non-scrollable and uses absolute positioning. Therefore automated content transformation requires the development of algorithms to perform the translation.

The example presented shows how a scrollable table can be transformed into an ITV environment with absolute positioning and no defined table structure. For small tables, the whole object may be displayed on a single screen; but if the table is too large to fit within the ITV display constraints, the table structure has to be reorganized into a suitable form with minimal loss of information.

Figure 3 - Original table in XHTML

Transformation of a table

Figure 3 illustrates a XHTML Basic [6] table that does not fit within a single ITV screen. The table structure is manipulated to be displayed across multiple screens, by duplicating each table header, with corresponding values from a single row, onto each screen.

The XSL translation process is executed in two parses.

The First Parse

The first parse performs a set of computations to determine the table size based on the number of characters in each cell and the associated font size requirements as illustrated in Figure 4.

Figure 4 - Output of the first parse

MHEG requires every displayable object to be encapsulated in a bounding box with fixed screen position values. In this case, the bbx and bby values denote the minimum bounding box values to display the text in each cell.

The Second Parse

The values generated are then fed into a second parse that either generates the whole table on a single screen or for larger tables, splits the table and generates the individual screens. Figure 5 illustrates the translated screen in MHEG containing values from the first row of the table.

Figure 5 - Table after transformation


XSL can be a useful tool for providing automated transformation between Web-based material and an ITV environment. Similar techniques could be used in developing an integrated platform where content might be seamlessly exchanged between heterogeneus devices.

There are limitations to the approach described, mainly due to variations between semantic models used within the various environments. Currently, Javascript is used to provide additional functionality where required. Use of common models such as DOM [7] events and SMIL [8] timing and synchronisation may offer a solution where integration of content between platforms is more straightforward.

We intend to investigate using devices such mobile phones and wireless PDAs to extend this work. Many of the issues raised here are applicable in these environments.


  1. Piesing J., The DVB Multimedia Home Platform, Proceedings of the IEE Colloquium on Interactive Television, Digest 99/200, 1999
  2. W3C, Extensible Mark-Up Language (XML) 1.0, W3C Recommendation,, Oct 2000
  3. W3C, Extensible Stylesheet Language (XSL) Version 1.0, W3C Working Draft,, Mar 2000
  4. W3C, Extensible Stylesheet Language Transformations (XSLT) Version 1.0,, Nov 1999
  5. ISO/IEC IS 13522-8 (MHEG), Information Technology : Coding of Multimedia and Hypermedia Information, Part 8: XML Notation for ISO/IEC 13522-5 (MHEG XML), ISO, Dec 1999
  6. W3C, Extensible HyperText Mark-Up Language (XHTML) Basic, W3C Working Draft,, Feb 2000
  7. W3C, Document Object Model (DOM),
  8. W3C, Synchronized Multimedia Integration Language (SMIL 2.0) Specification, W3C Working Draft,, Sep 2000
This work is part of the COMIT (Convergence of MHEG with Internet Technologies) EPSRC funded collaborative project involving the University Of Kent At Canterbury, Philips Research Laboratories and The Open University. Further details on COMIT are available at :