ITV: An Industry at the Crossroads


After decades of hype, interactive TV is becoming a reality. As digital TV and access to broadband gain momentum, ITV technology barriers are eroding and costs are falling, and the advent of personal video recorders, such as UltimateTV and TiVo, is sparking new interest in consumer-controlled TV.

Powerful technology is making it possible to deliver targeted advertising and interactive content. As deployments increase worldwide, ITV's projected growth is soaring. According to Ovum, a global analyst and consulting company, 357 million homes worldwide will have access to ITV by 2005, generating revenues of about $44.8 billion.


Although the industry is growing, it now faces a difficult challenge: making it practical and cost-effective for developers to create content and services across multiple platforms. Lack of agreement on standards among several groups — technology providers such as hardware manufacturers and middleware suppliers; content tools developers; content creators such as TV producers and games developers; and broadcasters — threatens to delay large-scale ITV deployment.

Without an industry-wide standard that is easy to work with, content authored once for a particular format or platform must be re-authored for another. An expensive and time-consuming process that results in a proliferation of content versions for developers to create and network operators to support, thereby discouraging deployments and reducing revenue for everyone involved.

In fact, in its August 2001 report "iDTV Developer Survey: Middleware Platforms and Standards," Strategy Analytics estimated that application developers will be using an average of four different middleware platforms in 2002 compared with 3.1 today.

Several ITV standards exist or are now in development. It is up to the industry to adopt one. The biggest challenge is determining the best way to enable ITV programming. Currently, the ITV industry is divided over a variety of approaches to use — a presentation engine, an execution engine or a combination that enable various ITV content and services on one device. At the most abstract level, both approaches convert data into visible ITV content, but differ in the way the content is created and processed.


Presentation engines interpret and render content created using mark-up languages such as HyperText Markup Language or Extensible Markup Language usually supplemented with scripting language such as ECMAScript. Authoring such "declaratively scripted" content is easy and inexpensive and is far and away the predominate method for creating Web content today.

Presentation engines such as Web browsers can compensate for differences between applications and receivers by ignoring data declarations they don't support or interpreting them with the best approximation the receiver can support. Enhancements to the authoring languages to support TV content have extended these benefits to the ITV environment.

Today's market favors presentation engines and applications that are as small as possible because they have the ability to run on as many receivers as possible and offer the ability to repurpose content and authoring tools from the Web to the TV and vice versa.


On the other hand, the execution engine renders content written using a procedural language, such as C, C++, Java or emerging languages like C#. In a procedural style, the application engine in the receiver knows a small set of tiny operations, but these operations can be combined by the programmer in a large number of ways to achieve everything that is possible in the device.

This approach, while flexible, is more verbose and generally more complicated to author than a declarative scripting application since the content is written in a programming language, just like a computer program.

Execution engines require a very precise match between the application delivered in binary computer code and the instructions supported by the execution engine. A single unrecognized bit will often cause an error that terminates the application, so extensive testing and debugging on each target system is often required.

It is important to bear in mind that neither approach is inherently more powerful than the other, although they tend to be optimized for different types of applications. Execution engines tend to be more efficient for computation, data processing, signal processing and operating system functions. Presentation engines tend to be more efficient for interchangeable applications that involve text, graphics, animation and user interaction. For content developers, the key difference lies in the required complexity and level of convenience for authoring the task at hand, as well as the functionality of the intended receiver.


Adopting industry standards allows the creation of content and services that work across multiple hardware, software and network platforms. Programming can be developed one time and then deployed on any TV network. Adopting open standards ensures innovation and greater competition through increased technology options and suppliers in the value chain resulting in a choice of middleware solutions for network operators, and authoring tools for developers.

To decide on a set of common standards for ITV application engines, industry decision-makers should consider several factors:

  • Is the standard based on non-proprietary technology?
    An open standard should be based on technology that is widely available, and provided on terms consistent with processes like those followed by International Standards Organization and other such concerns.
  • Is the standard "light"?
    Many network operators want to begin by introducing basic interactive services on their existing infrastructures of current-generation set-top boxes. The standard must allow very efficient implementations that run well in the limited memory and CPU speed of existing set-top boxes.
  • Is the standard scalable?
    The technology should easily scale from current-generation to advanced systems. It should be elegant enough to render attractive displays but strong enough to enable more complex two-way transactions.
  • Does the standard make content creation easy?
    The more time or skill level it takes to create and test ITV content, the more it costs. The industry will quickly gravitate to the lowest cost method of producing interactive content and will only use more expensive means if there is no alternative and the return on investment remains viable.
  • Does the standard have a shared cost infrastructure?
    The technology should take advantage of the economies of scale of existing interactive infrastructures such as the Internet, as well as existing content, data servers, developers, tools and business models. By repurposing these resources for ITV applications, early deployments do not bear the entire cost of building that infrastructure.
  • Will the standard facilitate the portability and interoperability of services?
    Content, services and applications should not have to be completely re-authored for different broadcast or distribution networks. With different hardware on different networks with different capabilities, some repurposing work will always be necessary; does the standard make this easier or more difficult?


At this time, available technology doesn't fully meet all of the requirements listed above. A solution comprising a presentation engine and an execution engine, both based on open standards, will ultimately offer developers the most flexibility but imposing such a standard today will burden receiver implementations with unacceptable costs and complexity in the current developing marketplace.

Presentation engine-only approaches based on Internet technologies that can be light and flexible, are easy to develop, provide a shared cost infrastructure and offer proven portability of services. This will provide the most cost-effective solutions for enabling ITV in the near term, but may have limits for future applications.

Further delay over standards will result in fewer products for consumers and network operators to choose from, will increase deployment costs and could set the industry back years. The only practical solution is for the industry to agree on a new evolutionary approach that standardizes different levels of functionality, or profiles, sufficiently well defined to insure content interoperability, and designed to be migration-friendly and open-standards driven.

Paul Mitchell is senior director at Microsoft TV Platform and president, board of directors of the ATV Forum.