Fairfax welcomes the Department of Communications Information Technology and Arts Discussion of Options paper regarding the Review into the Scope of Datacasting and Enhanced Services ("Options Paper"). We are pleased that the Department has recognised as viable options, the proposals we made in our initial submission to this review and has deflected attempts to define datacasting out of existence.

In this submission we take the opportunity to re-emphasise the reasons why the definition of datacasting must be broad and allow both :

We also address why the "datacasting only" option (Enhanced Programming Option 3) is the simplest and fairest way of allowing commercial broadcasters to transmit programming enhancements without offending the ban on multi-channelling.

1 Datacasting Definition

Before discussing the detail of any of the proposed options we must emphasise that the most important feature of a datacasting definition is that it is broad and allows datacasting to develop to its full potential. To ensure this, datacasting should remain defined as all non-broadcasting uses of the broadcasting services bands and the definition of broadcasting should be refined to more accurately reflect the well-known and accepted characteristics of broadcasting.

Datacasting is the new broadband digital highway to every home in Australia, not just some new category of service to be found an appropriate niche in the Broadcasting Services Act. As recognised in the independent report from Malcolm Long and Chris Winters, datacasting has the potential to provide an enormous variety of services, many of which have yet to be invented. It would be a tragedy if a narrow definition of datacasting, designed to protect a narrow range of interests, forestalled the development of such services.

It is exactly because the concept of broadcasting is well understood while the concept of datacasting is still undefined, awaiting the innovative new services of the future, that this review should focus on refining the definition of broadcasting, rather than guessing what datacasting may be and then creating a self-fulfilling prophecy by constraining it to those predictions.

Moreover, datacasting can provide online services to the disadvantaged majority of Australians who do not have a computer with internet access and to those in the bush and regional areas who may never gain access to broadband cable or xDSL technologies and for whom a satellite connection has substantial up front costs (eg dish as well as set top box). Datacasting also provides broadband access to the large majority of Australian households without a broadband pay TV connection. These are important policy objectives in their own right, ones which underlie the Universal Service Obligation in the Telecommunications Act 1997 and many recent government initiatives to extend internet and broadband access.

Finally, and perhaps most crucially, the regulatory and competition policy effects of a narrow versus a wide datacasting definition should be considered. The definition of datacasting is less relevant to the current free to air broadcasters because they are allowed to both broadcast and datacast; thus they are less restricted in their use of their 7MHz of spectrum. In contrast, the definition of datacasting affects the growth prospects of an entire industry – albeit an embryonic one - and the customers it may serve. Rather than exploring new media and providing a whole innovative range of services to meet ever-changing customer demands, datacasters could be forced to follow a ‘home brew’ definition to develop services which are far more constrained than those offered overseas. This short-changes Australian consumers in the services they can obtain, limits export opportunities for Australia’s content creators, and limits - perhaps destroys - the business cases of potential datacasters.

Because of the huge upside and small downside of a broad definition and the uncertainty as to what datacasting may become, any definition of datacasting should always err on the side of being too broad rather than too narrow. The consequences of too broad a definition are limited, but the consequences of too narrow a definition are that innovation is stifled, and a new industry is stunted or killed. In essence, a narrow definition of datacasting may be anti-competitive, and this, especially in light of the Competition Principles Agreement, is as important a policy consideration as any of those explicitly stated in the Digital Conversion Act.

In order to ensure that the datacasting definition is broad, the broadcasting definition must be refined to reflect all of broadcasting’s essential characteristics. This means that the broadcasting definition must exclude both services which do not have the appearance of television (Datacasting Option 1) and services which are interactive (Datacasting Option 2).

Fairfax agrees with submitters such as News Ltd and Telstra that an important distinction between datacasting and broadcasting is whether the service is interactive (Datacasting Option 2). Fairfax also shares the concern, expressed in the Options Paper, that this approach alone might exclude some potentially successful datacasting services which do not involve interactivity. These would include non-interactive breaking news or financial ticker type information or other point-to-multipoint data download type services. It is for this reason that services which do not have the appearance of television (ie multimedia services which include only a limited amount of audio or video) must also be excluded from the definition of broadcasting.

However, Fairfax considers that data "carouselling" and other datacasting models which involve the user interacting with elements implanted in the set top box, rather than directly with the transmission at source, do fall within the interactivity category. The essential characteristics of broadcasting relevant to this issue are that a broadcaster has total control over what the viewer sees and all viewers see the same content at the same time. Data carouselling has neither of these characteristics - the user controls what he or she sees from a wide variety of content and users will rarely see the same content at the same time. It is irrelevant (and often invisible) to the consumer whether this interactivity is achieved using a back channel to the transmission source or not. It is artificial to define services on the basis of the location of the hardware with which the viewer interacts and it would infringe the principle of technological neutrality to define interactivity on a technological basis rather than on the basis of the characteristics of the service.

While it is true that the concepts of interactivity and on-demand are being extended by technology, there is no question that broadcasting services are completely non-interactive. Thus a definition which excludes all interactive services from the definition of broadcasting, irrespective of how that interactivity is achieved, will avoid the problem of changing technology highlighted in the Options Paper. This reinforces the point that this review should not attempt to limit datacasting by establishing allowable and non-allowable categories of interactivity but should focus on refining broadcasting by focusing on all its well understood characteristics.

Providing exceptions to the definition of broadcasting for both interactive services and services which do not have the appearance of television will significantly reduce the problems associated with drawing a line in the sand under the latter. This is because interactive services change the appearance of the transmitted material in ways that are not under the service provider’s control. The examples used in the Options Paper highlight this point. The Options Paper suggests that a service which involves the provision of normal television programs and a range of web pages on the same screen at the same time "would be difficult to classify as either broadcasting or datacasting" based simply on the appearance of the transmitted material. Yet such material could be classified on the basis of whether it is interactive or not. This again reinforces the point that both interactivity and the appearance of the transmitted material should be exceptions to the definition of broadcasting.

In any case, the need to define a precise boundary, based at least in part on the appearance of the transmitted material, is a task that cannot be completely avoided. This is because even if the interactive and/or subscription options for datacasting were adopted, defining an enhanced programming category to ensure that it is not used for multi-channelling will require the same difficult distinctions, based on the appearance of the transmitted material, to be drawn. This is why it is far better to decide the issue only once, in relation to datacasting, and require that all enhanced programming be done under the datacasting definition.

The Options Paper raises two other concerns regarding the interactivity option which Fairfax considers to be easily addressed.

The first is the apparent desire to exclude video-on-demand services from the definition of datacasting. This appears to stem from an extension of the government’s moratorium on the issue of additional FTA television licences (until 2007) to encompass protection of FTA broadcasters from all forms of video based competition.

Video-on-demand has been outside the definition of broadcasting since the commencement of the Broadcasting Services Act in 1992. Indeed the Act specifically excludes "a service that makes programs available on demand on a point-to-point basis" to enable video-on-demand to be provided outside broadcasting regulation. The reasons for this exclusion remain sound. Where a program is provided on demand, on a point-to-point basis, it loses those qualities of mass audience and service provider control that are essential to the concept of broadcasting and becomes more like the choice of a video tape from a video store.

The only basis on which video-on-demand can be considered similar to current FTA television services is that they both provide video programming. Yet this is equally true of video stores and movie theatres, and video-on-demand programming is more likely to resemble these than a FTA broadcast. Clearly the FTA broadcasters would have no justification for eliminating competition from these sources. Similarly, they have no justification for seeking to prevent datacasters providing video-on-demand.

The government decision not to immediately grant additional FTA TV licences does not constitute a broad policy of protecting FTA broadcasters from all forms of video based competition. Indeed, because this prohibition is inherently anti-competitive, the Competition Principles Agreement requires that it be applied narrowly and only to the extent that:

Thus not only is there no legitimate policy objective supporting the extension of this moratorium beyond its strict terms, but such an extension would be contrary to the government’s microeconomic policy reform agenda.

The second unnecessary concern is the Options Paper’s view that datacasters should be precluded from providing internet access on the basis it may allow users to access broadcasting services provided via the internet.

There is simply no basis for this concern because the argument supporting it is patently circular. One cannot argue that datacasting would allow users to access broadcasting services via the internet because the very purpose of this review is to determine the definition of broadcasting service (and datacasting service) which will indeed determine whether the first statement is true. Even under the current definition of broadcasting service it is unsettled whether "internet radio" services are broadcasting services.

There is no underlying policy reason why datacasters should be prohibited from providing access to broadcasting services. The only relevant policy consideration is the licence condition preventing FTA broadcasters transmitting a commercial radio service. This is no justification for imposing restrictions on all datacasters. As noted above, the prohibition on additional FTA licences must not be expanded beyond its strict terms for competition policy reasons.

In any case, providing access to a broadcasting service as a result of a user command while surfing the internet is quite different from providing a broadcasting service itself. At most a datacaster providing internet access will be retransmitting a broadcasting service, not providing the broadcasting service itself. Retransmission is treated quite differently in the Broadcasting Services Act from providing a broadcasting service.

Moreover, prohibiting provision of internet access via datacasting is profoundly contrary to public policy. It would be perverse result indeed if the very technology most likely to bring high speed internet access to regional and rural areas was the only technology over which the provision of internet access was prohibited. Indeed, while datacasting is less suited to the provision of internet access in city areas because of the high cost of spectrum and the capacity constraints of a 7MHz channel, it is well suited for rural and regional areas where spectrum is cheaper and lower populations densities prevent capacity constraints arising.

Finally, a "walled garden" approach would not only deny consumers access to the wide range of material which has made the internet so popular, but may limit the financial viability of the datacaster. While popular a couple of years ago, the "walled garden" approach has been abandoned by mainstream consumer network access providers, most notably MSN and AOL, and is mostly used today in corporate intranets. In any event, STBs will likely be equipped with PSTN modems which would enable exactly the same content to be delivered to the same device via a telephone line. Datacasters should be able to dynamically manage the bandwidth to enable users to have internet content in the most efficient and user-friendly manner possible, whether via spectrum or dialup. Artificial distinctions such as barring the internet from delivery via datacasting spectrum will only distort the manner of delivery at the expense of both consumer and producer.

While Datacasting Option 3 avoids some of the hard issues in defining the distinctions between broadcasting and datacasting, it is wholly artificial, would unreasonably constrain the range of datacasting services and will cause market distortions.

There is no policy or theoretical justification for restricting datacasting to a subscription only service, even in part. Subscription is not a basis on which one can distinguish between broadcasting and datacasting because broadcasting includes both free to air and subscription services. Datacasting also will include both free to air and subscription services unless artificial restrictions are put in place. There are a range of successful revenue models including advertising, transaction and subscription in related areas such as the internet. To force datacasting into only one of these categories would prevent it adapting to changes in this most dynamic market and thus may undermine its success.

Moreover, requiring datacasting to be subscription based will cause market distortions because it will lead to commercial decisions being made based on artificial regulatory restrictions rather than market forces. In particular it will encourage datacasters to charge subscription fees for services which may otherwise be free, thus disadvantaging consumers and reducing the takeup of datacasting.

FACTS’ submission to the Datacasting Review proposed 8 separate restrictions on datacasting, ostensibly to prevent datacasters providing a FTA television service. As noted above, the only restriction needed to achieve this limited policy aim is to impose, for non-interactive services only, reasonable limits on the amount of time full-motion video can be transmitted. FACTS expansion of this restriction to include interactive services is anti-competitive and unnecessary because FTA TV is passive not interactive.

FACTS’ proposal to ban datacasting of entertainment programming is overly onerous, and is as ridiculous as suggesting that news content should be banned from broadcasting. Given that much of the material currently available on the internet is entertainment programming (ie games sites, online games, music sites, film star fan sites, adult sites, sports sites, chat rooms etc) it is not sustainable to assert that the provision of entertainment is the exclusive domain of broadcasting. Preventing the datacasting of such material would kill many of the most attractive and natural datacasting services, such as internet access, online games and high speed download of music, books and games software. Moreover, defining datacasting by program content, such as "entertainment" or "drama", would impose onerous additional obligations on the ABA to be content police in relation to datacasting content. Content neutrality is the only tenable regulatory posture for this issue.

FACTS’ proposals to restrict video frame rates and to limit video to one quarter of the screen would make datacasting unattractive for consumers and anchor datacasting against the already outdated benchmark of low speed internet technology. Neither of these restrictions advance the limited policy objective of preventing datacasters providing a commercial television service because Fairfax’s proposed video time limit (for non-interactive services) already achieves this goal.

FACTS’ proposal to limit datacasting to subscription services would represent unreasonable interference with the commercial decisions of prospective datacasters and is based on the false assumption that being available free of charge is a distinguishing feature of broadcasting. This is clearly incorrect as broadcasting includes both free and subscription services and internet content is available free of charge to anyone with an internet connection.

2 Enhanced programming definition

The Options Paper’s analysis of the enhanced programming options clearly reveals that Fairfax’s "datacasting only" option provides the best solution to this issue. Not only is this option simpler to implement, fairer, more competitively neutral and less open to abuse than the other options; the alternatives are clearly unworkable.

As we pointed out in our initial submission to this review, there is no legislative requirement for an enhanced programming category to be created. While certain Government announcements indicated that FTA broadcasters would be able to broadcast program enhancements, this does not imply that an additional category need be created. Indeed legislative efficiency dictates that if the activities envisaged by the Government can be achieved within existing categories then no additional category should be created.

The datacasting only option allows a full range of enhanced programming to be provided by broadcasters. The multimedia limb of Fairfax’s datacasting definition will allow text, data, still images and up to 10 minutes of unrepeated video every half hour. This is sufficient to show all of the programming enhancements proposed by FACTS in their submission to the enhanced programming review ("financial information relating to the finance news, bushfire alerts relating to the news, features on stars of a movie[,] replays and sports statistics relating to a sports broadcast ... additional information on garden items, a feature on Bill Clinton related to a current affairs program and progress summaries on a sports match in the form of data.").

The only restriction would be that the full video "features" would have to be less than 10 minutes long and be followed by around 20 minutes of non video programming. This should not be onerous on the FTA broadcasters as FACTS states, in its enhanced programming submission that "[t]here will be long periods every day when the [enhanced programming] channel does not look like television, as it may just be text." Indeed, limits on the amount of video that can be shown are the best way to ensure that non-interactive enhanced programming does not become a television channel in its own right which consumers will watch continually without reference to the "main" program.

The other enhanced programming options proposed in the Options Paper must be rejected because they are unworkable and wide open to abuse.

As noted earlier, the program only option (Enhanced Programming Option 1) raises the same problems of boundary delineation based on the appearance of the transmitted material as were highlighted by the Options Paper in relation to Datacasting Option 1. However these problems are multiplied because an enhanced programming category will require a second set of precise definitions which will inevitably overlap with the first and because, as we stated in our initial submission "it is very difficult to imagine any enhanced programming which is not, on analysis, either multicasting or datacasting." This view is reinforced by the fact that, as noted above, all the enhanced programming options proposed by the FTA broadcasters will fit within Fairfax’s proposed definition of datacasting.

Enhanced Programming Option 2, which would define enhanced programming solely based upon the nature of the link to the main program is even more problematic. As indicated in the Options Paper, the fact that it would allow broadcasters to provide services which would otherwise be treated as datacasting under the enhanced programming category creates a host of problems:

Moreover, relying solely on the nature of the link to the main program to determine whether a service is enhanced programming or not will either impose an unmanageable burden on the ABA, or lead to rampant abuse of the category to multi-channel and datacast without paying the datacasting charge. As the Options Paper recognises, all the options for properly defining "incidental and directly linked programming" are vague and ambiguous. As a result either the ABA will have to assess whether particular material is sufficiently incidental and directly linked on a case by case basis, or broadcasters will be free to follow their own interpretation, which will invariably allow them to do anything they like. Neither option is satisfactory.

Finally, and perhaps most critically, neither Enhanced Programming Options 1 or 2 properly address the issue of the boundary between enhanced programming and multi-channelling. The original enhanced programming discussion paper makes it clear that this is a crucial boundary issue and defines multi-channelling to mean "the provision by a broadcaster of two or more separate streams of television programs in real time on a continuous basis over substantial time periods." Yet neither options 1 or 2 prevents the broadcast of continuous video material 24 hours a day under the enhanced programming category. This alone is sufficient reason for Enhanced Programming Options 1 and 2 to be rejected as contrary to Government policy.

The only concern raised by the Options Paper in relation to the "datacasting only" option was that it could not be used "if it was decided that broadcasters should be able to provide secondary television programs as enhancements (ie full video material)." This "concern" is in fact a ringing endorsement of the "datacasting only" option because it highlights that this option is the only one which successfully prevents multichannelling - the broadcast of a secondary full video television program in addition to the main program - and thus the only one which complies with the Government policy banning multi-channelling by FTA broadcasters.

To allow commercial FTA broadcasters to transmit a secondary full video television program in addition to their main program would constitute a major policy shift which would be tantamount to lifting the ban on multi-channelling. Such a major policy shift is outside the scope of this review and thus cannot be a material concern with the adoption to the "datacasting only" option.

3. Regulatory Approach

The Options Paper reveals a predictable division in relation to datacasting regulation between potential datacasters, who want light handed regulation by the ACA consistent with regulation of other online services, and broadcasters, who want heavier handed regulation by the ABA consistent with broadcasting services. In choosing between these two options the critical question is whether datacasting is more similar to broadcasting services or to online services, especially in the area of degree of influence.

The answer to this question is quite simple. Datacasting is, by definition, not broadcasting. Yet most, if not all, current online services would fit within the current definition of datacasting and are likely to fit within any new definition (if they were provided using the broadcasting services bands). Indeed the only similarity between datacasting and broadcasting is that they both use the broadcasting services bands. Common sense and the policy principle of technological neutrality suggest that this is not a suitable basis for regulating datacasting like broadcasting.

Moreover, the degree of influence of datacasting will be far more similar to that of other online services than broadcasting. This is true both because of the common nature of online and datacasting services (interactive and multimedia type services) and the fact that they are in an early stage of development and require special equipment to access.

Therefore, light handed regulation by the ACA is the only logical and consistent method of regulating datacasting services.