Thursday 27 January 2011

Digital Economy Act Costs - an unlawful Prisoners Dilemma

On 17 January 2010, the draft Online Infringement of Copyright (Initial Obligations) (Sharing of Costs) Order 2011 was laid before parliament. It is the first substantive item of secondary legislation made under the Digital Economy Act 2010 and concerns about the simplest thing possible: how much copyright owners and ISP's will have to pay under the Initial Obligations Code (IOC). Despite its simplicity it has been misdrafted and it may also be unlawful.

Background

In case you haven't been following (and why should you?), the Initial Obligations Code is intended to allow copyright owners to send "copyright infringement reports" or CIR's to ISP's, that allege, roughly speaking, that the copyright owner has reason to believe that there has been some copyright infringement associated with one of the ISP's IP addresses. Some of these reports (which and how many to be determined by later legislation) must then be reported by the ISP to its subscriber. If a subscriber collects enough CIR's over a certain period of time (again to be determined) their IP address will go on a list which the copyright owner can then demand from the ISP which will (in theory at least) allow the copyright owner to match up different CIR's to the same subscriber.

There is no such thing as a free lunch. All this has to be paid for. OFCOM which polices the system will need funds to do so, ISP's will incur costs processing CIR's and there will also be an appeal system for subscribers who believe that a CIR was given wrongly which will have to be paid for. The Order is intended to split the cost 75:25 as between copyright owners and ISP's.

I want to avoid engaging in a discussion as to whether the principle of the act is a useful way to encourage people to stop infringing copyright, or a hopelessly misguided waste of time for everybody. I think my views on that are sufficiently well known. What interests me is whether there is anything technically wrong with the order itself, and I think there may be.

A technical miscalculation

Copyright owners are supposed to pay ISP's 75% of the costs that (OFCOM thinks) they will incur in handling the CIR's. OFCOM sets a rate per CIR that copyright owners must pay, which it can change from time to time. Copyright owners pay in advance, and this is where things come unstuck.

Paragraph 2 of the Order's schedule requires the copyright owner to notify in advance any ISP to which it intends to submit CIR's and give an estimate of the number of CIR's it expects to be submitting. Before the start of each notification period, the copyright owner must submit another estimate for the coming notification period and then pay the rate set by OFCOM multiplied by:

(a)the number of copyright infringement reports which the qualifying copyright owner estimates it will make to the qualifying internet service provider under the Code during the notification period; less
(b)the difference between the number of copyright infringement reports which that qualifying copyright owner estimated it would make to that qualifying internet service provider under the Code in the previous notification period and the number it actually made to that qualifying internet service provider in that period, if lower.
Read that carefully. It means that if the copyright owner systematically under estimates the number of CIR's, it will never have to pay the difference. Indeed there is nothing to stop the copyright owner giving an estimate of "1" in each year and paying the fee for exactly one report.

The proportion of fees payable by any copyright owner (or ISP for that matter) to OFCOM are also based on the copyright owner's estimates. The actual number of CIR's submitted never enters into the calculation. This amounts to a zero-sum game between all copyright owners. Each will have an incentive to submit the lowest estimates possible.

Unless I have completely misread the Order this seems quite mad and must (surely) be the result of a misdraft.

UPDATE:

I have been told that OFCOM stated at a stakeholders' meeting that the initial obligations code would limit the number of CIR's a copyright owner could submit in any notification period to the number it had estimated at the start of that period. This would be a change from the draft code published last year and is as suggested by Malcolm Hutty in the comments below.

This does avoid the strong incentive to underestimate the number of CIR's to be submitted, but it does not leave the scheme in a satisfactory state. It is still the case that fees payable to OFCOM are based on estimates and not reality. While that will penalise a copyright owner that overestimates, it will also penalise any ISP that receives an over-estimate of CIR's. The ISP's have no recourse in such a situation.

There is nothing in the Act that prevents the costs regime from dealing with reality and apportioning costs according to actual use (with balancing payments where necessary). OFCOM's suggestion looks like a quick and dirty bug fix, rather than a well thought through statement of policy.

Is the Order lawful?

According to James Firth, it appears that a different concern has been raised by the European Commission. Does the rule on cost sharing comply with the the "Authorisation Directive" (2002/20/EC).

The Authorisation Directive is, as EU directives go, pretty straightforward. The idea behind it is that regulation of those wishing to offer public communication networks should be extremely lightweight. Member States may regulate, but only within fairly tightly defined limits. In particular they may only impose conditions on the operator of a public communications network that fall within one of a list of type of condition (found in Annex A of the Directive).

The list is fairly long, but much of it is pretty common sense. There are only two which could plausibly allow the UK via OFCOM to make ISP's pay for the operation of the Initial Obligations Code.

First, which is what is being referred to in James's post, the UK (via OFCOM) could require ISP's to pay "administrative charges" in order to be authorised to operate a public communications network. Article 12 of the Directive makes it quite clear that "administrative charges" must be just that — charges for the administration of the general authorisation scheme for public communication networks. That does not include an obligation to pay for an appeals body to hear subscriber appeals or money to OFCOM to monitor and run a scheme for assisting copyright owners enforce their rights.

Second, it is permissible to impose a condition restricting the transmission of "illegal content" (such as content which infringes copyright) and "harmful content" (such as indecent images of children). It seems to me that the CIR/initial obligations scheme of the Digital Economy Act 2010 is not a "restriction" on the transmission of illegal content, but something quite different. What is more, a requirement to pay fees cannot be a "restriction" even on a fairly relaxed reading of the Directive.

In other words I cannot see anything in the Authorisation Directive that could allow the UK to impose a requirement to pay for the initial obligations scheme that is compatible with the Authorisation Directive. Unless I am missing something, the Order, as it stands is unlawful.

Interestingly, the 1709 blog asks whether the Order complies with the Technical Standards Directive and suggests that it does, because the imposition of fees and costs under the Order is a restriction permitted by the Authorisation Directive. Alas, whatever view one might take about technical standards, I don't believe the Order complies with the Authorisation Directive.

Conclusion

The Order isn't really news and I am sure the government believes they are complying with the Authorisation Directive. We disagree and only time (and perhaps the ECJ) will tell, but it does worry me that in this already contentious area the first piece of subordinate legislation is already buggy.

Monday 3 January 2011

A new kind of copyright? Graphical user interfaces in the ECJ.

A decision of the European Court of Justice published before Christmas concerning whether a graphical user interface is protected by copyright causes me concern. I have not seen any detailed analysis of the decision which may be due to Christmas and New Year stupor. I plead the same excuse for what follows.

The case Bezpečnostní‌ softwarová‌ asociace‌ –‌ Svaz‌·softwarové‌ ochrany v Ministerstvo kultury C-393/09 has a rather involved background an excellent account of which is provided by Martin Husovec on his blog. Very roughly speaking, a Czech organisation called the Security Software Association (BSA) wished to set up a collective licensing scheme for computer software which included the right to transmit works by cable television. The point in issue was whether the broadcast of the graphical user interface of a computer program could infringe copyright (and would therefore require licensing).

Copyright in computer programs was harmonised across the EU under the Software Directive (91/250/EEC). Article 1(2) applies the directive to "the expression in any form of a computer program". The first question referred to the ECJ was whether a graphical user interface could be described as a computer program "in any form". The ECJ said "no". One fact influencing the court was that Article 10(1) of the TRIPS Agreement requires the protection of computer programs "whether in source or object code". Clearly source and object code are examples of forms of expression of a computer program.

Similarly the 7th recital to the directive states that the term "computer program" also includes "preparatory design work leading to the development of a computer program provided that the nature of the preparatory work is such that a computer program can result from it at a later stage". The common element of these examples, thought the court, was that they lead to the reproduction or creation of a computer program. A graphical user interface does not enable the reproduction or recreation of a compute program, therefore it is not the expression of a computer program "in any form".

This all seems reasonable so far. Although a graphical user interface might contain one or more images or graphical works that would attract copyright as an "artistic work" in the normal way; attempts to claim copyright in an abstraction of the user interface — such as its mode of operation — have tended to founder in the English Courts in cases such as Nova Productions v Mazooma Games [2007] EWCA Civ 219 (concerning features of the play of two computer Pool games) and in Navitaire v Easyjet (concerning a clone of a air travel booking system).

However, the court went on to consider — although it had not been asked to do so — whether the graphical user interface of a computer program might be protected by the "ordinary" law of copyright. Here the court appears to have committed the logical fallacy of affirming the consequent. To explain why I need to say a something about earlier developments in European Copyright Law.

The Information Society Directive (Directive 2001/29/EC) ("INFOSOC") has partially harmonised other forms of copyright in the European Union. Article 2 INFOSOC requires that Member States create in their domestic law a "reproduction right" for "authors, of their works". Although described as a "right" it amounts to a prohibition on anyone else reproducing (in whole or part) a work without the author's permission. In other words it is a copyright. Although INFOSOC does not define "works" (or indeed "reproduction" or "reproduction in part"), the European Court of Justice in the earlier case of Infopag v Danske Dagblades Forening C-5/08 held that the directive only applies to a work that is "original in the sense that it is the author's own intellectual creation".

So to the apparent fallacy: in Bezpečnostní‌, the ECJ reasons thus:

The Court has held that copyright within the meaning of Directive 2001/29 is liable to apply only in relation to a subject-matter which is original in the sense that it is its author’s own intellectual creation (see, to that effect, with regard to Article 2(a) of Directive 2001/29, Infopaq International, paragraphs 33 to 37).
Consequently, the graphic user interface can, as a work, be protected by copyright if it is its author’s own intellectual creation.
In other words the court deduces from a statement of the form "P implies Q" a conclusion that "Q implies P". Infopaq establishes that being the author's "own intellectual creation" is a necessary condition. That does not mean it is a sufficient one.

Now the judges of the ECJ are, for the most part, bright and well educated. They will be well aware of the dangers of affirming the consequent. I think it is safe to assume they don't mean quite what they appear to be saying. In particular, the idea that any form of intellectual creation is protected by the general law of copyright would be quite revolutionary.

For example, INFOSOC does not exclude from copyright protection inventions that could form the subject of a patent, or designs that might be protectable under one of the European design rights. Although both might well be some author's "own intellectual creation", I doubt very much that copyright is intended by anyone to extend so far.

In English law merely being creatively original is not enough for a work to obtain copyright protection. For example in Creation Records v News Group Newspapers [1997] EMLR 444, the judge held that it was not even arguable that the scene for the Oasis album cover of Be Here Now could be protected by copyright. It simply did not fit into any existing protected category — and the record company attempted to argue that it was a dramatic work, a work of artistic craftsmanship or even a collage.

So I assume (in hope) that all the ECJ are guilty of is a failure to show working and that they have in mind some other criteria that must be satisfied before a work is protected by copyright. What those criteria might be the ECJ do not say which leaves me with the following questions: if a graphical user interface is protected by copyright, what kind of a work is it and what about it is protected?

It may be possible to gain some inkling as to what is in the ECJ's mind from their answer to the second question posed to them. They were asked if graphical user interfaces were protected by copyright in computer programs (the first question), would television broadcasting of the graphical user interface be an infringement of that copyright (by communication of that work to the public)? Although the ECJ had answered "no" to the first question, ever helpful they considered whether, on the assumption that graphical user interfaces were protected by "ordinary" copyright, TV broadcasting of them could constitute infringement. Again the ECJ thought "no" because the viewers "cannot use the feature of that interface which consists in enabling interaction between the computer program and the user".

It seems from this that the ECJ have in mind not the graphical elements of the user interface, but something about its operational behaviour. A sort of copyright in interactivity, which cannot of course be infringed by TV broadcasting because you cannot interact with the subject of the broadcast. If that is what the ECJ have in mind, then I am troubled.

First, what kind of a copyright is this? Is it an artistic work, a literary work or something else (a work of artistic craftsmanship perhaps)? For the purposes of English law this matters. Different species of work are protected in different ways, even if the differences are sometimes rather subtle. The ECJ do not tell us the answer to this question.

Second, where is the graphical user interface "fixed". In English law a work is only protected if and when it is fixed in some permanent form. If I make a really excellent speech, I will have copyright in the speech when, but only when, it is recorded. Similarly if a group of musicians create an exciting musical work while jamming together in the studio, the work will not be protected until fixed in some form. Is the graphical user interface fixed in the computer program's source code? If so, copying the code may be an infringement of the copyright in the graphical user interface. If it is not fixed in the code, where is it?

Lastly, there are numerous special rules that apply to copyright in computer programs. For example there are rights to make back-up copies, to test and to decompile for certain purposes. If the ECJ is right, these would not apply to a user interface which represented an author's own intellectual creation. In English law, moral rights do not apply to computer programs, but would (presumably) apply to a suitable user interface. Both of these conclusions could be awkward.

Not all is bleak. The ECJ emphasize that when assessing whether a graphical user interface is protected by copyright as its author's "own intellectual creation", a national court should ignore components of an interface that are "differentiated only by their technical character". It will not be every, or every aspect, of a user interface that will be protected. Despite this limitation I suspect that this decision, if taken to its logical conclusion and followed by later cases, will require some careful rethinking about copyright in the computer industry. Just what we need.