Liability for Hyperlinks in Australia
Information Technology Law: Internet Governance
Robert Progl, LL.M., holds a Master’s Degree in Media, Communication and Technology Law (UNSW, Sydney, Australia). He is a member of the Bars of Munich, Germany and practises intellectual property law and unfair competition law at the Law Chambers 'Progl Rechtsanwalt' in Munich, Germany.
His coordinates are: telephone: +49-89-40908035; facsimile: +49-89-40908036; E-mail: firstname.lastname@example.org; web site: https://www.progl.de.
The Internet made a rapid development in the past decade.
From connecting only a few university computers in the beginning it grew
to a global information and communication infrastructure for millions
today. This overwhelming success of the Internet was initiated by the
introduction of the WWW in 1990, when Tim Berners-Lee provided standards
for addressing, linking and transferring content over the Internet, thus
creating "an intuitive interface" that would permit "the
computers, networks, operating systems and commands [...] to become invisible"(1) so that users could communicate and access information directly. This
ability to connect content on different computers 'seamlessly' gave the
WWW its name as 'web' of interconnected computers. The 'hyperlink' as
technical means to perform the interconnection between websites is therefore
a core element of the WWW, a conditio-sine-qua-non for its effectiveness,
which results from the integration of isolated content into the immense
array of other information forming the web. As legal restrictions on the
freedom to link to any other page available would frustrate the very idea
of the web, Tim Berners-Lee claims a natural-law of non-liability for
links in the Internet.
Terms and techniques
Due to different techniques of linking, the question whether links are unrestrictedly permissible can not be answered in a general sense, but rather depending on the nature of the individual link. In the course of legal discussion special terms developed to distinguish different ways of linking, but because of technical misunderstandings these notions are not used homogenous or are misleading. A closer look at the terms and their technical background is therefore essential.
Hyperlinks in general
Webpages consist of an arrangement of HTML(3) commands which provide display instructions to the web browser on the user's computer. By interpreting these tags the web browser generates the page displayed on the screen. A hyperlink is the technical construction under HTML to make a reference to another data source. The basic hyperlink is a HREF(4)-tag followed by the URL(5) of the linked object. This causes the web browser to open the linked page when the user interacts(6) with the displayed link(7) (e.g. pointing, clicking, touching with the mouse).
"Surface"- and "Deep"-Links
These terms are used to describe links by referring to the target of the link. A surface-link points at the "entry"-page(8) of a website, whereas a "deep"-link(9) points at a specific page located deep in the hierarchical structure of the website. Generally conflicts arise about the legality of deep-links, when advertisements on the entry-page are bypassed(10). However, the notion of deep is purely fictional and not consistent with the architecture of the WWW. Each resource on the web has its own identification(11) and from a technical point of view a measurable distance (by clicks or in the actual structure of the files on the server) between an entry-page and an intended document can not be determined, notwithstanding the ambiguity of deepness of database driven websites.
Embedded (or Inline-)Link
The content of a website doesn't necessarily have to be located on the same webserver as the webpage itself. By using HTML commands(12) in the web page the user's browser can be instructed to request the referenced content directly from another host and embed it into the displayed web page. As no user interaction is needed and the display does not differ from the content retrieved directly from the webserver hosting the webpage, most users are unaware of the different origin of the embedded object. Some authors describe this kind of linking as Inline-Link. But as most HTML commands are 'inline' - even links to other parts of the same webpage or website - this term is not suitable as a descriptive term. Yet even the qualification as 'link' is arguable, because technically no reference to another content is made and no user interaction activated the link.
A common way to design web pages is to use frames, which divide web pages in several independent segments. These segments can be addressed individually to change their content without affecting the display of the other segments.
In this context framing describes the display of remote content located on another host in one frame surrounded by or side-by-side with parts of the linking web page in the other frames. Another web page can therefore be displayed as an integrated part of the initial webpage(13). However this procedure is technically not a hyperlink (nor are embedded objects), as no reference to another object is made, but rather the content is incorporated from outside into the own web page.
Framing occurs typically together with deep-linking, when the destination page is displayed inside the own frame set thereby substituting the navigation/advertisement content in the omitted frames of the linked webpage by own content or advertisments.
Copyright law provides the backdrop for most hyperlink disputes, however the infringement of copyrights by hyperlinking is rather unlikely. The Copyright Act 1968 (Cth) ('the Act') grants the originator of protected material a bunch of rights to protect his intellectual property. In relation to hyperlinks the exclusive economic rights of reproduction, adaptation and communication to the public may be relevant as well as the moral rights which confer the rights to claim authorship and to protect the integrity of the original work to the author. The right to performance in public is expressly exempted by s27(2) and s27(3) of the Act in terms of online communication.
Copyright protected material
The copyright protection of the content of the linked webpage like text, pictures, sounds, videos etc.(14) is assumed for the purpose of this essay. However, the protection of the underlying HTML-code under copyright law is arguable. From a technical point of view HTML documents are no computer programs, but documents consisting only of a combination of text blocks, which are provided with structure and are accompanied by other embedded objects. As no functions are solved by use of a sequence of commands, but only text is included into tags to determine its display HTML is only a markup language (Hypertext Markup Language) and not a programming language.
Copyright law can be infringed directly by a person, who not being the owner of the copyright, and without the license of the owner of the copyright, does in Australia [.] any act comprised in that copyright(16).
The single most important copyright right implicated by the transmission and use of works on the Internet is the right to reproduce the work in material form(17). It is arguable, however, whether this is an appropriate way to approach the protection of digital content. Traditional copyright law was designed to deal primarily with the creation, distribution, and sale of protected works in tangible copies. In a world of tangible distribution, it is generally easy to know when a reproduction has been made. In the Internet, however, it is often difficult to know precisely whether a copy of a work has been made and, if so, where it resides at any given time within the network. Information is sent through the Internet by breaking up the data into smaller units (packets) and sending these packets individually on its way. As these packets pass through the different servers it is arguable whether a reproduction of the work occurs. At least in the computer of the user a full digital copy is made in the RAM or in the cache folder respectively. Notwithstanding the transient nature of the copy in the computer's memory the amendment of s10(1) Copyright Act 1968 (Cth) includes now also digital reproductions which are not in tangible form as in material form(18). Looking a the technical side of the different links described above an infringement of the reproduction right by the link setter cannot be seen. On the one side the link itself contains the URL of the linked content and this URL is transferred to the user's computer. But copyright protection in the URL itself is rather unlikely, because it is only a method of referencing and comparable to a mere address - the URL as such has no the required level of originality(19) which would be necessary to acquire protection under copyright law. The linked content, however, is not provided by the link setter, but sent directly from the webserver hosting the linked content to the user's computer. The reproduction is therefore only done in the user's computer and no copy of the content is located on or transferred from the linker's own webserver(20).
Right to make an adaptation
S 31(1A)(vii) Copyright Act 1968 (Cth) grants the owner of the copyright in an original work the same exclusive rights in any adaptation made as in the original work. An adaptation is defined in s 10(1) of the Act as a version [.] or a translation [.] of a work which itself represents an original work of authorship. The display of webpages linked to via deep-links, framing and embedded objects often strips off parts of the linked page's content or adds additional content to it which may constitute the making of an adaptation by creating a new version of the linked material. But just by linking to a third party's content the substance of the original work located on the copyright owner's webserver is not transformed in any way by the link. One may say that the HTML code of the link 'instructs' the user's computer to perform the reproduction/adaptation and therefore the link setter may be responsible for the resulting display. How the page appears in the end on the display of the user, however, is not under the control of the link setter, but rather very dependent on the used browser. Whereas older browser versions are not capable of displaying frames and CSS(21), Lynx(22) doesn't even display graphics and on a PDA(23) or mobile phone the display is always deprived of most of the display features. In case of using CSS the outcome is even more unpredictable and on different browsers the display differs quite substantially. Because a specified display of a webpage 'as original' does not exist it is almost impossible to determine which difference in display is based on the way of linking and what is only a technical error or feature of the used display device. Therefore the user determines to a great extent the display of the linked material. The display on 'the majority of browsers' is too vague to be a fit basis to determine copyright infringements of a hyperlink, because then basically every hyperlink could constitute an infringement of the adaptation right under certain circumstances. Next to the question of who controls the way of displaying the linked content the concept of adaptation in the online environment involves more ambiguity in terms of level of originality of authorship and the determination of when the displayed content amounts to a new version of the work. In the case of framing the display of the linked site which is surrounded by a frameset of the linker's weppage will in most cases not represent a new original work of authorship, because no editorial change of the linked content is made by just juxtaposing a navigation bar or an advertisement banner next to it. The frames around the linked pages are not permanent additions, but easily separable by a single click on a mouse button. In a U.S. case where 'add-on' software altered the output of a computer game cartridge, this was not deemed as an unauthorised adaptation because it did not incorporate the copyrighted material in permanent form(24). Thus, framing does most likely not produce a new version of the work(25), however, dependent on the actual case the outcome may be different. When deep-linking to another webpage the display of this destination page is not altered at all. The entry-page of the other website, if something like that exists at all(26), can hardly be deemed as part of the linked webpage, because it is an independent document with an independent URL. The omission of the original frame set may constitute a new version, if the content of the frameset had any permanent editorial content. But as described above, frames are not a permanent part of a webpage. Next to the fact that in most cases only advertising and navigation bars are excluded, which are probably not essential enough to give the stripped document the quality of a new version, the final way of display is performed and determined by the user's computer and not by the linker. When embedding a third party's object inside the web page it is questionable what the original work in fact is. If the single object, e.g. a graphic, is the original work, then the process of embedding is not altering the display as such. But if the whole webpage which originally incorporates the linked object is deemed to be the original work, then the use of a graphic of that page may not be enough to create a new version of that webpage as only a very small part of it is displayed. In relation to computer programs the adaptation of a work is closer defined in s 10(1) of the Act as a translation/a version of a work being a computer program (whether or not in the language, code or notation in which the work was originally expressed) not being a reproduction of the work. Thus, already the translation of the source HTML-code into object code by the browser may be a new version of the work. Notwithstanding the fact that the translation is done solely by the user's browser, s 21(5) of the Act states that the processing of the source-code into object-code is not an adaptation, but rather a reproduction. A direct infringement of the adaptation right by hyperlinking is therefore rather unlikely, but courts may decide differently based on the facts of each particular case.
Right to communicate the work to the public
The Copyright Amendment (Digital Agenda) Act 2000 introduced a new right to communicate to the public with respect to online communications. It defines communicate as meaning to make available online or to electronically transmit [.] a work or subject matter [to the public]. The link itself clearly does not electronically transmit the linked content. Initially only the URL is transmitted with the webpage, but the following transmission of the linked material after the activation of the link can't be attributed to the link setter, because only the linked web server and the user's computer engage in transmission of data without any control of the linker over the content of the transmission(27). Further it is questionable whether setting a hyperlink in any form makes available online the linked content. The Revised Explanatory Memorandum gives as example the upload of material to a server which is connected to the Internet. But by placing a link on a webpage the content itself is not uploaded, rather the link only points to material already uploaded and made available online by someone else. Thus, a direct infringement of the right to communicate to the public is not constituted by hyperlinking. As an intermediary result it can be stated that hyperlinking is unlikely to constitute a direct infringement of copyright.
Authorisation of infringement
A person infringes copyright under ss 36(1), 101(1) of
the Copyright Act 1968 (Cth) if he is not the owner of the copyright and
authorises any infringement of the copyrights comprised by the Act. The
act names three requirements to be considered to determine authorisation
in case a direct infringement is done: the power to prevent the doing,
the [commercial] character of the relation between user and linker and
whether reasonable steps were taken to prevent [.] the infringement(28).
A direct infringement of the user by using hyperlinks in course of browsing
the WWW is rather unlikely in most of the cases, because the Copyright
Act 1968 (Cth) provides an exemption for prima facie copyright infringements
made during the technical process of communication. Under s43A of the
Act a temporary reproduction of a work or an adaptation of a work is not
an infringement if it is made as part of the technical process of making
or receiving a communication, unless the communication as such is an infringement.
The temporary reproductions and adaptations made in the memory of the
computer and on caching devices when browsing and viewing copyrighted
material are therefore not infringements of copyright. This result leads
to the question which communication (as such) constitutes an infringement
for the purpose of s43A of the Act. The definition of communication described
above gives guidance that any upload of a copyrighted document without
the authorisation of the copyright owner onto a computer connected to
the Internet as well as the electronic transmission of that copyrighted
material is an infringement of the right to communicate. Consequently,
the browsing of infringing content is a direct infringement by the user
by making unauthorised reproductions and adaptations in his computer's
memory. The browsing via deep-linking, framing and embedding objects are
no infringing communication per se, because if the determination of the
infringing character of communication is based on the fact that reproductions
and adaptations are made the outcome is a vicious circle. Moreover, the
intention of s43A of the Act is to de-criminalise the Internet user when
browsing the WWW - a narrow interpretation of the exemption clause would
prohibit the intended result, because next to hyperlinks any use of software
which alters the display of a webpage, e.g. ad-removers, browser add-ons
etc., would constitute a copyright infringement of the user. At least
with respect to infringing content an authorisation of the infringing
conduct of the user by providing the hyperlinks appears possible, when
hyperlinks are deemed as facilities for the making of the infringement.
The leading case of authorisation in Australia states that:
Implied license and fair use
Next to the economic rights the use of hyperlinks may affect also the moral rights of an author. Copyright Law(32) protects not only the commercial utilisation of the work, but also the personal and intellectual relation of an author to his work. In contrast to the economic rights the moral rights are personal rights and cannot be assigned(33) or waived, but remain always with the individual being the originator which is not always the website operator. The right to claim attribution of a work means the right to be identified as the author of the work when an 'attributable act' is done to the work or a substantial part of it(34). This attributable act is any kind of reproduction, adaptation or communication of the work. The counterpart of it is the right not to have authorship falsely attributed, which is infringed when a person's name is inserted in or affixed on the work, in a way which implies falsely that the person is the author of the work. With respect to hyperlinking a regular surface-link is not likely to constitute an infringement of those rights, because the change of authorship can easily be recognised by the change in the address-field of the browser and may be inferred from the displayed homepage. In the case of embedded objects, however, the author of the linked content (esp. in the case of graphics) is not easily conceivable, but rather the display suggests to the user that the author of the webpage is also the copyright owner of the embedded object and that they are transmitted from the same webserver, if the object itself doesn't contain any copyright notices. In some circumstances hyperlinking via deep-links and framing may amount to an infringement, when the correct author or at least the fact of change of authorship is not easily recognised, because the layout or design of the linked webpage is similar to the linking webpage or the copyright notice is omitted. The right of integrity of authorship grants the author the right not to have his work subjected to 'derogatory treatment'. This allows him to refuse any treatment resulting in the mutilation or material alteration of his work which could be deemed prejudicial to his honour or reputation. With respect to hyperlinks this right is likely to be infringed, when the linked content is displayed only in parts or if displayed in an context which is apparently inappropriate or undesired by the author. As a direct adaptation of the original work is not needed, but rather circumstance effecting the display of the work from outside are sufficient, framing and embedding objects may negatively affect the honour or reputation of the author. Unlike the economic rights hyperlinks may infringe the moral rights of an author giving him the power to object to certain kinds of links.
A hyperlink is generally composed of two elements - the HTML code, which is invisible for the user, and the visible part displayed on the screen, which usually consists of words, pictures and/or logos. The use of a protected trademark as the visual part of the link may constitute an infringement of trademark law. But also the special links subject of this essay may proof to create an infringing outcome if the linked content involves protected trademarks. A registered trade mark is infringed when a person uses as a trade mark a sign that is substantially identical with, or deceptively similar to, the trade mark in relation to goods or services in respect of which the trade mark is registered(35). The scope of protection is therefore restricted to commercial uses of the trademark in the registered class. Famous marks, however, are not bound to a certain registration class, but are protected against any dilution or unlawful exploitation of goodwill. In the case of a surface-link the use of a trademark as a visualisation is not infringing trademark rights, because the sign is used to accurately identify the website location of the registered trade mark's owner and no likelihood of confusion is given when the new page appears after the user activated the link. An infringement would only be likely to arise in the event that the website owner used the hyperlink to indicate some connection with his or her own goods and services, e.g. to set up the hyperlink such that it connected the user to another page of the linker's own website, rather than a webpage of the trade mark owner. If the linked page contains a protected trademark, the conduct of framing and embedding object may create a display of the linked content causing confusion about the origin of the offered goods or an affiliation of the trade mark owner and the webpage owner. By juxtaposing the linked trademark and the linker's trademark in a confusing way so that the user may assume that the trademark owner approved the use or is in any other way connected to the link setter a breach of the Trade Marks Act 1995 (Cth) is constituted. In U.S. legislation involving alleged breaches of trademark laws the linker setter always raised the constitutional right to free speech (First Amendment) as a defence, because the referencing to other material is deemed to be an act of speech. However, as in Australia no fundamental rights exist such a defence is unlikely to be successful.
Unfair trade practises
Most conflicts about hyperlinking arise in the commercial environment, when competitors use unfair practises to use each others webpages for own advertising or exploitation of the competitor's business reputation and goodwill. However, the Trade Practises Act 1974 (Cth) and tort law grant a wide protection against such conduct.
Misleading and deceptive conduct
Ss 52, 53 Trade Practises Act 1974 (Cth) ('TPA') prohibit misleading or deceptive conduct in trade and misrepresentations as to affiliations, approvals and endorsements. It is a breach of the TPA to represent that a company or products have a commercial connection with another company or its products if no such connection exists.
With respect to deep-links, however, the judge in the Ticketmaster v Tickets case(36) did not find a misleading character. The court concludes that deep linking by itself (i.e., without confusion of source) does not necessarily involve unfair competition. In dicta the court analogised hyperlinks to traditional indexing techniques, thus suggesting that they are benign and indeed helpful, and thus hardly tortuous. The court stated:
Where there are no trade mark registrations for a name or sign, an action may be brought under passing off. Passing off in the Australian context is a tort under the common law, which protects the goodwill and the reputation attached to a name or mark against the use of those (or deceptively similar) marks to confuse the relevant public, which as a result causes damage to the business reputation. The legal issues resemble those discussed above under the Trade Marks Act, especially framing in combination with deep-links may be used to establish reverse passing off, which is accomplished when the linker removes the name or trademark on another party's product and sells that product under a name chosen by the linker.
Misappropriation is a common law form of unfair competition developed to protect material which is not copyrightable. In a misappropriation claim, a defendant is accused appropriating a plaintiff's creation that is not protected by trademark, copyright or patent law. A typical misappropriation claim involves that the owner of the linked material has made a substantial investment of time, effort, and money in creating the content to the extent that it may be deemed as 'property' and that the linker has appropriated that material at little or no cost, such that a court can characterize conduct as 'reaping where it has not sown'. This must cause damages such as a direct diversion of profits or a loss of royalties. Hyperlinking via deep-links, framing and embedding objects may be tortuous when used to incorporate third party's online databases into the own webpage, when the latter are not protected by copyright laws.
The interest of the public in smooth (network-) communication and in unhindered flow of information prohibits a general judgement over the lawfulness of hyperlinks. Despite the narrow view of the copyright owner's lobby hyperlinks are not means of infringing copyright, but rather a way of observing copyright. The copyright owner disseminates himself by uploading the copyrighted material, a link only points to the server of the copyright owner and the communication remains fully under his control - if he deletes the content from his server, the link points into emptiness(37). In case of infringing copies available on the WWW the web site operator hosting the material is the right addressee for legal remedies as the removal of the content simultaneously also disables all links to it. Unfair competition law and trademark law suit better to limit the negative effects of hyperlinking in commercial activities as their ambit is limited to competitors and commercial interests rather than providing a general prohibition of hyperlinks for everybody. With the introduction of more sophisticated Internet services as XML and semantic links the actual view of linking and WWW will be outdated and a new concept of approaching the issues of linking will be needed. The traditional approach of copyright to protect reproductions is not the appropriate way of regulating the online environment which is based on ubiquitous reproductions as a technical necessity. A change to honour the effort and cost in producing the material in conjunction with general usage fees for the Internet may be more adequate. Hopefully present litigation doesn't bias future legislation to disregard notions of fair use and public access to information or else the biggest invention of the millennium may lose its power and usefulness.
|© 2002 Rechtsanwalt Robert Progl, LL.M.|