1. Introduction

In the first action of its kind in Hong Kong, the operator of a financial web-site, HK Finance.com Limited, has issued a Writ against the operator of a rival site, Prosticks.com Limited, alleging that unauthorised hypertext links into HK Finance.com's web-site are unlawful.

This action is based on a claim by HK Finance.com that 3 hypertext links taking users from a site operated by Prosticks.com to the sub-pages of HK Finance.com's site, by-passing its home-page and missing the advertising on its home-page, infringe its copyright. Although the courts in Hong Kong have not previously ruled on whether deep-linking amounts to a copyright infringement, this issue has been considered by courts in the U.K, the U.S. and, most recently, the Netherlands.

In each of these cases, the complainant has been the operator of a content based web-site which derives revenue from advertisers on its home-page and, accordingly, has sought to restrict access to its site except through its home-page. Whilst the notion of "freedom to surf" being paramount is attractive, it is also easy to have sympathy with operators of content based web-sites who wish to prevent deep-links into their sites.

The purpose of this article is to consider the ability of operators of such sites to use copyright law in Hong Kong and the U.K. to prevent deep-linking into their sites.

2. The cases so far

2.1 Shetland Times Ltd v Wills1

The primary argument considered by the court in the Shetland Times case was whether the internet is a cable programme service within the meaning of section 7 of the Copyright, Designs and Patents Act 19882 and whether the contents of web-sites are "cable programmes" which are capable of protection. In essence, the argument put forward by The Shetland Times went as follows:

1. All web-sites, including The Shetland Times' site, are "cable programme services" for the purposes of the Copyright, Designs and Patents Act and that the contents of web-sites are "items included in a cable programme service" which are defined as "cable programmes"3. Specifically, it was alleged that the headlines of the news stories featured on The Shetland Times' web-site were cable programmes.

2. The defenders' web-site was also a cable programme service, and the inclusion of headlines (cable programmes) from The Shetland Times' site on the defenders' site (a cable programme service) amounted to an infringement of copyright.

Lord Hamilton of the Court of Session in Edinburgh held that there was a prima facie case that the incorporation by the defenders of headlines from The Shetland Times' web-site amounted to a copyright infringement. However, this finding was not critical to the success of the action, because it had also been alleged that the defenders' copying of headlines from The Shetland Times' site gave rise to an independent infringement of copyright (copying a literary work). The judge held that this allegation was at least arguable and, having considered the balance of convenience, granted an interim interdict in favour of the pursuer.

Both arguments put forward in the Shetland Times case relied on the linking site including material (headlines) from the target site. If such material in itself constituted a copyright work (as was argued by The Shetland Times), then the copying of the material would constitute a stand-alone copyright infringement, and there would be no need to rely on an argument that such material amounted to a cable programme.

The only reason for relying on an argument that material which constitutes a link from a linking site to a target site is a cable programme is if that material is not capable of being protected by copyright in its own right. For example, the headline "Plane crashes" is clearly not sufficiently original or substantial to enjoy copyright protection as a literary work, although it may nevertheless still be an "item included in a cable programme service" in which copyright subsists. This argument was not, however, fully considered by the court in the Shetland Times case as it settled before trial.

2.2 Ticketmaster Corp. et al v Tickets.com Inc.4

In August this year, the California Central District Court in the U.S. handed down its decision in an application by Ticketmaster for a preliminary injunction restraining Tickets.com from using hypertext links from Ticket.com's web-site at www.tickets.com to the sub-page of the www.ticketmaster.com web-site where users could purchase the tickets for specific events.

It should be noted that the ruling of District Judge Hupp in this application was made on the basis that it would not have precedent value. It should also be noted that much of this decision dealt with the use of "spiders" by Tickets.com to crawl and extract information from the Ticketmaster site. In fact, the only allegation of copying on the part of Tickets.com related to the uniform resource locators (URLs) of the web-pages on Ticketmaster's web-site to which Tickets.com had set up deep-links. In this regard, the court considered it unlikely that such material would be protectable, as the URLs appeared to the court to contain only functional and factual elements and no original material.

The application for a preliminary injunction against Tickets.com was dismissed.

2.3 PCM v Kranten.com5

More recently, in a case reminiscent of Shetland Times, the Rotterdam District Court held that Kranten.com's links into the sub-pages of PCM's web-site do not infringe its copyright.

PCM, a major newspaper publisher in the Netherlands, had applied for an interim injunction to restrain Kranten.com from deep-linking into articles set out in the individual pages of its web-site. As with the Shetland Times case, the links consisted of headlines from the target site in which copyright was alleged to subsist.

In this case, however, the court considered that Kranten.com's copying of headlines from PCM's site was subject to a fair dealing defence of reporting current events and, further, that deep-linking was in itself lawful.

3. Alternative claims for copyright infringement

Of the three cases referred to above, the claimant in only one (Shetland Times v Wills) was successful. This claim was based on a very particular copyright argument, which will only apply in very limited circumstances. However, this is not the only possible basis for a claim against deep-linkers for copyright infringement.

3.1 Making a copyright work available over the internet

Section 26 of Hong Kong's Copyright Ordinance sets out an infringing act which is not included in the U.K's Copyright, Designs and Patents Act 1988. In this regard, section 26 states:

"(1) The making available of copies of the work to the public is an act restricted by copyright in every description of copyright work.

(2) References in this Part to the making available of copies of a work to the public are to the making available of copies of the work, by wire or wireless means, in such a way that members of the public in Hong Kong or elsewhere may access the work from a place and at a time individually chosen by them (such as the making available of copies of works through the service commonly known as the INTERNET).

(3) References in this Part to the making available of copies of a work to the public include the making available of the original."

The essence of sub-sections (1) and (2) is that it is an infringing act to make available a copy of a copyright work on the internet, and sub-section (3) extends this prohibition to original copyright works. The main act covered by this section is the uploading of a copyright work onto the internet without authority. In this regard, it is irrelevant whether a copy of the work is already available elsewhere on the internet (for example if the copyright owner has already uploaded a copy of the work onto his web-site).

Whether or not the scope of section 26 is sufficiently wide to cover unauthorised hypertext links will depend on the interpretation which the court gives to the words "make available". In this regard, it may be argued that a hypertext link automatically makes available any copyright works on the target site.

Support for this argument comes from the reference in section 26(2) to making the work available in such a way that members of the public in Hong Kong or elsewhere may access the work at a time individually chosen by them. The very purpose of a hypertext link is to make a particular web-page available to a user immediately, without the user having to type in the URL. In this way a hypertext link allows users to choose the time at which they access the work, which would bring the use of hypertext links to within the definition of "make available" set out in section 26(2). This argument would be particularly strong in the event that the link takes the user to a restricted web-page which is available only to subscribers after the provision of personal information or payment of a fee.

However, the counter-argument to this view is that the work in question is already available on the target site and the hypertext link merely facilitates access to the target site and any copyright works contained on that site. This counter-argument is persuasive as otherwise one would be left with a rather bizarre situation of a web-page being simultaneously "made available" on the internet both by being uploaded onto the internet and by being linked to. This is consistent with Article 6 of the WIPO Copyright Treaty 1996 upon which section 26 of the Copyright Ordinance is based6. The better view appears to be that the act of uploading constitutes the "making available" and that, except perhaps where the link is to a restricted web-page, the hypertext linking simply facilitates access to a web-page which is already available.

3.2 Copying the URL of the target site

3.2.1 Technical background

The URL of a web-site (eg http://www.twobirds.com) consists of three distinct elements:

* http://- this part of the URL tells the user's browser to use HyperText Transfer Protocol (http) to locate the web-site.

* www - this is a host entry which informs the user's browser that the host web-server is on the world wide web.

* twobirds.com - this is the unique domain name for the operator of the site which the domain name server (DNS) translates to an IP address consisting of 4 sets of numbers (bites) - for example the IP address which is equivalent to twobirds.com is

Accordingly, when a user types the URL http://www.twobirds.com into a browser, Bird & Bird's web-site stored on a web-server with the IP address is located. Sub-pages for this site can then be located by reference to additional characters after the URL for the site's homepage. For example, articles relating to the internet can be found at http://www.twobirds.com/library/internet/index.8htm (the characters at the end of this URL, .8htm, constitute a fourth element of the URL which is the filename of the HTML file which contains the web-page). All such sub-pages are stored on the same web server and located in accordance with a system set up by the creator of the web-site.

Simple web-sites with only a few pages will not require very long URLs, and these can be created manually. However, web-sites with very large contents (such as, for example, on-line newspapers) require more sophisticated indexing systems. In such circumstances, an indexing server is used to create the URL for each page of the site which can then be indexed and stored in a searchable data base.

3.2.2 Subsistence of copyright in a URL

As indicated above, the element of a URL for a web-page after the domain name sets out a directory listing where that page can be located on the relevant DNS server. When considered as a whole, every URL is unique. However, this originality is not necessarily reflected by a URL being protected by copyright and, in this regard, there are a number of issues involved in considering whether copyright can subsist in a URL:

1. Is the URL a literary work capable of being protected by copyright? Works which lack sufficient substance are incapable of attracting copyright protection regardless of whether they are unique. By way of example, copyright will not generally subsist in a book title, an advertising slogan or the business name of a trading company, regardless of the amount of resources were expended in its creation7.

In many cases, it is obvious that a URL is too short to constitute a copyright work. For example, it is unlikely that copyright could subsist in either the URL http://www.twobirds.com or http://www.twobirds.com/library/internet/ index.8htm. However, the URLs of sub-pages on web-sites which contain a large amount of information (such as on-line newspapers) are often a significant length (in excess of 100 characters) and, in such circumstances, it is at least conceivable that a court would consider that such URLs have sufficient substance to enjoy copyright protection.

2. Has the URL taken skill and effort to create? It is a prerequisite to copyright protection in any literary work that at least a modicum of skill and effort have been expended in its creation. The structure of simple web-sites will usually be rather basic and, as such, the element of the URL setting out the directory structure can be created without any real skill or effort.

However, the directory structure of larger sites can be extremely sophisticated in order to ensure that the web-site is searchable in accordance with a number of criteria. In such circumstances, the URL of each individual web-page may well take considerable effort to create. Such URLs are more normally created automatically by the index server in which case the human effort involved in their creation may be minimal. Nevertheless, it is submitted that a court should consider the skill and effort test to be an objective one, and base its decision on whether the URL would have taken skill and effort to create had the author been human8.

3. Is the URL merely functional/factual? In the Ticketmaster case, the judge rejected the plaintiff's argument that copyright subsisted in the URL on the basis that it was merely functional and factual. In this regard, it appears that the judge considered that URLs were similar to conventional addresses in which copyright does not subsist.

However, the directory element of a URL differs fundamentally from a conventional addresses in that it sets out the location of the relevant web-page within a web-site structure created by the web-site developer, rather than in a pre-existing environment. Conventional addresses are by and large unoriginal and take little if any skill and effort to create. In contrast, the URL for a web-page within a web-site which is structured in a unique way, may well satisfy both these criteria.

4. Is the URL merely a meaningless set of characters? Many long-form URLs consist of a seemingly random series of numbers, letters and symbols which are incomprehensible to humans. However, this does not in itself prevent copyright subsisting in that work9, although, in order to attract copyright protection, the work must be created in a systematic rather than a random manner, otherwise it would fail the "skill and effort" test.

3.2.3 Ownership of copyright

As indicated above, URLs are often generated automatically by the index server. In such circumstances, the owner of any copyright which subsists in the URLs for the site's pages would be "the person by whom the arrangements necessary for the creation of the work are undertaken"10.

This person is likely to be either the designer of the web-site and/or the person who is responsible for uploading the relevant web-page onto the web-server (which triggers the index server to create the URL). These person(s) may well be part of an independent web-site development or hosting company, in which case an assignment would be necessary in order to vest ownership of any copyright in the URL to the company's whose web-site it is.

3.2.4 Infringement of copyright

A hypertext link from one web page to another can appear in many different forms - for example, a headline from the target site (as with the Shetland Times case), plain text, or a photograph or other pictorial depiction. In all cases, however, the URL of the target site must be stored on the server of the linking site. The user then clicks on the relevant icon on the linking site which triggers the URL for the target site to be inserted into the user's browser which then downloads the target site onto the user's computer.

In each of these instances, the URL for the target site must be stored on the server of the linking site. By section 23(2) of the Copyright Ordinance (section 17(2) of the Copyright Designs and Patents Act 1988) such storage in an electronic form is deemed to be a copying and, accordingly, it is an infringing act11.

4. Potential defences

In the event that proceedings for copyright infringement are issued against the operator of a web-site which has set up a deep-link into a target site, there are a number of defences which may be put forward in addition to contending that no infringing act has occurred.

4.1 That the link is licensed

The internet is by its very nature a vast web of sites which are linked together. Accordingly, there is a strong argument that any person who uploads a site on to the internet thereby grants an implied licence for other parties to create hypertext links into the site.

However, it is generally a straightforward process to exclude any such implied licence by way of an express statement to the contrary. In this regard, the courts in the U.K. and Hong Kong will generally accept that a disclaimer by notice is effective12 in the event that the party relying on the disclaimer has done what was reasonably sufficient to give the other party notice of its terms13.

It is submitted that it would generally be sufficient for the operator of a web-site to include on its home-page a prominent reference to terms and conditions of use which may be accessed by means of a hypertext link to a sub-page of the site (this may be done by way of a hypertext link placed in a prominent position at the top of the home-page). The operator of a linking site would generally visit the home-page of the target site at which time he would be put on notice of the terms and conditions which may include a prohibition on deep-links into the site.

However, there would remain a chance that the courts would not accept this as being sufficient, particularly if the deep-links are set up automatically. The courts may also take the view that by uploading a web-site onto the internet, the operator of the site grants a licence to access the site from which it is not entitled to derogate14. In the circumstances, the only means of making absolutely certain that the disclaimer is effective15 would be to require users of the site to acknowledge that they are aware of and accept its terms as part of a contractual relationship between the web-site operator and the user. This would typically be done by way of an entry page which requires users to click on an "I Agree" icon. However, since such a page would deter some users from entering the site, this option is unpopular with operators.

Absence any defence that a web-site operator has granted a "licence to access" from which it is not entitled to derogate, there is one other, very simple way in which an implied copyright licence may be revoked, namely to send express notice of revocation to the relevant party. Accordingly, once the operator of a site becomes aware of an authorised link into its site, a notice can be sent to the operator of the linking site expressly revoking any implied licence which may have been granted.

4.2 That the link is fair dealing

4.2.1 News reporting

Fair dealing in a copyright work for the purpose of reporting current events does not infringe copyright in the work, provided that it is accompanied by a sufficient acknowledgement (although no acknowledgement is required in the case of the reporting of a current event by means of a sound recording, film, broadcast or cable programme)16. As indicated at section 2.1 above, a web-page is arguably a cable programme for the purposes of relevant legislation and, if this is correct, no acknowledgement would be required.

In the case of a link into a web-site which sets out current events (for example an on-line newspaper), the operator of the linking site may seek to rely on the defence of fair dealing for the purpose of reporting current events (although, interestingly, the defender in the Shetland Times case did not seek to rely on this defence). However, this defence does have limitations, as follows:

(a) the defence is limited to events which are current. Accordingly, the operator of a site which links into the site of an on-line newspaper would only be able to rely on this defence whilst the story which is being linked to is current.

(b) the defence is limited to "fair" dealing. The question of whether a particular hypertext link is fair would depend on the nature of the allegation of copyright infringement, i.e. whether:

(i) it is alleged that material on the linking site has been copied (as with the case of Shetland Times Ltd v Wills - see section 2.1 above);

(ii) it is alleged that the operator has made material on the linking site available over the internet - see section 3.1 above; or

(iii) it is alleged that the URL of the relevant page has been copied - see section 3.2 above.

Judicial guidance on what constitutes "fair" for the purposes of this defence can be obtained from the judgment of Lord Denning in Hubbard v Vosper17. However, without going into a full analysis of that case, it appears that a defence of fair dealing is unlikely to be available in respect of (i) and (ii) above when the link takes the user to the whole of a news story set out on the target site. The question of whether the defence is available in respect of (iii) is more difficult and it will depend on whether the court considers that a deep-link which cuts out the home-page of the target site is fair. As can be seen from the cases set out above, the current trend worldwide is towards allowing such links and, accordingly, such a defence may well be available in respect of links to current news stories.

4.2.2 Criticism or review

In the event that a link is set up as part of a review of material on the linking site or any other material then the link will not constitute an infringement of copyright, provided that it is accompanied by a sufficient acknowledgement (although, as with fair dealing for the purpose of reporting a current event, no acknowledgement is required in the case of the reporting of a current event by means of a cable programme)18.

It should be noted that it is not sufficient for the linking site merely make to reference to the work in question and then set up a link to the target site, as this will not constitute a criticism or review for the purposes of this defence. It should also be noted that, as with the defence of fair dealing for the purpose of reporting a current event, the dealing must be fair (see section 4.2.1 above).

5. Summary

Although the Shetland Times case initially gave some indication that the courts would be willing to use copyright laws to restrict the use of deep-links, the recent decisions in Ticketmaster Corp. v Tickets.com and PCM v Kranten.com appear to show an international trend in the opposite direction. These cases are, however, unlikely to be the final word on the subject and we shall have to wait to see whether copyright laws can be applied in any more original ways to allow operators of web-sites which create their own content to protect themselves against deep-linkers.

Ultimately this issue may be settled through legislation, in the event that Article 6 of the WIPO Copyright Treaty 1996 is updated to take account of developments in the use of the internet. Any such amendment would have to balance the need to reward creators of content for their efforts with the desirability of avoiding over-regulating the internet. However, with no such amendment on the horizon we can expect to see continuing skirmishes both on a legal and technical level between web-site portals and the sites they link to. This is clearly not an ideal scenario, but with internet operators vying to be one step ahead of each other it is almost inevitable.

1 [1997] FSR 604
2 Equivalent to section 9 of Hong Kong's Copyright Ordinance (chapter 528)
3 Copyright, Designs and Patents Act, s.7(1); Copyright Ordinance, s.9(1)
4 U.S. District Court, Central District of California, 10 August 2000
5 Rotterdam District Court, 22 August 2000
6 The notes to the draft wording for Article 6 of the WIPO Copyright Treaty 1996 state that the relevant act which is sought to prohibit is the act of making available a copyright work by "providing access to it". In this regard, the notes indicate that it is the initial act of making the work available which is prohibited, although the notes do not consider the question of whether a hyperlink to a work which is already available would also be prohibited.
7 See, for example, Exxon Corpn. v Exxon Insurance Consultants International Ltd. [1982] RPC 69, CA in which the court held that copyright did not subsist in the business name "EXXON".
8 See The Modern Law of Copyright and Designs by Laddie, Prescott & Vitoria 2nd Ed. - Para 20.63
9 See DP Anderson & Co. Ltd. v Lieber Code Co [1917] 2 KB 469 where copyright was held to subsist in a series of codes which weremeaningless in any known language. A comparison may also be drawn with computer programs, the source code of which is incomprehensible to most humans. Original computer programs are now deemed to be protected as literary works (see Copyright (Computer Software) Amendment Act 1985 and section 4 of the Copyright Ordinance). However, even before such legislation came into force computer programs were considered to be protected by copyright - see Copinger and Skone James on Copyright -14th Ed. - para. 3-19.
10 Copyright Ordinance, s.11; Copyright, Designs and Patents Act 1988, s.9
11 Section 65 of the Copyright Ordinance provides for an exception to the application of section 23 in cases where the making of a transient and incidental copy is technically required for the viewing or listening of the work by a member of the public. However, the storage of a URL on a web-server in order to form a hypertext link is not transient and nor is it required for the target page to be viewed by the user and, accordingly, the operator of the linking site would not be able to take advantage of this provision.
12 Subject to the Control of Exemption Clauses in Hong Kong and the Unfair Contract Terms Act in the UK
13 See Chitty on Contract Vol. 1, 29th Ed, Paras. 12-013 and-12-014
14 See British Leyland Motor Corporation v Armstrong Patents Ltd. [1986] 1 AC 578 in which the House of Lords held that the manufacturer of cars (the parts of which were protected by copyright in their underlying drawings) automatically granted a licence to owners of the cars to repair them and that the manufacturer was not entitled to restrict this right by relying on copyright. Similarly, the court may consider that uploading of a web-site may carry with it a "right of access". However, the Privy Council in Canon Kabushiki Kaisha v Green Cartridge Company [1997] 3 WLR 13 made it clear that scope of any such defence would readily be expanded. In the circumstances, any such defence may be limited in application to the right to access the home-page of a web-site and it may have no application in relation to the sub-pages of a site, particularly if access to such pages is generally restricted.
15 Also subject to the Control of Exemption Clauses in Hong Kong and the Unfair Contract Terms Act in the UK
16 Copyright Ordinance, s.39(2)&(3); Copyright, Designs and Patents Act 1988, s.30(2)&(3)
17 [1972] 2 Q.B. 84
18 Copyright Ordinance, s.39(1)&(3); Copyright, Designs and Patents Act 1988, s.30(1)&(3)

First published in Copyright World in October 2000.