Alpin, Tanya --- "When are compilations orginal? Telstra Corporation Ltd v Desktop Marketing Systems Pty Ltd" [2001] DigTechLawJl 1; (2001) 3(1) Digital Technology Law Journal 1

When are compilations original? Telstra Corporation Ltd v Desktop Marketing Systems Pty Ltd [1]

Tanya Aplin

Robinson College, Cambridge

Contents

• Introduction
• Facts
• Issues
• Decision
• Subsistence
• Infringement
• Comment
• Notes

Introduction

1. Telstra Corporation Ltd v Desktop Marketing Systems Pty Ltd [2] is a recent first instance decision of the Federal Court of Australia concerning the standard of originality required for compilations to attract copyright protection. Its significance lies in the fact that Finkelstein J refused to follow the approach of the U.S. Supreme Court in Feist Publications, Inc v Rural Telephone Service Co. Inc [3] ('Feist'). Instead, his honour held that sufficient labour or expense in gathering data, known alternatively as 'industrious collection' or 'sweat of the brow', could qualify a compilation as original.

   Facts

2. Telephone directories were at the centre of the dispute in this case. The applicant, Telstra Corporation Ltd ('Telstra'), together with its wholly owned subsidiary, Pacific Access, is in the business of producing, on an annual basis, white pages directories and yellow pages directories for each of fifty-five areas within Australia. The first respondent, Desktop Marketing Systems Pty Ltd ('Desktop') produced its own versions of white and yellow pages directories on CD-Rom. The second respondent was a director of Desktop. Telstra's objection to Desktop's business activities arose from the simple fact that Desktop's products used primary data extracted from its directories.

3. To appreciate the nature of Telstra's claim of copyright infringement against Desktop, it is useful to look in more detail at the nature and production of their respective works.

4. Each white pages directory produced by Telstra follows a standard format. The first segment of a directory contains 'information pages', which pages include useful general information. The second segment is the true 'directory' part of the publication, in which the name, address and telephone number of every subscriber in the area, barring those with a 'silent' number, is listed.

5. Every white pages directory is created by a combination of manpower and machine. The telephone number and address of every telephone subscriber is entered manually into a computer program. This program classifies entries into those that do or do not require manual editing and then transfers the information to a database. The database program verifies the listings and sorts them into alphabetical order. An entry that is rejected by the program is manually verified. The information is then transferred from the database to a graphics compilation and typesetting system. Editors proof every line of the proof pages for errors. Corrected proof pages are then sent to the publisher for printing, binding and publication.

6. The major difference in the format of a yellow pages directory is that businesses are grouped under headings that describe business activity. These headings are listed alphabetically and each business is listed under the appropriate heading in alphabetical order, together with its address and telephone number. There is also the option of purchasing an enhanced entry, namely, a listing formatted in such a way as to attract special attention. The sale of enhanced listings is what makes the production of directories (both white and yellow) lucrative for Telstra.

7. Compiling a yellow pages directory also requires the appropriate heading for a particular listing to be chosen. This is usually done by the subscriber, with occasional guidance from a Telstra salesperson. Headings are selected from unpublished heading books, which have been developed over many years and which are revised annually.

8. Desktop produced three CD-Rom directories, 'CD-Phone Directory', 'Phonedisk' and 'Marketing Pro'. CD-Phone Directory contained the name, address and telephone number of every telephone subscriber in Australia. Names were listed alphabetically within each postcode number and postcodes were ordered sequentially. A search by name could be conducted. Phonedisk contained the same data as CD-Phone Directory, but permitted the data to be searched by telephone number or postcode. The results of a search could be saved or printed.

9. Marketing Pro was Desktop's most sophisticated and commercially successful product. It consisted of eight different editions representing different geographical areas (one national edition, four state editions and three city additions). As well as grouping each business by type, businesses were also categorised by reference to the American Standard Industrial Classification Code and its Australian and New Zealand counterparts. Advanced searching capabilities were available. There were twenty-seven different ways to search the data, amongst the more important being by business name, business activity, industry classification and area. Search results could also be saved and printed.

9. The primary data for Desktop's products was extracted from the applicant's directories, however, it was not directly extracted by the respondents. Rather, a Sydney company arranged for every directory published by Telstra to be sent to the Philippines, whereupon a team of persons entered the data from these directories into a computer. The data was changed in various respects; for example, abbreviations were expanded, the name of the State was included and the data was formatted so that it could be read easily by a computer.

10. Desktop received this data in electronic form and loaded it onto its own computer system. It conducted limited verification of the data, however, it did insert additional information, namely, the appropriate industry code, number of employees, fax number and Ausdoc DX number. It used virtually identical headings to those used by Telstra in each of its three CD-Rom products.

11. Telstra brought an action against Desktop claiming infringement of certain directories published between 1996 and 1999 and in certain headings books produced between 1996 and 1998. Desktop denied that copyright subsisted in the directories or headings books or that it had infringed copyright. The second respondent was alleged to have authorised the infringement by Desktop, however, no finding was made on this issue.

    Issues

12. Telstra claimed that each directory and headings book was an original literary work, being a compilation, pursuant to sections 32 [4] and 10(1)[5] of the Copyright Act 1968 (Cth) ('Act'). The respondents conceded that each condition for subsistence of copyright, apart from the requirement of originality, had been satisfied. The respondents claimed that Telstra's products did not demonstrate sufficient skill in their selection or arrangement of data and thus were not original works. Telstra, however, contended that industrious collection of data could suffice to make its products original and that, in any event, they were the result of intellectual skill.

13. The respondents admitted copying from the applicant's works, however, they denied that the CD-Rom products were substantially similar to Telstra's directories and headings books.

    Decision

    Subsistence

14. Finkelstein J did not doubt that creative selection or arrangement of data would confer copyright on compilations. [6] The question was whether copyright could subsist in the presence of industrious collection of data, despite the absence of intellectual labour. In answering this question, Finkelstein J reviewed a series of English cases [7] on compilations decided prior to the Copyright Act 1911 (UK), which Act was adopted in Australia by virtue of the Copyright Act 1912 (Cth). These cases, he said, supported an 'industrious collection' approach to originality: "According to [them] copyright protection will be given as a reward for the author's investment of time and money, even if there be no creativity in the work." [8] The Copyright Act 1911 (UK) was the first statute to grant express protection to compilations as a form of original literary work. [9] In his honour's view, subsequent English decisions did not reject the pre-1911 approach to compilations. H Blacklock & Co Ltd v C Arthur Pearson Ltd [10] was an example of continued support for an 'industrious collection' test of originality and the cases in which compilations were denied copyright [11] did not undermine this test, since they were examples of insufficient skill and labour.

15. According to Finkelstein J, the effect of the English authorities was that copyright could subsist in compilations where sufficient intellectual effort was applied to the selection or arrangement of the facts or where the author has engaged in sufficient work or incurred sufficient expense in gathering the facts. The requisite amount of intellectual effort, labour or expense was a question of degree. Copyright protection did not extend to the facts themselves, so that a third person could publish the same facts, provided he collected the facts himself and did not copy them from another work. [12]

16. The respondents urged his honour to depart from the English authorities and to follow Feist. In Feist the US Supreme Court held that original compilations required independent creation by the author and a minimal level of creativity in the selection or arrangement of data. Finkelstein J discussed arguments of principle and policy in respect of a Feist approach, [13] but he felt bound by the pre-1911 English authorities, which had been incorporated into Australian law in 1912. Subsequent Australian legislation had not eschewed the English approach and Australian cases could not be explained solely by reference to creative selection or arrangement. Thus, his honour concluded, it was not possible "to jettison the old law and replace it with the principles expressed in Feist". [14]

17. Finkelstein J had little difficulty in concluding that each directory and headings book, when considered as a whole, [15] was original. Sufficient skill had been applied in producing the information pages of each directory and Lamb v Evans [16] supported the finding that the headings books were original. [17]

    Infringement

18. Telstra alleged that Desktop infringed copyright in their directories and headings books by reproducing a substantial part of these works. [18] Finkelstein J stated that in considering whether a substantial part of a work has been reproduced the originality of the part allegedly copied is an influential factor. This was clear from cases such as Ladbroke (Football) Ltd v William Hill (Football) Ltd [19] and Data Access Corporation v Powerflex Services Pty Ltd . [20] His honour added:

19. "Once it is accepted that 'industrious collection' will suffice to confer copyright, it follows that the directory portion of the white pages directories, the directory portion of, and the headings in, the yellow pages directories, and the headings in the headings books, are original and essential features of the respective works. This was not seriously in dispute." [21]

20. However, Finkelstein J indicated that, had he preferred the creativity approach to originality, he would have found that the telephone listings and headings in the applicant's works lacked any trace of creativity. [22] Presumably, if this had been the case, a real issue would have arisen over whether a substantial part of Telstra's works had been copied.

21. The respondents admitted copying from the Telstra's directories and headings books, however, they denied that their CD-Rom products were sufficiently similar to the applicant's products to constitute a reproduction. The respondents claimed that the inclusion of additional information and a different visual presentation of data made their products sufficiently dissimilar to preclude a finding of infringement. In terms of additional information, addresses were given in full (ie, street, suburb, state and postcode), facsimile numbers were sometimes included and the Marketing Pro product included the position of the business and its industry codes. In terms of the visual presentation of data, information did not appear within columns on a page, but as a separate 'record' of information. In each record, information was listed next to appropriate headings: eg, 'Company Name', 'Address', 'Telephone' and 'Facsimile'. Because of the record format, the relevant headings appeared with each business entry, as opposed to just once.

22. These differences did not persuade Finkelstein J of non-infringement. His honour explained that sufficient similarity did not relate solely to physical appearance of data, but to "whether the substance of the information that has been taken from Telstra's works...has been reproduced in the CD-Roms." [23] Finkelstein J had little trouble in finding that the "significant recorded facts...are the same, or substantially the same, as they appear in Telstra's works" and, thus, that the respondents had infringed copyright in those works. [24]

    Comment

23. Finkelstein J was correct in his view that prior to the Copyright Act 1911 (UK) courts protected compilations that were a result of 'labour and expense'. This is clear from the cases he relied upon, along with Jarrold v Houlston , [25] Collis v Cater, Stoffell and Fortt Ltd [26] and Walter v Steinkopff.[27] Courts were also prepared to protect compilations that resulted from intellectual skill. [28] The express reference to 'original' works in the 1911 and 1912 Acts did not mark the introduction of a higher standard of originality than was previously the case. [29]

24. However, Finkelstein J's conclusion that originality may be conferred either by creative selection or arrangement of data or by industrious collection of data oversimplifies the way in which courts have dealt with this issue since 1911. First, courts have avoided approaching the originality issue in this specific, alternative manner, but have tended to ask whether sufficient 'skill, judgement and labour' has been expended in producing the compilation. [30] In other words, courts have generally avoided protecting compilations solely on the basis of expenditure of effort and money. [31] At the same time, however, courts have protected compilations that have displayed marginal skill in their selection or arrangement, but which have involved considerable labour or expense. [32] Thus, the application of the 'skill and labour' test is a highly flexible one and it may operate as a pretext for protecting mainly investment. Second, courts have regularly looked beyond the physical process of presenting and ordering information when assessing whether a compilation is original. Instead, they have been prepared to consider skill and effort applied in gathering or calculating data for a compilation. [33] Thus, the effect of Anglo-Australian authorities may be more accurately stated as follows: compilations will attract copyright where there has been sufficient skill and labour expended in the collection, selection or arrangement of information [34] and the expenditure of skill and labour need not be proportionate.

25. In terms of whether a substantial part has been copied, it is correct to say that the originality of the part taken is highly relevant. [35] Where a compilation has attracted copyright due only to its creative selection or arrangement, failure to copy the collocation of material has resulted in non-infringement. [36] Where the compilation has been copied wholesale, courts have found infringement without having to isolate whether skill and labour in either collecting, selecting or arranging information has been appropriated. [37] Where persons have verified, rearranged or embellished data taken from earlier compilations this has nevertheless amounted to infringement. [38] These cases may be seen as examples of appropriation of substantial skill and labour expended in collecting information. The finding of copyright infringement in Telstra may be justified on the same basis.

26. The interesting question to emerge from Telstra is whether Australian law should follow Feist . This issue is likely to be at the heart of the respondents' appeal to the Full Federal Court.

27. Finkelstein J concerned himself with two main arguments of principle: is an 'industrious collection' test contrary to the nature of copyright and does it lead to the protection of raw data? After a lengthy digression into 18th and 19th century English copyright law, Finkelstein J observed that copyright has largely protected creative works of art and literature. Therefore, protection of compilations resulting from 'industrious collection' arguably conflicts with the nature of copyright. However, his honour commented that compilations necessarily involve a different type of creativity because they consist of gathering, selecting and arranging information and the English cases have recognised this different order of creativity. [39] One could also add that Australian copyright law has long protected subject matter that is the result of more labour and expense than skill, or at least the product of different (technical) skill. [40] Thus, the scope of protection has been of a neighbouring rights kind. Future debate might therefore consider whether investment-driven compilations should be recast in this vein. [41]

28. Another criticism of the 'industrious collection' approach to originality is that it undermines the fundamental principle that copyright law does not protect raw facts. [42] This criticism held sway with the Supreme Court in Feist . [43] However, it is important to recognise that courts have, on occasion, refused copyright protection where this would otherwise protect raw data. [44] Further, as has been emphasised by judges [45] (including Finkelstein J) and commentators, [46] third persons remain free to collect primary data. Of course, this statement does not hold true where the first compiler is also the sole source of data. Examples of sole source information providers are racing authorities who fix racing lists, [47] football authorities who compile fixture lists, [48] broadcasters who fix radio programmes [49] and telephone providers who compile lists of subscribers. [50] In these situations, there is legitimate concern about whether second comers are wrongly inhibited from creating value-added or derivative products. But there are also mechanisms by which such monopoly effects could be regulated, namely, compulsory licensing [51] and competition law.
    
    [52]

29. As a matter of policy, Finkelstein J highlighted two main objections to following Feist: the Feist standard of originality is an uncertain one and reduced protection for compilations will act as a disincentive to producing them.

30. The Supreme Court in Feist described the originality requirement for compilations as "a minimal level of creativity" in the selection or arrangement of data. [53] It further commented that the requisite level of creativity is "extremely low" and "the vast majority of works" will meet this test. [54] Finkelstein J remarked that the meaning of the Feist criterion is uncertain [55] and pointed to Key Publications, Inc v Chinatown Today Publishing Enterprises [56] and Bellsouth Advertising & Publishing Corp. v Donnelly Info Publishing Inc [57] as examples of the courts struggling to apply the standard consistently. [58] The uncertain application of the Feist standard does not necessarily sound against its adoption in Australia. Rather, it emphasises that courts would have to provide clearer guidance on its scope. For example, courts would have to clarify whether or not skill in generating, calculating or collecting information is to be entirely ignored.

31. A more important consequence of following Feist is that compilations will receive less protection. Finkelstein J commented that a 'creativity' approach to originality would exclude compilations in which functional or commercial considerations determined selection, and in which data was listed in obvious ways, such as alphabetically, chronologically, or sequentially. [59] Probably of greater concern for database producers, however, is the fact that only creative selection or arrangement of materials would be protected. Thus, second comers would be able to extract data from a compilation, provided they rearranged the data and did not copy a substantial part of the first compiler's selection. [60] In the other words, on a Feist approach, the respondents in Telstra would probably avoid infringement.

32. Anxiety over the impact of Feist-type protection on incentives to create compilations, whether entirely justified or not, is very likely to generate pressure for Australia to adopt sui generis protection for databases. The experience in the United States tells us so: one need only look to the number of bills that have entered Congress post-Feist . [61] In addition, a Feist approach could rekindle the efforts of the European Union to make its sui generis database scheme an international standard. [62] The likelihood of national and international pressure to introduce sui generis database protection is not a reason to reject Feist. What it does suggest, though, is that debate should be cast in wider terms to include the advantages and disadvantages of existing copyright protection of compilations as compared with possible neighbouring rights or sui generis protection of databases. [63] Until this wider debate gets seriously underway, it seems sensible to preserve the status quo and to decline to follow Feist , as Finkelstein J indeed did. Whether the Full Federal Court will share this view remains to be seen. [64]

    Notes

    [1] This case note was originally written for publication in the EIPR. The author would like to thank the editors of the EIPR for allowing re-publication of this note in the DTLJ.

    [2] [2001] FCA 61.

    [2] (Finkelstein J).

    [3] 499 US 240 (1991).

    [4] s32 states that copyright will subsist in unpublished and published original literary works.

    [5] s10(1) defines 'literary work' to include a table or compilation, expressed in words, figures or symbols.

    [6] Paras 44, 47.

    [7] Matthewson v Stockdale [1806] EngR 112; (1806) 12 Ves 270; 33 ER 103; Kelly v Morris (1866) 1 Eq 697; Morris v Ashbee (1868)

    [7] Eq 34; Walter v Lane [1900] AC 539.

    [8] Para 50.

    [9] See s1 and s35(1) of Copyright Act 1911 (UK).

    [10] [1915] 2 Ch 376.

    [11] Leslie v J Young & Sons [1894] AC 335; G A Cramp & Sons Ltd v Frank Smythson Ltd [1944] AC 329.

    [12] Para 64.

    [13] Paras 75, 83-84.

    [14] Para 85.

    [15] Adopting the approach in Ladbroke (Football) Ltd v William Hill (Football) Ltd [1964] 1 WLR 273.

    [16] [1892] 3 Ch 462 (Chitty J); [1893] 1 Ch 218 (CA).

    [17] Paras 90-91.

    [18] See s31(1) and s14 of Act.

    [19] [1964] 1 WLR 273.

    [20] [1999] HCA 49; (1999) 166 ALR 228.

    [21] Para 93 (emphasis supplied).

    [22] Paras 93, 102.

    [23] Para 109.

    [24] Ibid.

    [25] [1857] EngR 756; (1857) 3 K&J 707; 69 ER 1294.

    [26] (1898) 78 LT(NS) 613.

    [27] [1892] 3 Ch 489.

    [28] Hotten v Arthur [1863] EngR 803; (1863) 1 H&M 603; 71 ER 264.

    [29] Sands & McDougall Pty Ltd v Robinson [1917] HCA 14; (1917) 23 CLR 49; Express Newspapers plc v News (UK) Ltd (1990) 18 IPR 201.

    [30] Eg Macmillan & Co Ltd v Cooper (1923) 1B IPR 204; Ladbroke (Football) Ltd v William Hill (Football) Ltd [1964] 1 WLR 237; Sands & McDougall Pty Ltd v Robinson [1917] HCA 14; (1917) 23 CLR 49; Skybase Nominees Pty Ltd v Fortuity Pty Ltd (1996) 36 IPR 529.

    [31] Cf obiter dicta in Football League Ltd v Littlewoods Pools Ltd [1959] Ch 637, 656 (Upjohn J) and Australian Consolidated Press Ltd v Morgan [1965] HCA 21; (1964) 112 CLR 483, 487 (Barwick CJ).

    [32] Eg BBC v Wireless League Gazette Publishing Co [1926] Ch 433; Purefoy Engineering Co Ltd v Sykes Boxall & Co Ltd (1955) 72 RPC 89; Kalamazoo (Aust) Pty Ltd v Compact Business Systems Pty Ltd (1984) 84 FLR 101; T R Flanagan Smash Repairs Pty Ltd v Jones [2000] FCA 625; (2000) 48 IPR 19.

    [33] Eg Ladbroke (Football) Ltd v William Hill (Football) Ltd [1964] 1 WLR 237; Football League Ltd v Littlewoods Pools Ltd [1959] Ch 637; Winterbottom v Wintle [1947] WALawRp 21; (1947) 50 WALR 58; Mirror Newspapers Ltd v Queensland Newspapers Pty Ltd (1982) 59 FLR 71.

    [34] Autospin (Oil Seals) Ltd v Beehive Spinning (A Firm) [1995] RPC 683, 698.

    [35] Data Access Corporation v Powerflex Services Pty Ltd [1999] HCA 49; (1999) 166 ALR 228, 248; Cantor Fitzgerald International v Tradition (UK) Ltd [2000] RPC 95, 130-131.

    [36] Warwick Film Productions Ltd v Eisinger [1969] 1 Ch 508; John FairFax & Sons Pty Ltd v Australian Consolidated Press Ltd (1960) SR(NSW) 413.

    [37] Collis v Cater, Stoffell and Fortt Ltd (1898) 78 LT(NS) 613; Express Newspapers Plc v Liverpool Daily Post [1985] 1 WLR 1089; Winterbottom v Wintle [1947] WALawRp 21; (1947) 50 WALR 58; Mirror Newspapers Ltd v Queensland Newspapers Pty Ltd (1982) 59 FLR 71; Kalamazoo (Aust) Pty Ltd v Compact Business Systems Pty Ltd (1984) 84 FLR 101.

    [38] Morris v Ashbee (1868) 7 Eq 34; Waterlow v Rose (1989) 17 IPR 493; Elanco Products Ltd v Mandops Ltd [1979] FSR 46; Autocaps (Aust) Pty Ltd v Pro-Kit Pty Ltd [1999] 46 IPR 339.

    [39] Para 84.

    [40] Eg broadcasts, sound recordings and published editions.

    [41] This has already occurred to a limited extent: see Ricketson, 'Copyright and Data Bases' in Hughes, G (ed), Essays On Computer Law (Longman Professional, 1990), 67-98, 96-98 and Garrigues, 'Databases: A Subject-matter for Copyright or for a Neighbouring Rights Regime?' (1997) 1 EIPR 3.

    [42] Feist , 353.

    [43] 350-354, 359-360.

    [44] Smith's Newspapers Ltd v The Labour Daily [1925] NSWStRp 15; (1925) 25 SR(NSW) 593; Odham's Press Ltd v London and Provincial Sporting News Agency Ltd [1935] Ch 672; Victoria Park Racing & Recreation Grounds Co Ltd v Taylor [1937] HCA 45; (1937) 58 CLR 479.

    [45] Kelly v Morris (1866) 1 LR Eq 697, 701-702; Walter v Lane [1900] AC 539, 549 (Halsbury LC) and 551 (Lord Davey).

    [46] Eg Ginsburg, 'Creation and Commercial Value: Copyright Protection of Works of Information' (1990) 90 Colum L R 1865, 1922; Thakur, 'Database Protection in the European Union and the United States: The European Database Directive as an Optimuum Global Model?' (2001) 1 IPQ 100, 119-120.

    [47] Eg Mander v O'Brien [1934] SAStRp 12; (1934) SASR 87.

    [48] Eg Football League Ltd v Littlewoods [1959] Ch 637.

    [49] Eg, BBC v Wireless League Gazette Publishing Co [1926] 1 Ch 433.

    [50] Eg Feist .

    [51] Ginsburg (1990), 1924-1936.

    [52] Wyburn, 'Copyright, Databases & Misuse of Market Power' (1997) 15 Copyright Reporter 46; Magill TV Guide [1995] 4 CMLR 78.

    [53] 345, 348, 358.

    [54] 345.

    [55] The standard of 'author's own intellectual creation' contained in Article 3(1) of the EU Database Directive has also been criticised for its ambiguity: see Karnell, 'European Originality: A Copyright Chimera' in Kabel et al (eds), Intellectual Property and Information Law (Hague: Kluwer, 1998), 201-209.

    [56] 945 F. 2d (2d Cir. 1991).

    [57] [1993] USCA11 1579; 999 F. 2d 1436 (11th Cir. 1993).

    [58] Paras 75-78.

    [59] Para 74.

    [60] Ginsburg, 'No "Sweat"? Copyright and Other Protection of Works of Information After Feist v Rural Telephone' (1992) 92 Colum L R 338, 348-349.

    [61] Database Investment and Intellectual Property Antipiracy Bill 1996 (HR 3531); Collections of Information Antipiracy Bill 1998 (HR 2652); Collections of Information Antipiracy Act 1999 (HR 354); Consumer and Investor Access to Information Act 1999 (HR 1858). All have fail to be introduced.

    [62] The EU was a main proponent of a Database Treaty, along the lines of the sui generis provisions of its Databases Directive, at the 1996 WIPO Diplomatic Conference.

    [63] Eg see Ricketson (1990) and Davison, 'Sui Generis or Too Generous: Legislative Protection of Databases, Its Implications for Australia and Some Suggestions for Reform' [1998] UNSWLawJl 63; (1998) 21(3) UNSW LJ 729.

    [64] The Full Federal Court is listed to hear an appeal from Desktop in December, 2001.