T 1134/06 (Internet citations) 16-01-2007

Reasons for the Decision

1. The appeal complies with Articles 106 to 108 and Rule 64 EPC and is therefore admissible.

2. Determining the prior art: standard of proof

2.1 Following accepted practice before the Boards of Appeal (see the decisions cited in Case Law of the Boards of Appeal of the EPO, 4th edition, 2001, section VI.J.5), the standard of proof to be applied in establishing the facts on which a decision is to be based - whether in inter partes or ex parte proceedings - is that of "balance of probabilities". However a much stricter standard of proof is applied in respect of a prior use or a prior oral disclosure which have to be proved beyond any reasonable doubt (see in particular decision T 472/92 (OJ EPO 1998,161) considering an alleged prior use).

2.2 Furthermore, as explained in T 750/94 (OJ EPO, 1998, 32), point 8 of the reasons, the principle of free evaluation of evidence applies in examining and evaluating the available evidence when considering an issue of fact. "Each item of evidence ... must be given an appropriate weight", corresponding to its likelihood of being true which is evaluated in accordance with all the relevant surrounding circumstances, including whether [it] is corroborated by other independent evidence". When an issue of fact concerning a past event is to be resolved, the available supporting evidence may include contemporaneous records, or it may consist of personal statements. In each case, "a first question which must always be considered is the reliability of the source of evidence, and thus its probative value". In summary, the probability that an allegation of a fact is true, and therefore the weight that evidence in its support is to be given when deciding whether the fact is true, is closely related to the reliability of the source from which the evidence supporting the alleged fact stems. As a first step therefore the reliability of the source must be ascertained.

2.3 When establishing the state of the art under Article 54(2) EPC for the purpose of assessing novelty and inventive step on the basis of traditional publications issues of fact (other than pertaining to what is disclosed) are normally not explicitly dealt with. This is because the majority of traditional publications such as published patent applications or specifications emanate from sources which are perceived as independently verifiable, hence considered as reliable. The evidentiary weight of such sources normally resides in the large number of physical, existing copies of the publication, or because the source itself has the character of public authority, such as a patent office. Therefore, publication date and content are assumed to be truthful with a high level of confidence. Anyone who challenges these assumptions in a particular case carries a heavy burden of proof. Thus, in this situation, the balance of probabilities is squarely in favour of such publications.

2.4 However, where the state of the art is established on the basis of information made available to the public by means of oral description, by use, or in any other way as also considered in Article 54(2) EPC, the assumptions that held true for a traditional publication are no longer necessarily correct, and further evidence regarding the circumstances of disclosure will be necessary. The Board holds that absent directly applicable rules or guidelines, the criteria to be applied for establishing a disclosure made available to the public through the Internet as in the present case should be the same as those introduced by the jurisprudence of the Boards of Appeal for establishing a prior use or a prior oral disclosure. To establish prior use or a prior oral disclosure the questions "When", "What" and "Under which circumstances" must be answered (see in particular T 328/87 (OJ EPO 1992, 701) and T 93/89 (OJ EPO 1992, 718)). Accordingly, in considering a disclosure made available to the public "in any other way", especially a disclosure on the Internet, these same questions must be answered as with prior use or oral disclosure. These questions are to be decided using the same strict standard of proof in respect of a prior use or a prior oral disclosure. Considering further T 750/94 mentioned above, the Board is of the opinion, that the circumstances which allow the determination whether a disclosure on the Internet was made available to the public may, where appropriate, also include factors that have a bearing on the reliability of the information. They may relate to, for example, the manner in which information has been procured and date stamped, how it has been preserved by the source, and whether it has remained unaltered since deposit.

3. The Internet as Source for Prior Art

3.1 In the case at hand the prior art document D3, as is indicated in the search report, has been retrieved from the Internet. The decision, point 1.) identifies the particular resource as the "Internet Archive", and the document has been retrieved using the "Internet Archive Wayback Machine".

3.2 The Board notes that the Internet, which is a global network of interconnected networks exchanging data via a common protocol, is a relatively new medium, which has seen phenomenal growth over the past decade or so. Its most popular use is via the World Wide Web which provides easy access to a vast, ever growing amount of information stored thereon to a very large number of Internet users. Internet and web are increasingly being used by industry for advertising and disseminating technical information, and have become a popular resource of technical information. However, there are recognized reliability and security issues concerning the Internet and the web and information retrieved from them. This is mainly due to the inherently transient nature of such media and of the data stored thereon. Internet and web are in continuous flux with data easily added, modified and removed. Web sites are updated and come and go with breathtaking speed. Linking between sites changes continuously. Internet and web are moreover very democratic media, are relatively unregulated and allow unrestricted access in large areas. They are notoriously insecure against unauthorized access. It is thus at the present state of affairs often very difficult to establish with a high degree of reliability what exactly appeared on a web site and when. Often the only certainty is the real-time certainty of the availability and content of a website in the particular moment it is viewed.

Recently, attempts have been made to preserve web data in the form of web archives, but these are still very much in their infancy. The Internet Archive (www.archive.org), is a US-based non-profit, private archiving initiative that provides an open resource storing past "images" of websites, which are retrievable using its search engine, the Internet Archive Wayback Machine. From available information (see e.g. http://www.archive.org/about/faqs.php#The_Wayback_Machine and an article by B. Howell in the Journal of Internet Law, February 2006 at http://www.strozllc.com/docs/pdf/PROVING_WEB_HISTORY_Jrl_of_Internet_Law_Feb%202006.pdf) the following details emerge: Though the Archive itself is non-profit, it receives its data from a for-profit company, Alexa Internet, which uses proprietary web crawl technology to harvest web data for its business interests. This data it donates to the Archive after a six month delay. The web crawls are not comprehensive: only non-password protected sites deemed of interest are captured, while site owners may further block the crawler, or request exclusion or retroactive removal of pages from the archive. The "images" are not necessarily instantaneous snap shots of a website, but may be assembled in the course of successive crawls. Links may not be preserved, or, if intact, may connect to different material than at the time of capture. Moreover, the archiving format may not be true to the original format of the site. The archived site may not be displayed in the format of the original or with all its functionality intact. Security structures are also not clear: the archive uses remote shell communication to its storage servers, but the extent of access once connected is not clear. Though there is a registration portal, users can gain access anonymously. Finally, there are known proprietary issues of the information retrieved and the terms of use advises the user that they use any content at their own risk.

The Internet Archive is not an archive in the classical sense. The givens of established archives such as authenticity and integrity of archived material - how truthful is archived material to the original, and how inviolable is the material against modification - cannot be taken for granted. This affects the evidentiary value of material retrieved from this resource. The Internet Archive itself for a fee offers an authentication service in the form of an affidavit, but states (see http://www.archive.org/legal/): "Before asking the Internet Archive to authenticate your documents, we ask that you please seek judicial notice or simply ask your opposing party to stipulate to the documents' authenticity. Of course, the best source of such information is the party who posted the information on the URLs at issue, and the second-best source of such information is someone who actually accessed the historical versions of the URLs".

3.3 Recent considerations in case law of the evidentiary value of information retrieved from the Internet, and the Internet Archive in particular, may be of interest. However, such case law is sparse.

3.3.1 Cases before the EPO that have some bearing on the use of the Internet as source of disclosure are T 91/98 and T 373/03 (both unpublished).

In T 91/98 the Board ruled that a date mentioned in the heading of an entry, which was retrieved from the Lexis-Nexis database (an Internet resource) much later than the relevant dates in question, could not be equated with the distribution date and, in particular, need not be right. Affidavits and declarations were also not found to establish the date of availability with the required degree of certainty.

In T 373/03 neither the author date of a PBS document recovered from the Internet nor its date of creation as found embedded in the code were accepted by the Board as proving a date of availability before the relevant dates in question.

In both cases the information retrieved from the Internet was not excluded per se. However, both cases demonstrate the difficulties in determining the exact date of availability of such information.

Case law at national level, though in no way binding on this Board, is also of interest:

The Bundespatentgericht in case BPatG 17W (pat) 1/02 (see GRUR 2003 Heft 04, pp 323-325) confirmed in later BPatG 17W (pat) 47/00, ruled that the Internet was not a reliable source for determining the state of the art. This applied also to web archives such as the Internet Archive.

3.4 Of further interest in the present case is the approach adopted in the PCT International Search and Preliminary Examination Guidelines, March 2004 (PCT/GL/ISPE/l). The Board notes that these provisions are relevant only to the PCT procedure, and are thus in no way binding or decisive in the present case. Nevertheless, the approach adopted in this preliminary international search and examination procedure in accordance with a treaty to which the EPO is also a signatory does have some bearing on the question of how to deal with Internet citations.

Sections 11.13 to 11.20 specifically deal with disclosures retrieved from the Internet. Sections 11.14 and 11.15 are headed "Disclosures Made on web Sites of Trusted Publishers" and indicate that dates of publications retrieved from such sites should be taken at face value" while "the onus is on the applicant to prove otherwise". Sections 11.16 to 11.20, on the other hand, deal with disclosures made on "Web Sites of Unknown Reliability" and provide a number of strategies for revealing the publication dates of such less reliable web sites, specifically mentioning commercial Internet archiving databases, such as the Internet Archive Wayback Machine as one way of obtaining such information. According to section 11.18, the electronic document which establishes the publication date for the Internet disclosure and which must mention the URL of the web site in question and its date of publication, should be printed out and the print out then cited as a further L citation. Nevertheless, further section 11.23 headed "Difficulty in Establishing Date of a Document" in reference to the Internet Archive Wayback Machine allows for the rebuttal of evidence therefrom by "sound reasons" contrasting with the much stricter term "proof" used in section 11.14 for trusted web sites. These guidelines thus firstly recognize the issue of reliability inherent in the Internet. More particularly, they also appear to indicate a different perception of the evidentiary value of citations from the Internet Archive.

3.5 In this respect, the Board notes that the EPO Guidelines are silent on Internet citations. Section B XII.5 provides general guidance for the situation where there are doubts as to date of availability. In reference to the Guidelines B-X, 9.2(viii), it recommends that "additional documents providing evidence in the matters in doubt may be cited". Paragraph (b) of section B-X, 9.2(viii), which refers back to section XII,5, relates to the use of L-citations, i.e. "documents which establish the publication date of another document".

3.6 In summary, the Internet is an increasingly used and popular source of technical information, and will therefore be of increasing interest for the purposes of establishing the state of the art under Article 54(2) EPC. Its use is surrounded by serious concerns as to its reliability. Problems arise in particular when establishing the date of availability of an in Internet disclosure, which in search or examination will normally be retrieved from the web long after the relevant dates of an application. Even if an earlier date can be established, it is not certain how authentic the disclosure is, i.e. how faithful it may be to what was available at the date in question. This will depend on its history and the circumstances of its preservation.

4. Standard of Proof and Internet Disclosures

4.1 From the above discussion the Board draws the following conclusions: a disclosure on the Internet may be comprised within the state of the art as defined by Article 54(2) EPC. If an Internet disclosure is to be used as prior art a strict standard of proof should be adopted. Thus, the fact that an Internet disclosure is state of the art under Article 54(2) EPC should be proved "beyond any reasonable doubt". The particular facts and evidence required will depend on each individual case, but will normally have to meet the criteria established by the jurisprudence of the Boards of Appeal in respect of a prior use or a prior oral disclosure, i.e. answer the questions of when the Internet disclosure was made available to the public, what was made available and under which circumstances was it made available to the public. Concerning the latter question, it will in most cases be necessary to address the main concern of reliability surrounding the Internet, in particular so as to establish whether and in how far a retrieved disclosure is true to the disclosure appearing at that date.

4.2 In certain cases, where a web site belonging to a reputable or trusted publisher publishes online electronic versions of paper publications, content and date can be taken at face value, and the need for supporting evidence can be dispensed with. It can also be envisaged that if a web site operates under recognized regulations and standards, which allow date and content of information retrieved therefrom to be established with a high degree of certainty, further evidence may also not or no longer be required. Of course, it should be clear for both the examiner and the public whether an internet source is considered as "reputable" or "regulated". This again calls for clearly defined guidelines.

Where a disclosure has been retrieved from a resource such as the Internet Archive, further evidence concerning the history of the disclosure, whether and how it has been modified since the date it originally appeared on a web site will be necessary. This could be in the form of an authoritative statement from the archivist. Alternatively, an appropriate statement as to the content, either from the owner or author of the archived web site which included the disclosure may suffice.

5. Internet Disc1osure D3

5.1 As noted D3 is a printout of a composite set of web pages retrieved from the Internet on a date well after the filing date of the present application (the first page bears two dates recognizable as such: 06-05-2003 and 05/11/03, i.e. 6 or 5 November 2003). Evidence regarding its date of availability is provided in the form of two URLs and a date mentioned in the FAQ set of pages included in D3, on the basis of which the Examiner concludes that "the availability of D3 has been made sufficiently plausible". No evaluation of the reliability of the Internet Archive is provided, and the argumentation appears based on implicit assumptions that the date given by the Internet Archive as archiving date is correct, and that the content on that date is in exact correspondence to that subsequently retrieved and printed out at search.

5.2 That these assumptions are problematic is clear when accessing the two URLs (at the time of writing this decision).

5.2.1 The Examiner infers a date of availability of 20 June 2000 from the number appearing in the first URL (URL1): http://web.archive.org/web/20000620174023/http://www.boersenspiel.de which, according to the Internet Archive FAQ page is in format yyyymmddhhmmss giving date and time the webpage was crawled. The second URL (URL2): http://web.archive.org/web/*/http://www.boersenspiel.de is cited to indicate that the relevant web site http://www.boersenspiel.de/ was in existence as far back as 30 May 1997.

5.2.2 The first URL, URL1, which is barely legible at the top of the printout of the first set of pages making up D3 (the HELP and FAQ sets bear URLs which are the same apart from a number), links to a different home page with a different URL number. The HELP link on this page connects to the correct HELP page, but the FAQ link thereof produces an error message. Thus, D3 is no longer retrievable from the stated URL. The Appellant also noted that he has been unable to obtain definitive documents on the subject.

5.2.3 The second URL, URL2, leads to a webpage listing hyperlinked dates, each of which appears to connect to a respective archived version of the page as harvested on that particular date. The date of 6 June 2000, the assumed date of availability of D3, is not included in the list. Moreover, the majority of the entries from the earliest entry of 30 May 1997 up to the priority date are asterisked. This asterisk "*", according to the Internet Archive FAQ page, indicates that the archived web page is not identical to the immediately preceding archived version. The list thus shows that this webpage has changed continuously since its first entry. This conclusion calls into question the date of 8 December 1996 mentioned in the FAQ page of D3 as the date since which the Börsenspiel has been running: if the web pages have changed, has the game itself also changed since that date?

5.2.4 On the basis of the information provided by the Internet Archive it would thus have been necessary to investigate further what was disclosed in the archived webpage of earliest date. Though such further investigation might reveal that the main features in question were the same in any of the archived versions of the web pages predating the present application, this is by no means self-evident.

5.2.5 Proof of the above assumptions regarding reliability will depend on suitable further evidence which may take the form of an authoritative statement regarding the authenticity of information retrieved from the Internet Archive and of the date it was harvested. Alternatively, the web pages in this particular case themselves offer various opportunities for obtaining corroboratory evidence by directing an appropriate request to the company that originally published the Börsenspiel (VoCalweb) or the company that currently manages the game (F.A.Z. Electronic Media GmbH). Even absent direct provisions in the Guidelines, Article 114(1) EPC certainly empowers the Examining Division to complement the retrieved documents with further supporting evidence, in a manner analogous to the use of the L-citations (see point 3.6).

5.3 It is to be noted further that the archived web pages of D3 or those accessible via the second URL are introductory pages to the Börsenspiel that describe the various advantages and properties of the game. The parts of the site where the game is actually played are accessible only upon registration and were thus excluded from the crawl according to the information available from the Internet Archive FAQ page. It is thus not possible to access these past playing sites and play the Börsenspiel game as it was played at the corresponding archiving dates. In other words, the game itself, or rather versions of the game at the time of archiving, were not archived. In this respect the archived web pages constitute no more than circumstantial evidence of the existence of a game with the properties and advantages described in the web pages at their archiving date, but do not disclose the game itself.

In this instance further evidence to the effect that the game had in fact been played at the relevant dates in the past, e.g. a personal statement to this extent in particular from the past or present game publishers (or a member of the public who played the game at that date) could conclusively prove the existence of and the prior use of such a game.

6. The Board shares the opinion of the Examining Division that it is plausible that the web sites of www.boersenspiel.de as archived at www.archive.org relate to a game which was played at the archive dates prior to the valid dates of the present application, and which would thus belong to the state of the art under Article 54(2) EPC. However, though this may be likely, the reasonable doubts which are due to the inherently unreliable nature of the Internet and the Internet Archive must be removed before the past game can indeed be taken into consideration in the present case for assessing patentability.

7. The Board therefore decides to exercise its discretion under Article 111(1) EPC and to remit the case to the Examining Division for further examination in order that it may carry out a further investigation to obtain the necessary evidence outlined above. In as far as a reasonable effort to obtain such further evidence is unsuccessful, the examination should be carried out without considering D3 or the archived websites of www.boersenspiel.de as included in the state of the art.