Friday, 31 March 2017

T 545/08: a lecture: internet disclosures and balance of probabilties

Subject of this appeal is the date of public availability of an internet disclosure and whether the Examining Division should have allowed evidence with respect to the publication date of this disclosure into the proceedings. Although the Board describes somewhere in its decision that the Boards of Appeal of have a huge backlog, this Board spent quite some time on writing a nice "lesson" on determining the date of availability to the public of internet disclosures and how the balance of probabilities must be applied in those cases. Of course, the Board has to provide a well-reasoned decision - somehow I have the feeling that the lesson is also written for the Examining Division because it seems that the Board was not completely happy with their work.
Reasons for the Decision


2. The contested decision is based on the assumption that document D1, an Internet disclosure, belongs to the prior art. Documents presented by the appellant in the oral proceedings in order to show that there was no sufficient basis for this assumption were not admitted as late-filed. Consequently, no consent was given under Rule 86(3) EPC 1973 to amending the application in accordance with the main request, i.e. reverting to the originally filed claims, and the then auxiliary request was found to lack inventive step over document D1 as closest prior art.
3. The appellant maintains that the Examining Division committed a procedural violation by relying on a document without properly ascertaining its status as prior art and by not admitting the appellant's evidence to the contrary (see section XI above). In the following, the Board will therefore assess whether, in the light of the general principles concerning the standard of proof with respect to Internet disclosures (see points 4 to 11 below), the Examining Division acted correctly when raising objections based on document D1 (see points 12 to 19 below) and refusing to admit documents E1 to E3 in the oral proceedings (see points 20 to 27).
Internet disclosures and standard of proof
4. The boards of appeal have dealt with issues relating to the reliability of Internet disclosures on several occasions. The crux of the matter is adequately summarised in the following passage of the Guidelines for Examination in the EPO (November 2016) [in the following: "Guidelines"], G-IV, 7.5.1:
"The nature of the Internet can make it difficult to establish the actual date on which information was made available to the public: for instance, not all web pages mention when they were published. Also, websites are easily updated, yet most do not provide any archive of previously displayed material, nor do they display records which enable members of the public - including examiners - to establish precisely what was published and when."
5. In decision T 1134/06 of 16 January 2007 (point 4.1), the Board held that, in order to establish that an Internet disclosure formed part of the state of the art, the same strict standard of proof as for prior use or prior oral disclosure, i.e. "beyond any reasonable doubt", had to be applied (see also decision T 373/03 of 2 September 2005, point 3.3). Circumstances which allowed the determination of availability to the public might also include factors that had a bearing on the reliability of the information, e.g. the manner in which information had been procured and date-stamped, how it had been preserved by the source, and whether it had remained unaltered since deposit.
6. Although this decision was cited with approval in some other decisions (see T 1875/06 of 8 January 2008, point 9; T 19/05 of 15 February 2012, point 2.3.2), its conclusions did not remain uncontested. Decisions T 286/10 of 21 May 2014 (points 2 to 2.3) and T 2227/11 of 18 February 2016 (point 2) did not find any justification for the application of the strict standard of proof beyond any reasonable doubt. Rather, only a sufficient degree of probability leading to the boards' conviction that a fact has occurred was considered to be required. The yardstick of balance of probabilities was also applied in the context of Internet disclosures in decisions T 990/09 of 3 July 2012 (points 10 to 12) and T 2339/09 of 17 November 2011 (point 2).
7. It is no easy task to try to reconcile the different views expressed in the appeal case law about the proper standard of proof relating to the issue of availability to the public, in particular with respect to Internet disclosures. However, any such attempt has to respect the European Patent Convention's overarching principle of free evaluation of evidence (see decision G 1/12, OJ EPO 2014, A114, point 31, with further references), which would be contradicted by laying down firm rules of evidence defining the extent to which certain types of evidence were, or were not, convincing. Instead, the question whether a fact can be regarded as proven has to be assessed on the basis of all the relevant evidence (G 3/97, OJ EPO 1999, 245, point 5).
8. While the above-described standards of proof clearly differ on a conceptual level, in most cases adhering to one or the other in judicial practice does not need to lead to divergent results if the standard of "balance of probabilities" is applied with some qualification. In particular, the latter should, in the Board's view, not be misunderstood as implying that a fact has to be regarded as already proven when it has a probability which is "just tipping the balance slightly", for instance, a probability of 51% (if it were at all possible to express probabilities in this context in mathematical terms).
9. Although the standard of balance of probabilities has occasionally been paraphrased as meaning that one has to assess whether one set of facts is more likely to be true than the other (see T 381/87, OJ EPO 1990, 213, point 4.4.b)), such a literal interpretation of "balance" does not, in the Board's view, properly reflect the long-standing judicial practice of the boards of appeal in determining whether a particular piece of information has become publicly available. It may not be possible to identify a single decision where a patent was refused or revoked on the basis of prior art, be it a written disclosure, an oral disclosure or a prior use, because its availability to the public had been found to be just slightly more likely than its non-availability. Even in the above-cited case T 381/87, which appears to have been seminal for the development of the "balance of probabilities" standard, the board found (see point 4.4 a)) that on the evidence - two letters from a librarian of a renowned institution written with full regard as to the potential serious nature of their contents - it was "clearly much more likely that document (A) was placed on the open shelves of the Library on 26 November 1981, than that it was not so placed" and that "[a]ny uncertainty [was] of a minimal nature".
10. There are numerous decisions which refer to and apply the standard of balance of probabilities, but interpret it in a manner which does not leave any doubt that just slightly tipping the balance of probabilities is not sufficient when assessing the public availability of prior art. An often-cited case is T 750/94 (OJ EPO 1998, 32), which in point 4 of its reasons stated the following:
"When an issue of fact is being examined and decided by the EPO on the balance of probabilities, the more serious the issue the more convincing must the evidence be to support it. If the decision upon an issue under examination may result in refusal or revocation of a European patent, for example in a case concerning alleged prior publication or prior use, this means that the available evidence must be very critically and strictly examined, for example in order to ascertain whether or not something happened (the alleged prior publication or prior use) before the relevant filing or priority date. [...] A European patent should not be refused or revoked unless the grounds for refusal or revocation are fully and properly proved".
Corresponding reasoning can be found e.g. in the more recent decision T 526/12 of 31 August 2015, point 2.2, assessing the public availability of an Internet disclosure.
11. Thus, the facts on which any finding of public availability is based must be established with a sufficient degree of certainty in order to convince the competent organ of the EPO in view of all the relevant evidence that they have indeed occurred. This holds true even if the determination is made on the basis of probabilities and not on the basis of absolute certainty ("beyond any reasonable doubt"). It is thus correctly stated in the Guidelines (G-IV, 7.5.2) with respect to Internet disclosures:
"The standard for assessing these circumstances is the balance of probabilities. According to this standard, it is not sufficient that the alleged fact (e.g. the publication date) is merely probable; the examining division must be convinced that it is correct."
The objection based on document D1 raised in the written proceedings
12. It is a general principle that, when raising objections, the burden of proof lies initially with the examining division. This means that objections must be reasoned and substantiated, and must show that, on the balance of probabilities, the objection is well-founded (see Guidelines, G-IV, 7.5.3). With respect to the publication date of a cited document, at least prima facie evidence is required. Prima facie evidence means evidence which is sufficient, on its own, to establish a fact or to raise a presumption of the truth of a fact unless controverted (see T 750/94, point 6; T 526/12, point 1.4). Thus, not every indication or hint qualifies as prima facie evidence.
13. If the objection is properly raised, it is then up to the applicant to prove otherwise or to at least submit evidence to displace the prima facie evidence. If the applicant successfully challenges prima facie evidence concerning a fact, e.g. the nominal publication date of a document, the burden of proof shifts back to the examining division to establish that the document was made available to the public (see e.g. T 929/94 of 7 July 1998, point 2.1).
14. In the present case, according to the international search report, document D1 was retrieved from the Internet on 15 December 2003, i.e. more than 18 months after the priority date of the application (24 May 2002). Neither the report nor the first communication pursuant to Article 96(2) EPC 1973, which cited document D1, contained any explanation as to why the document was considered to have been publicly available before the priority date.
15. In the context of Internet disclosures, such an absence of explanations is acceptable only where, in view of the circumstances of the case, it is self-evident - or at least easily understandable - why the document is considered to have become publicly available before a specific date. Those circumstances may include the nature, in particular the reliability, of the website from which the document was retrieved as well as intrinsic evidence deduced from information provided in the document itself.
16. Document D1 was retrieved from an Internet site with the address It consists of 13 pages. Its front page contains the title ("Oracle9i, ETL from the Database Out for Free"), a date ("January, 2002"), a firm logo with the text "innovative consulting IntelligentSolutionsforyourBusiness" and an address. The last page gives some information on the author (Daniel White), who is said to be a technology consultant with Innovative Consulting, a Pennsylvania-based professional services firm specialising in strategy consulting and implementation of high-performance business intelligence and data warehousing solutions. The last page also carries a copyright notice ("© 2001 Innovative Consulting Inc. ...)". There are no passages in the document indicating that, prior to its publication on the Internet, copies of it were made available to members of the public.
17. The only circumstantial evidence on which the International Searching Authority and the Examining Division apparently implicitly relied for its assumption that document D1 was made publicly available in January 2002 is the date on the cover page of the document. However, this date is not only imprecise, but also unqualified. In particular, the document does not specify the date as its date of publication. The indication of the date can therefore at best be considered as prima facie evidence that the document was created in January 2002, but not that it was made available to the public on that date (see also T 373/03, point 3.3).
18. Furthermore, the commercial website from which document D1 was retrieved cannot be considered, at least not without further investigations, as a source generally deemed to provide reliable publication dates, such as the websites of scientific publishers. Thus, irrespective of the standard of proof to be applied, i.e. "balance of probabilities" (see points 9 to 12 above) or "beyond any reasonable doubt", neither the information provided in document D1 itself nor its source of retrieval allowed, on their own, to conclude with a sufficient degree of certainty that the document was publicly available in January 2002. There was no prima facie evidence to rely on.
19. It follows that the Examining Division was not entitled to consider document D1 as prior art in its first substantive communication without providing further explanations and evidence as to the document's public availability before the priority date (see also T 1961/13 of 16 September 2014, point 5.1.6). Thus, the objection was not properly raised. It could not generate an obligation on the applicant's side to submit evidence against the assumed publication date in the written proceedings.
Refusal to admit documents E1 to E3 as evidence
20. While prior to the oral proceedings the appellant did not argue against the Examining Division's assumption that document D1 was prior art and, in its reply to the first communication, even amended the original claims in order to overcome the objections raised, it contested the assumption at the oral proceedings and submitted documents E1 to E3 (see below) as evidence to the contrary.
21. Document E1 contains a version of document D1 retrieved as a pdf file from the Internet site It also contains a printout of the document's properties in German, according to which it was created ("erstellt") on 15 January 2002 at 15:32 and modified on 15 January 2002 at 16:06. The printout names "L Eliott" as creator ("Verfasser"). The appellant argued that document E1 supported its view that January 2002 was only the creation date, and not the publication date, of D1.
22. Document E2 was said to be a printout of a search carried out by the applicant on the website of the Internet Archive Wayback Machine with respect to the URL of D1 and for the time period from 1 January 2002 to 31 December 2003. The search results show no record of the URL of D1 in 2002 and only one page in 2003. The appellant argued that, since the Wayback Machine had not archived document D1 during the period between its creation and the priority date of the application, this was evidence that the document had not been published in that period.
23. Document E3 was said to be a printout of various web pages of the website archived by the Wayback Machine. The appellant argued that they did not provide any hint that document D1 had been published on this website, so that D1 should not be regarded as prior art.
24. The Examining Division did not admit the new evidence pursuant to Rule 71a EPC 1973 for the reason that otherwise it could not come to a final conclusion without further investigations, so that the end of the procedure would be delayed. Moreover, prima facie none of the evidence submitted appeared to clearly establish that D1 had not been published before the priority date. The fact that the pdf file had been created on 15 January 2002 could instead be seen as a step towards the publication of the underlying Word document, because the pdf format was more suitable for publication. The fact that the Wayback Machine had not archived the pdf version of D1 under the URL used in the search report was not proof that the document had not been published earlier on the Internet using a different URL, because websites were frequently reorganised, or published in another form (e.g. as printed paper). Finally, no convincing conclusions could be drawn from the absence of a reference to D1 on printouts of some archived pages of the website The Examining Division further noted that D1 had been authored by a consulting company, which could be expected to disseminate D1 to its clients, i.e. to make it available to the public.
25. It is well established in the case law of the boards of appeal that Articles 113(1) and 117(1) EPC enshrine the basic procedural right to give evidence in appropriate form, specifically by the production of documents, and to have that evidence heard. Failure to consider evidence will normally constitute a substantial procedural violation in that it deprives a party of this basic right (see e.g. T 1110/03, OJ EPO 2005, 302, point 2.4, and T 2294/12 of 12 January 2016, point 1.1.3). Nevertheless, according to Article 114(2) EPC, facts and evidence which are not submitted in due time by the parties concerned may be disregarded. Rule 71a(1) EPC 1973 (which corresponds to Rule 116(1) EPC) stipulates that, when the EPO issues a summons for oral proceedings, a final date for making written submissions in preparation for the oral proceedings has to be fixed and that new facts and evidence presented after this final date need not be considered.
26. The discretionary power given by Rule 71a EPC 1973 (and Rule 116 EPC) is governed by Article 114(2) EPC. In ex parte proceedings, this discretion has to be exercised by considering all the relevant factors which arise in a particular case and by balancing the applicant's interest in obtaining proper patent protection for his invention against the EPO's interest in bringing the examination procedure to a speedy close by the issue of a decision (see T 755/96, OJ EPO 2000, 174).
27. When exercising its discretion in the present case, the Examining Division should have been aware that, although it had cited document D1 at the beginning of the examination procedure, it had never provided any further explanations and evidence as to the document's public availability before the priority date, so that the objections based on this document had not been properly raised before. Moreover, the arguments provided by the Examining Division for the exercise of its discretion are not persuasive. In particular, it is difficult to see why the fact that D1 was authored by a (consultant of a) consulting firm was taken as a hint for a specific publication date. Nor can this fact be a sufficient basis for the assumption that D1 was disseminated to clients in the relatively short period between the creation date of the pdf file (15 January 2002) and the priority date of the present invention.
28. Under these circumstances, the Examining Division was not allowed to reject the appellant's further evidence as late-filed and to continue to rely on document D1 as prior art. It thereby infringed the appellant's right to be heard (Article 113(1) EPC in conjunction with Article 117(1) EPC). This procedural deficiency was causal for the contested decision, as the Examining Division itself recognised that, when admitting the evidence, it could not have come to a final conclusion without further investigations.
Consequences of the procedural deficiencies

For these reasons it is decided that:
1. The decision under appeal is set aside.
2. The case is remitted to the department of first instance for further prosecution.
This decision T 545/08 (pdf) has European Case Law Identifier:  ECLI:EP:BA:2017:T054508.20170324. The file wrapper can be found here. Photo "Gold Scales of Justice" by Kittisak obtained via (no changes made).

No comments :

Post a Comment