c) Archives and internet publications 

In T 151/99 the board regarded it as highly plausible in the light of the available evidence that a master's thesis had become available to at least one member of the public. If the reference was in a document published before the priority date of the patent in suit, then it could be assumed that the thesis had also been made available to the public before the said date. This decision is analysed more recently in T 538/09 (master's thesis of the first examiner as prior art) and T 582/18 (master's thesis from Carnegie Mellon University in Pittsburgh stating June 2011 on its cover, the priority date being July 2012 – standard of proof: balance of probabilities). Note that T 538/09 also cites T 1134/06 and T 750/94 in that an important criterion in evaluating evidence is the reliability of the source: fulfilled for University of Utah.

In addition, note that the strict standard of proof (beyond reasonable doubt) laid down as the rule for internet publications in T 1134/06 no longer applies since T 286/10 and T 2227/11, both found the balance of probabilities to be the usually applicable standard. Confirmed, for example, by T 884/18, in which availability to the public was considered to be established on the balance of probabilities by the date stamps on evidence (screenshots of internet page) from the Wayback Machine, and T 3257/19.

In T 2165/18 the opposition division had found there was a lack of novelty over the undated user manual E1, which had been published online (www.janserwis.pl). The board observed that dates obtained from reputable and trusted editors or archives, which included www.archive.org in particular, were accepted by the boards as reliable. Janserwis, however, was not a known reputable editor or archive, so the board considered the probative value of a witness statement made by two of its company members to be insufficient. The opponent had also invoked the prior use of the mass-market software product the user manual was for, but this was unsuccessful as it did not answer the crucial question what version of the manual had been supplied then. Finding that a witness's testimony alone was not enough to fill in the gaps in the evidence as to when user manual E1 had been published and which software version it had accompanied, the board held that its date of public availability had not been established. This decision was subject to a petition for review particularly on the issue of testimony; the petition was rejected by R 15/21.

In T 1469/10 the board held that the publication dates indicated on the documents published by the ETSI 3GPP organisation, which was regarded as a reputable standardisation body having clear and reliable rules for their publications, were of a high probative value and thus might serve as prima facie evidence that a document was published on the date indicated, and regarded the documents at issue thus as representing the state of the art under Art. 54(2) EPC.

In ex parte case T 3257/19, concerning the publication date of document D11, the appellant seemed to argue that the examining division had failed to establish beyond reasonable doubt the exact earliest date of public availability. However, according to the established case law (see T 884/18), this was no longer the applicable standard of proof required for assessing whether or not documents downloaded from the internet or standard preparatory documents belonged to the state of the art. In view of the available evidence and arguments, the board was sufficiently convinced that D11 was indeed made available by file upload to a public 3GPP server before the Meeting at issue took place, i.e., at any rate well before the relevant date so that D11 was to be considered prior art under Art. 54(2) EPC.

In T 832/22 the question at issue was to establish the publication date of D9, a user manual published on the internet and issued by the opponent. The board stated that if the deciding body was convinced beyond a reasonable doubt that an alleged fact had occurred, there was no need to decide which standard of proof was applicable. This was the situation in the present case in which the board considered with reference to D39 (Digital forensics technical report) that D9 was part of the state of the art under Art. 54(2) EPC. Indeed, the board saw no reason to question the conclusions in the report D39 with respect to the upload date. The credibility of the report D39 and the methodology applied was as such not challenged by the respondent (patent proprietor) either. The fact that D39 had been commissioned by the appellant (opponent) and was based on digital evidence from the appellant had, in the absence of any inconsistencies or other counter-evidence, no negative impact on its credibility. The further evidence did not contradict the conclusions based on the publication date based on D39. Even if – for the sake of the argument – it had to be decided which of the contradicting publication dates in D39 or D9’, two independent pieces of evidence, was the correct one, both of them were well before the priority date.

As a matter of principle, if an adverse decision of an organ of the EPO is to rely on a certain fact that fact should be proven beyond any reasonable doubt. In ex parte T 826/03 the board applied this principle when attempting to establish the publication date of a Canadian patent application, and whether it was prior art. It found that in view of the contradictory evidence on file and in the absence of any further clarifying information from the Canadian Intellectual Property Office, it was virtually impossible for the board to establish with certainty that the public had access to the application according to document D3 before the priority date claimed by the application in suit. It therefore considered document D3 as not belonging to the prior art within the meaning of Art. 54(2) EPC.