Determining the information content means interpreting what comprises the state of the art. The boards have established certain principles to be observed in this process.
According to the boards' settled case law, the technical disclosure in a prior art document must be considered as a whole (see T 56/87, OJ 1990, 188). The individual sections of a document cannot be considered in isolation from the others but must be seen in their overall context (see T 312/94, T 452/05, T 456/10). In T 1321/04 the board found that the terms used in patent documents should be given their normal meaning in the relevant art, unless the description gave the terms a special meaning. The patent document might be its own dictionary (T 523/00, T 311/93). Thus, if a special meaning could be derived from the patent document, only this meaning was ultimately decisive.
In T 312/94, the board held that it was a general legal rule for the interpretation of any document, in particular a patent application or patent, in order to determine its true meaning and thus its content and disclosure, that no part of such a document should be construed in isolation from the remainder of the document: on the contrary, each part of such a document had to be construed in the context of the contents of the document as a whole. Thus, even though a part of a document appeared to have a particular meaning when interpreted literally and in isolation from the remainder of the document, the true meaning of that part of the document could be different having regard to the remainder of the document (see also T 546/07; T 860/06).
According to established case law, it is a prerequisite for the acceptance of lack of novelty that the claimed subject-matter is "directly and unambiguously derivable from the prior art". In other words, it has to be "beyond doubt - not merely probable - that the claimed subject-matter was directly and unambiguously disclosed in a patent document" (see also T 1029/96). This also means that the novelty of the claimed subject-matter over a prior art document cannot be opposed only on the basis of equivalents to the features directly and unambiguously disclosed in the cited document (see also T 167/84, OJ 1987, 369; and T 928/93) (cf. T 1387/06).
In T 988/95 and T 218/00, the board held that a prior art document could therefore be considered noveltydestroying only if the claimed subjectmatter could be inferred directly and unambiguously from its teaching (see T 511/92) and if that teaching was likewise set out clearly and unmistakably (see T 450/89 and T 677/91).
In T 969/92 the board decided that, in order to determine what had been made available to the public, not only the main claim but also the remainder of a patent document had to be carefully considered for guidance as to what had really been taught in the prior document, i.e. its real express and implicit information content.
According to decision T 158/96, the information in a citation that a medicament was undergoing a clinical phase evaluation for a specific therapeutic application was not prejudicial to the novelty of a claim directed to the same therapeutic application of the same medicament, if such information was plausibly contradicted by the circumstances, and if the content of said citation did not allow any conclusion to be drawn with regard to the actual existence of a therapeutic effect or any pharmacological effect which directly and unambiguously underlay the claimed therapeutic application (see also T 385/07, T 715/03, T 1859/08).
In T 1859/08 the board first observed that independent claims 1 and 10 were in the form of medical use claims, the novelty of which was derived from the intended medical use. As a consequence, all the technical features of the therapeutic indication specified in the claims had to be taken into account when considering whether or not the claimed subjectmatter was novel. The board noted that prior art document D1 did not involve humans, whereas the claims before the board were directed to the treatment of breast cancer in a human patient. Further, D1 stated that the combination therapy disclosed in the study was "currently being explored". The board held that a mere statement that a combination therapy was being explored did not amount to a novelty-destroying disclosure. The "currently being explored" situation, where no clinical benefit was disclosed, fell within the rationale of decisions T 158/96 and T 715/03. According to these decisions, if a prior art document disclosed clinical investigations such as phase I, II or III studies (or stated that these investigations were ongoing), but failed to disclose the final result of these studies, it was not novelty-destroying. The board concluded that there was no description in the prior art documents of the treatment of a human patient, nor any disclosure of the biological effect. Therefore, the claims satisfied the requirements of Art. 54 EPC.
In T 943/93 the board held that a hypothetical possibility of operating within the claimed area was not per se sufficient in law to deprive this area of novelty, particularly if the skilled person had no technical motive and thus no practical necessity to work within this region (see also T 892/05).
Decisive for novelty in T 464/94 was a citation disclosing a preliminary test to transform plant protoplasts with selective markers. The opposition division had considered it probable that this document anticipated the patent in dispute. In the board's view, it was not justifiable to decide whether a document was prejudicial to novelty on the basis of probability. When a patent was revoked for lack of novelty, the department concerned had to be sure, having taken all the facts and arguments put forward during the proceedings into consideration, that the revocation was justified. If in doubt, further evidence had to be adduced otherwise the patent could not be revoked for lack of novelty (see also T 892/05).
In T 233/90 the board took the view that in a case where a document comprised in the state of the art under Art. 54(3) EPC 1973 referred to "a usual manner" of preparing a product, it was permissible to use documents of reference such as handbooks, encyclopaedias or dictionaries in order to determine what the skilled person would have understood by such a reference on the effective date of the prior document.
In T 410/99 the board stated that according to the boards' established case law, a prior art disclosure was noveltydestroying if it disclosed directly and unambiguously the subject-matter in question, account also being taken of a skilled person's common general knowledge at the publication date of the cited document in the case of prior art cited under Art. 54(2) EPC, or at the priority date of the cited document in the case of an Art. 54(3) EPC document (see e.g. T 511/92, T 412/91).
In T 412/91 the board decided that a prior art disclosure had to be read giving the information it contained the meaning that a skilled person would have given it at its publication date and disregarding information which would be understood by a skilled person to be wrong; however, any teaching which would not be recognised as wrong by a skilled person had to be accepted as state of the art.
In T 4/00 the board held that the points of a graph in a diagram in a prior art document did not represent a disclosure of the corresponding values read from the scales of the diagram, if the accuracy of the graph in the diagram could not be established.
In T 546/07 the board found that the skilled person reading a document did not construe individual points of the disclosure in isolation, but rather in the context of the document as a whole (T 312/94, T 860/06).
In T 1849/08 the board stated that what was decisive for the information content of a prior art document was what a person skilled in the art reading the document would understand from it. If a skilled reader of a prior art document containing an isolated statement that a diagram in a figure was "not to scale" could find nothing in the figure itself or in the rest of the document supporting this statement, he would consider the statement to be made in error and disregard it.