T 1724/10 (Document indexing/DASSAULT) 27-01-2015
Download and more information:
Method for indexing and searching a collection of internet documents
Amendments - added subject-matter (no)
Inventive step - (yes)
Substantial procedural violation (yes) -appealed decision sufficiently reasoned (no)
Reimbursement of appeal fee - (yes)
I. The patent proprietor (appellant) appealed against the decision of the Opposition Division to revoke European patent no. 1 462 952.
II. The opposition had been based on Article 100(a) (lack of novelty and inventive step), Article 100(b) and Article 100(c) EPC.
III. In the contested decision the Opposition Division found, inter alia, that claim 1 of the patent did not have a valid basis in the application as filed (Articles 100(c) and 123(2) EPC), and that the subject-matter of claim 1 did not meet the requirements of Article 56 EPC because a technical problem solved by the method of claim 1 could not be identified. The same objections applied to claim 1 of the first and third auxiliary requests. The subject-matter of the second auxiliary request was found not to comply with Article 56 EPC and to extend the protection conferred by the patent.
According to the minutes of the oral proceedings before the Opposition Division, a further new auxiliary request filed by the patent proprietor was not admitted into the proceedings.
The Opposition Division relied, inter alia, on the following prior art:
D6: Harmandas V., Sanderson M., and Dunlop Mark D., "Image retrieval by hypertext links", ACM SIGIR Forum, Vol. 31, No. SI, ACM, 1997.
D9: McBryan, Oliver A. "GENVL and WWWW: Tools for Taming the Web", Proceedings of the First International World Wide Web Conference, May 1994.
D11: Kobayashi, M., & Takeda, K. (2000), "Information retrieval on the web", ACM Computing Surveys (CSUR), 32(2), 144-173.
IV. With the statement of grounds of appeal, the appellant filed, as a main request, a set of claims 1 to 8 corresponding to the claims of the granted patent and two new sets of claims 1 to 8 as first and second auxiliary requests. The appellant furthermore requested reimbursement of the appeal fee.
V. With a letter dated 15 March 2011, the opponent (respondent) submitted the following document:
D23: Willett, Peter, "Recent trends in hierarchic document clustering: a critical review",
Information Processing & Management 24.5, 1988, 577-597.
VI. In a communication dated 18 September 2014 accompanying the summons to oral proceedings, the Board expressed the preliminary view that claim 1 of the patent (main request) appeared to comply with Article 123(2) EPC. Furthermore, it seemed from the submissions of the parties that the inventive step of the claimed subject-matter was to be assessed essentially on the basis of documents D6, D11 and possibly D23, and of the skilled person's general knowledge.
VII. In reply to the Board's communication, the appellant requested with letter dated 26 December 2014 that document D23 be disregarded as late-filed and not relevant to the present case.
VIII. On 27 January 2015, oral proceedings were held as scheduled before the Board. At the end of the oral proceedings the Chairman announced the Board's decision.
IX. The appellant requested that the decision under appeal be set aside and that the patent be maintained as granted or, in the alternative, as amended on the basis of the claims of either the first or second auxiliary request filed with the statement of grounds of appeal. The appellant further requested that the appeal fee be reimbursed.
The respondent requested that the appeal be dismissed.
X. Claim 1 of the patent as granted (main request) reads as follows:
"A computer-implemented process for searching among a collection of Internet or HTML documents (30, 32, 34, 36, 38, 40, 42), the collection comprising referencing documents and referenced documents referenced in the referencing documents by use of links such as Universal Resource Locators, wherein said referencing and referenced documents contain index terms, the process comprising:
- for a referencing document of the collection:
- recognising documents referenced in said referencing document; and
- aggregating (70) said referencing document (32) with documents (34) it references to form an aggregate document, said aggregate document formed being a single logical document associated with the documents forming said aggregate document;
- indexing said aggregate document, based on index terms contained in the referencing and referenced documents forming said aggregate document, to form an index, wherein an entry in the index table is associated with the referencing and referenced documents forming the aggregate document;
- searching (74) among aggregate documents by operating on said index; and
- providing (74), as a result, an aggregate document."
The first and second auxiliary requests are not relevant to the Board's decision.
XI. The appellant's arguments may be summarised as follows:
According to the decision under appeal the feature "by use of links such as Universal Resource Locators" extended the subject-matter of claim 1 of the patent beyond the content of the original application. In particular, the Opposition Division considered that this feature conflicted with the preceding claim feature "the collection comprising referencing documents and referenced documents referenced in the referencing documents". However, the terms "link" and "reference" were indifferently used across the description of the patent and should thus be regarded as synonyms. This was also consistent with the skilled person's understanding of "link" and "reference" in the context of the invention.
Hence, as it was directly and unambiguously derivable from the patent that the terms "link" and "reference" were synonyms and that links could be in the form of URLs, the feature "by use of links such as Universal Resource Locators" did not infringe Article 123(2) EPC.
Document D6 related to a model for the retrieval of images from a large World Wide Web based collection and focused in particular on how to exploit the linked nature of the Web to provide access to images based on an existing model developed for image retrieval from other hypermedia collections. Thus, document D6 concentrated upon the retrieval of a non-textual node (i.e. an image) and excluded de facto the primary aim of the present invention which was improving the search in context among web pages irrespective of the type of search documents or files, as specified in paragraph [0012] of the contested patent.
Document D11, published in June 2000, was a survey of notable studies covering a variety of topics in information management and retrieval. In fact, D11 was a reference document in the technical field of search engines which the person skilled in the art wishing to improve a search engine would consider and select as a promising starting point.
Document D11 did not disclose the step of aggregating a referencing document with referenced documents it references recited in claim 1 of the patent in suit. Indeed, at most, this prior art taught that clustering (i.e. grouping of similar documents together) might be used in order to expedite information retrieval. Specifically, document D11 taught that clustering was just one of several ways of organising documents to facilitate retrieval. In fact, clustering aimed at identifying groups of database documents that could be retrieved and processed together for a given type of user input query. In other words, the clustering discussed in document D11 related to the processing, upon a user query, of documents already indexed in a database. It did not involve the steps of "aggregating" and "indexing" recited in claim 1.
Moreover, clustering and indexing were different search tools and there was no reason to combine the teaching of a document (D6) relating to indexing with the teaching of a document (D11) directed to clustering.
Document D23, submitted by the respondent in reply to the statement of grounds, should not be admitted into the proceedings as it was late-filed and not relevant to the present case. As indicated in section 4.2, page 589, document D23 used tree searching algorithms that were state of the art in 1988 and consisted in hierarchically clustering documents in a tree-like structure and then travelling in the tree for assessing at each step the similarities between the search terms and terms that characterised the current document or cluster of documents. The purpose of clustering was to make it possible to travel through documents according to the top-down and bottom-up strategies discussed in section 4.1 of document D23. Furthermore, the methods disclosed in D23 were not applicable to very large collections of documents and in fact they had been abandoned by the time the opposed patent was filed.
As the use of clusters in document D23 had nothing to do with the aggregating step recited in claim 1 of the contested patent, the assumption that clustering and aggregating were synonyms was misleading. Moreover, there was no suggestion in document D23 of aggregating documents so as to form a single logical document which was subsequently indexed and which formed a single entry in the index table. As a matter of fact, document D23 did not refer to index tables which had only later become state of the art.
In summary, the subject-matter of claim 1 was not made obvious by any of the cited documents D6, D11 and D23, or their combinations.
The opposition division stated for the first time during the oral proceedings that the technical problem solved by the invention was speculative and not acceptable as a valid objective technical problem. The fact that the patent proprietor was taken by surprise by this objection and was not given the possibility to file amended claims in response to the Opposition Division's new and surprising argument constituted a substantial procedural violation. Furthermore, the Opposition Division failed to correctly apply the problem-solution approach and, in particular, failed to consider the right of the patent proprietor to reformulate the objective technical problem in a less ambitious manner, as it had been suggested. This amounted to a violation of the right to be heard and thus constituted a further substantial procedural violation justifying the reimbursement of the appeal fee.
XII. The respondent argued essentially as follows:
The Opposition Division had correctly analysed the meaning of "reference" and "link" in the context of the patent in suit and come to the conclusion that these terms had different meanings. Hence, the statement "by use of links such as Universal Resource Locators" violated Article 123(2) EPC.
The disclosure of document D6 differed from the claimed subject-matter only in that, according to the latter, the referencing document was aggregated with the documents it referenced. Indeed, clustering and aggregating were synonyms and the clustering practised in document D6 was in fact based on combining a referenced document with its referencing documents. The indexing step associated with clustering was explicitly mentioned in chapter 2.2 of document D6.
As found by the Opposition Division, it was not possible to define an objective technical problem on the basis of document D6. Thus, in view of this prior art, the claimed invention did not satisfy the requirements of Article 56 EPC.
On the other hand, although D6 was essentially directed to improving the retrieval of images from the Web, it was specified in section 2.2 that this approach was general enough to permit its application to any type of media. Starting from document D6, a problem addressed by the contested patent could be seen in extending the process of document D6 to textual documents containing index terms.
Document D11 disclosed a wide variety of clustering methods, some being based on semantics, others on hyperlinks. In particular, document D11 specified on page 161, right-hand column, paragraph 3 that the hyperlink structures were used to rank retrieved pages and could also be used for clustering relevant pages on different topics. Consequently, document D11 taught, at least implicitly, clustering as a synonym of aggregating a referencing document with referenced documents.
The combination of the teachings of documents D6 and D11 was obvious to the skilled person and would lead to the claimed invention.
Document D23, which was cited in the bibliography of document D11, related to clustering methods for large document databases. At the time of its publication, 1988, the Web, and thus hyperlinks, had not yet been invented. However, as specified in D23 (page 578, lines 6 to 17), the advantage of clustering to obtain better results in a document search was well known. Besides clustering, document D23 was concerned with the best searching strategies and specified that some sort of document-like representative was required to summarise the index term characteristics of the documents in a cluster. Thus, document D23 effectively disclosed the step of indexing an "aggregate document", as specified in claim 1 of the contested patent.
In summary, documents D11 and D23 disclosed document clustering and cluster indexing to obtain "better quality" search results. In particular, document D11 taught that clustering could be based on hyperlinks, whereas D23 pointed out on page 590, first paragraph, that small clusters were best for document retrieval. On the basis of this teaching, it would have been obvious to a person skilled in the art to use clusters limited to the referencing documents and the directly referenced documents.
Hence, the subject-matter of claim 1 resulted from an obvious combination of the teachings of documents D11 and D23.
1. The appeal is admissible.
Main request
2. Claim 1 of the contested patent relates to a "computer-implemented process for searching among a collection of Internet or HTML documents", and comprises the following features (itemised by the Board):
- the collection comprising:
(a) referencing documents and referenced documents
(i) referenced in the referencing documents by use of links such as Universal Resource Locators,
(ii) wherein said referencing and referenced documents contain index terms,
- the process comprising:
- for a referencing document of the collection:
(b) recognising documents referenced in said referencing document; and
(c) aggregating said referencing document with documents it references to form an aggregate document,
- said aggregate document formed being a single logical document associated with the documents forming said aggregate document;
(d) indexing said aggregate document, based on index terms contained in the referencing and referenced documents forming said aggregate document, to form an index,
- wherein an entry in the index table is associated with the referencing and referenced documents forming the aggregate document;
(e) searching among aggregate documents by operating on said index; and
(f) providing, as a result, an aggregate document.
Articles 100(c) and 123(2) EPC
3. According to the contested decision (point 16), it was not plausible that the terms "link" and "reference" were used as synonyms in the patent in suit. Hence, the statement "by use of links such as Universal Resource Locators" recited in claim 1 (feature (a)(i)) of the contested patent, but not in claim 1 of the original application, conflicted with the preceding statement "the collection comprising referencing documents and referenced documents referenced in the referencing documents" (see feature (a)), and consequently did not have a valid basis in the application as filed (Article 123(2) EPC).
3.1 On the contrary, in the appellant's view, it was directly and unambiguously derivable from the original application that the terms "link" and "reference" were synonyms and that links could be in the form of URLs.
3.2 With regard to the objection of added subject-matter (Article 123(2) EPC), the respondent maintained the view expressed in the opposition proceedings that the wording "by use of links such as Universal Resource Locators", as used in the context of claim 1, was not originally disclosed.
3.3 According to the original application, the "invention relates to the field of information retrieval, and more specifically to displaying results to a search query. It particularly applies to searches on the Internet, in Intranets, in mails, archives, files, databases or the like" (published application, paragraph [0001]).
3.3.1 A first embodiment of the invention is directed to a "process for displaying the results of a search among a collection of documents, the collection comprising a referencing document and a referenced document referenced in the referencing document" (ibid. paragraph [0012]).
3.3.2 A second embodiment relates to "a process for searching in a collection of documents. The collection comprises a referencing document and a referenced document referenced in the referencing document" (ibid. paragraph [0014]).
3.3.3 Paragraph [0004] specifies that a web page (accessed by the user through a given URL) "may also comprise a number of links to various types of documents, in the form of URLs embedded in the page. The links may bring the user to html pages, to audio or video files, or to other linked files".
3.3.4 Hence, the application as filed is generally concerned with documents, such as web pages, which contain links to various types of documents (i.e. "referenced documents"), such as other HTML pages, video and audio files or unspecified types of files. According to paragraph [0004], these links are in the form of URLs embedded in a web page (i.e. a "referencing document".
3.3.5 In paragraphs [0036] to [0040], which relate to the second embodiment of the invention (namely to a search process), it is not explained how a referenced document is actually linked to a referencing document. It fact, it seems implicit that various links may be possible, since the invention as originally disclosed applies to searches "on the Internet, in Intranets, in mails archives, databases or the like" (see paragraph [0001]).
3.3.6 In paragraphs [0040] and [0041], the second embodiment of the invention is described with reference to Figures 1 and 2, where the referencing document is an HTML document, identified by its URL, which contains links to two audio files (the referenced documents). It is implicit that the links to audio files may be expressed in the form of URLs.
3.3.7 Claim 1 of the patent indicates that the process is directed to searching "among a collection of Internet or HTML documents", and is therefore limited to documents of the kind shown in Figures 1 and 2.
3.3.8 Although it is not specified in the context of the second embodiment that the "referenced documents" are "referenced in the referencing documents by use of links such as Universal Resource Locators", it is immediately apparent to the skilled reader of the application that an HTML document ("referencing document") may contain links to one or more other documents ("referenced documents") in the form of URLs.
3.3.9 In summary, the Board finds that the feature "by use of links such as Universal Resource Locators" in claim 1 of the patent finds support in the original application. Hence, claim 1 does not violate Article 123(2) EPC.
Admissibility of document D23
4. Document D23 is cited in document D11 which, as argued by the respondent, the appellant used as closest prior art for the first time in the statement of grounds of appeal.
4.1 According to the respondent's reply to the statement of grounds of appeal, the late filing of document D23 aimed at proving that the appellant's reading of document D11 was not correct and that if D11 were to be considered as the closest prior art, the claimed invention would be obvious.
4.2 Document D23 deals specifically with the use of clustering methods for document retrieval and thus appears to contribute to the discussion on clustering started by the appellant's identification of document D11 as the prior art from which the inventive step assessment of the claimed subject-matter should start.
4.2.1 Hence, the Board accepts that the filing of document D23 was occasioned by the arguments submitted by the appellant in the statement of grounds and finds that this document is prima facie sufficiently relevant to be admitted into the appeal proceedings despite its late filing.
Article 54 EPC
5. In the contested decision, the Opposition Division found that none of the documents on file disclosed all the features recited in claim 1 of the contested patent.
5.1 The respondent did not contest the finding of the Opposition Division concerning the novelty of the subject-matter of claim 1 (Article 54 EPC).
Article 56 EPC
6. In the contested decision (see point 18), the Opposition Division identified the aggregating step recited in claim 1 of the patent (see feature (c)) and hence also the indexing step based on the aggregating step (feature (d)) as novel over the prior art. These features provided the technical effect of different aggregated documents and thus different search results.
6.1.1 In the Opposition Division's opinion, it was not possible, on the basis of these differences, to identify a technical problem solved by the claimed method, which could be the basis of an inventive step.
6.1.2 Furthermore, the Opposition Division held that the proposed solution could not be regarded as an alternative solution to the one proposed by the prior art firstly because a technical problem could not be identified and secondly because the results provided by the method of the opposed patent and the method of the prior art were different.
6.2 The Board finds the inventive step reasoning of the Opposition Division barely understandable and not plausible. In particular, the Opposition Division's conclusion that some technical features distinguished the claimed invention from the prior art, but that, on the basis of such technical differences, it was not possible to identify a technical problem solved by claimed method should, in the Board's opinion, put into question the choice of the closest prior art document, rather than directly point to a lack of inventive step.
6.2.1 Though not expressly stated, it appears from the contested decision that documents D6 and D9 represented the closest prior art according to the Opposition Division (see point 17 of the decision). In the appeal proceedings the respondent did not refer to document D9.
6.2.2 In the Board's opinion, the distinguishing features of claim 1 identified by the Opposition Division in paragraph 18 of the decision do not reflect the actual differences between the claimed process and the prior art document D6.
6.3 Document D6 relates to a computer implemented process for searching among a collection of Internet "documents" (see Abstract: "... a model for retrieval of images from a large World Wide Web based collection").
6.3.1 As illustrated in Figures 1. and 2., the collection comprises "referencing documents" ("textual nodes") and "referenced documents" ("non-textual nodes") referenced in the referencing documents by use of links such as Universal Resource Locators (see Section 2.2.1, first paragraph).
6.3.2 According to section 2.1, document D6 uses links in a hypermedia collection to calculate representations for non-textual nodes that permit their direct retrieval by textual query. The textual nodes linked to a non-textual node can be considered as forming a cluster. Cluster description techniques can then be applied in order to calculate a representation and establish the overall content of the documents forming the cluster. The representation can be subsequently assigned to the non-textual node, giving it a retrieval content equal to the combined content of the textual nodes connected to the non-textual node. In particular, as specified in section 2.1 of D6 (left-hand column, last paragraph) the "representation of a non-textual node can be calculated by considering each linked document L as a standard N-dimensional vector, where N is the number of index terms in the document base".
6.3.3 According to section 2.2 of D6, the process of providing a non-textual node (referenced document) with a representation (index term) comprises the following steps:
- recognising all the textual nodes (referencing documents) linked to the referenced document (non-textual node);
- using the content of these nodes to form a new node which is the representation of the non-textual node;
- the collection of representations is indexed.
6.3.4 D6 (section 2.2, page 2, right-hand column, lines 23 to 27) further specifies that with the index of these "new nodes" documents corresponding to the non-textual nodes can be retrieved in the same fashion as any standard text document. It is the task of the retrieval system's interface to associate the documents retrieved with their corresponding non-textual nodes and present them to the user.
6.4 The new node, formed on the basis of the content of the textual nodes linked to a non-textual node, could be regarded as constituting a "single logical document" within the broad meaning that this term appears to have in the patent.
6.4.1 Also the steps of claim 1 relating to indexing, searching and providing as a result an aggregate document appear to cover steps disclosed in or necessarily implied by document D6. In the prior art, however, the step of "aggregating" is directed to the index terms of the referencing documents and the resulting "aggregated index terms" are used for indexing a single referenced document (i.e. non-textual node) which can not be directly indexed.
6.4.2 As to the problem addressed in document D6, it differs substantially from the object of the present invention. Document D6 seeks to determine index terms which can at least approximate the content of a non-textual document on the basis of its links to textual documents, whereas the patent in suit relies for indexing on the index terms of both the referencing document and the referenced documents.
6.5 The respondent has argued that it was known from D11 to use clustering for improving the indexing of documents. In particular, document D11 referred to a clustering method which used words contained in a document and links from and to the document. In the respondent's view, the teaching of document D6, combined with clustering by means of links, as disclosed in document D11, would lead the skilled person to the present invention.
6.6 As pointed out above (point 6.4.2), the Board considers that document D6 is directed to the specific problem of indexing a non-textual document by means of the index terms of the textual documents linked to the non-textual document.
6.6.1 It is indeed pointed out in document D6 (section 2.2, page 2, right-hand column, lines 27 to 29) that the disclosed teaching can be applied to any type of media. However, D6 does not foresee the possibility of applying the same teaching to a collection of textual documents for improving indexing. Furthermore, as stressed by the appellant, there is no reason to assume that it would be obvious to apply the particular indexing method of document D6 to a cluster of documents when standard indexing based on the index terms of each document can be used.
6.6.2 In summary, the Board agrees with the appellant that the teaching of document D6 does not constitute a promising starting point for arriving at the claimed invention.
6.7 The respondent has furthermore developed an additional line of attack against the patentability of the claimed invention and argued that the contested patent effectively taught creating aggregate documents, namely clusters, on the basis of links, and indexing the aggregate documents on the basis of their index terms. In document D11, page 161, right-hand column, paragraph 3, it was disclosed that hyperlink structures could be used for clustering relevant pages on different topics. Thus, document D11 implicitly disclosed clustering, i.e. aggregating a referencing document with corresponding referenced documents.
6.7.1 In the respondent's view, it was known from document D23 (page 589, paragraph 3) that some sort of document-like representative was required to summarise the index terms of the documents in a cluster. This teaching corresponded to the feature of indexing recited in claim 1 of the contested patent. Considering that document D23 gave a further hint that small clusters were best for retrieval, the respondent concluded that it would have been obvious to the skilled person to use clusters limited to the referencing and the referenced documents and thus arrive at the claimed invention.
6.8 The appellant has essentially contested that the aggregating step of claim 1 was to be assimilated to clustering and pointed out that the combination of clustering and indexing at the basis of the respondent's argumentation went against the skilled person's experience and the actual development of search engines.
6.8.1 In the appellant's opinion, document D23 did not refer to index tables and to inverted indexes, which were a key feature of the present invention. The use of an index table, which was state of the art at the priority date of the patent, made it possible to reduce search time. On the contrary, document D23 presented search methods that did not use index tables, but relied on tree-searching algorithms. This had been state of the art at the time of publication of document D23 (1988) and consisted in hierarchically clustering documents in a tree-like structure.
6.8.2 According to the appellant, the methods disclosed in document D23 were not applicable to a very high number of documents and in fact had been abandoned at the time of filing of the opposed patent. In other words, the use of clusters in document D23 had nothing to do with the aggregating step referred to in the patent which created logical documents to be indexed and searched using an inverted index. In summary, the combination of documents D11 and D23 was artificial and could not lead the skilled person to the contested patent.
6.9 It is specified in document D23 (page 577, "INTRODUCTION", first paragraph) that clustering involves grouping of similar objects. According to section 2.1, cluster analysis methods are all based in some way on measurements of the similarity between a pair of objects, such as individual documents. The determination of similarity between a pair of objects involves the selection of variables used to characterise the objects, the selection of a weighting scheme for these variables and the selection of a similarity coefficient to determine the degree of resemblance between the two attribute vectors.
6.9.1 According to the contested patent, documents are aggregated not on the basis of their similarity or of shared characteristics, but only because a referencing document contains links to the other documents which may have a completely different content.
6.9.2 Thus, the Board agrees with the appellant that the step of aggregating a document with other documents merely because the former contains references to the latter, as specified in the patent, does not correspond to clustering, as disclosed in documents D11 and D23.
7. In summary, the Board comes to the conclusion that the subject-matter of claim 1 was not made obvious by the teaching of any of documents D6, D11 and D23, or by any of their combinations.
8. As none of the grounds for opposition prejudices the maintenance of the patent as granted according to the appellant's main request, there is no need to consider the auxiliary requests.
Reimbursement of the appeal fee
9. The appellant has essentially argued that during the oral proceedings the Opposition Division raised a new argument under Article 100(a) EPC and, in particular, stated for the first time that the technical problem solved by the invention was speculative and not acceptable as a valid objective technical problem. The patent proprietor was allegedly taken by surprise. Furthermore, in the appellant's opinion, the right to be heard was violated because the patent proprietor was not given the possibility to file amended claims in response to the new argument.
9.1 The appellant has further submitted that the Opposition Division failed to correctly apply the problem-solution approach and to consider the patent proprietor's right to reformulate the objective technical problem in a "less ambitious" way.
9.2 In the first paragraph of the section of the contested decision relating to Article 56 EPC, the Opposition Division merely identified the aggregating step and hence also the indexing step based on the aggregating step as novel over an unspecified prior art. From the preceding section concerning Article 54 EPC, it can be assumed that either document D6 or D9 constituted the prior art the Opposition Division intended to refer to. In fact, the minutes (section 6.) of the oral proceedings show that the inventive step discussion had focused on D6.
9.2.1 According to paragraph 18 of the contested decision, the aggregating and indexing steps recited in claim 1 provided the technical effect of different aggregated documents and different search results. The Opposition Division appears to have acknowledged the presence of a technical effect in the form of "different aggregated documents" and "different search results".
9.2.2 Having identified the differences between the prior art and the claimed invention, the Opposition Division concluded that it was not possible to identify a technical problem solved by the method of the application, which could be the basis of an inventive step.
9.2.3 In paragraph 20 of the decision, the Opposition Division summarised the different attempts of the patent proprietor to define a technical problem starting from the closest prior art document and, in particular, the patent proprietor's view that if no other problem could be formulated, one had to define the problem as how to find an alternate solution to the prior art and ask whether the skilled person would arrive from there at the claimed invention.
9.2.4 In reply to the patent proprietor's arguments, the Opposition Division stated in paragraph 21 of the decision that these argumentations were not convincing because it was not possible to consider the proposed solution as an alternative solution to the one found in the prior art firstly because a technical problem could not be identified and secondly because the results provided by the method of the opposed patent and by the method of the prior art were different.
9.2.5 Finally, in paragraph 21 of the contested decision, the Opposition Division reiterated that claim 1 did not meet the requirements of Article 56 EPC because a technical problem solved by the method of claim 1 could not be identified.
9.3 The Board notes that all the above conclusions of the Opposition Division relating to the lack of inventive step are presented in the contested decision without any prior analysis of the allegedly closest prior art document (D6), or of the problem it addresses. It is also not explained why the technical differences between the claimed subject-matter and the closest prior art identified by the Opposition Division did not allow the definition of a technical problem to be solved according to the problem-solution approach (see also point 6.2 above).
9.3.1 Similarly, the Board finds that the Opposition Division did not explain in the contested decision why, in its opinion, a synonymous use of the terms "link" and "reference" was not plausible and consequently the statement "by use of links such as Uniform Resource Locators" in claim 1 offended against Article 123(2) EPC (see point 16 of the contested decision). In particular, the Opposition Division did not provide any reasons for refuting the patent proprietor's arguments.
9.3.2 In the Board's opinion, the juxtaposition of seemingly contradictory conclusions (e.g. acknowledgement of technical differences but no possibility to define a technical problem), presented in the contested decision as apodictically evident and without any factual support, does not constitute a "reasoned decision" within the meaning of Rule 111(2) EPC.
9.4 In accordance with the case law of the boards of appeal (see Case Law of the Boards of Appeal, Seventh Edition, September 2013, IV.E.8), the Board considers that in the present case the failure to provide adequate reasoning in a decision is to be regarded as a substantial procedural violation justifying the reimbursement of the appeal fee.
For these reasons it is decided that:
1. The decision under appeal is set aside.
2. The patent is maintained as granted.
3. The appeal fee is to be reimbursed.