T 0328/09 (Electronic publishing/TIME BASE) of 14.1.2015

Summary of Facts and Submissions

I. The applicant (appellant) filed an appeal against the decision of the Examining Division to refuse European patent application no. 98901249.7.

II. The contested decision referred, inter alia, to the following documents which were filed with third party observations dated 18 April 2007:

D5: "Xsoft, A Division of Xerox", Internet citation, [online] 1996, retrieved from the Internet: URL: http://xml.coverpages.org/duCharme-sgmldbms.html#ID18>;

D6: "XSOFT PREMIERES ASTORIA; A SIMPLER WAY TO MANAGE 'MEGA-DOCUMENTS'", Internet citation, [online] 1996, retrieved from the Internet: URL: http://www.highbeam.com/doc/1G1-18079234.html>;

D7: "Xsoft Astoria", Internet citation, [online] 1996, retrieved from the Internet: URL: http://www.architag.com/tag/Article.asp?v=10&i=4&p=8&s=1>.

III. In the decision under appeal, the Examining Division held, inter alia, that claim 1 of the main request filed with letter dated 4 July 2008 did not involve an inventive step within the meaning of Article 56 EPC, having regard in particular to document D5.

Furthermore, the Examining Division decided not to admit a first auxiliary request filed at the oral proceedings on 8 July 2008, since, in the Examining Division's opinion, the applicant had already been given ample and sufficient opportunity to amend the application. Furthermore, this request could not prima facie overcome the objections under Article 56 EPC.

IV. With the statement of grounds of appeal dated 14 November 2008, the appellant filed, as main request, a set of claims 1 to 58 corresponding to the main request considered in the contested decision, and a new set of claims 1 to 39 by way of first auxiliary request.

V. In a communication dated 21 November 2014 accompanying the summons to oral proceedings, the Board raised objections under Articles 84 and 56 EPC and, in particular, expressed the preliminary opinion that both the main request and the auxiliary request did not appear to provide a basis for granting a patent.

VI. In reply to the Board's communication, the representative of the appellant, by letter dated 18 December 2014, informed the Board that the applicant had decided not to attend the oral proceedings.

VII. On 14 January 2015, the Board held oral proceedings as scheduled in the absence of the appellant. At the end of the proceedings, the Chairman pronounced the Board's decision.

VIII. With the notice of appeal the appellant requested that the decision under appeal be set aside in its entirety. With the statement of grounds of appeal, the appellant implicitly requested the grant of a patent on the basis of the main request or, if this was not possible, on the basis of the first auxiliary request, both requests having been filed on 14 November 2008.

IX. Claim 1 according to the main request reads as follows:

"A computer-implemented system for publishing an electronic publication using text-based data, said computer-implemented system characterised by:

means for providing a plurality of predefined portions of text-based data with each predefined portion being stored;

means for providing a plurality of linking means of a markup language;

means for modifying and storing at least one predefined portion so that the at least one predefined portion and a corresponding modified predefined portion are both stored;

means for encoding each predefined portion of said text-based data and said at least one modified predefined portion of text-based data with at least one linking means; and

means for organising and retrieving said plurality of predefined portions and said at least one modified predefined portion of said text-based data using a plurality of attributes (s1, s2, s3, s4, L, C, J), each attribute being a point on an axis of a multidimensional space (100);

wherein said plurality of predefined portions and said at least one modified predefined portion can be directly retrieved using said plurality of attributes to define the point in said multidimensional space that corresponds to one of said plurality of predefined portions or said at least one modified predefined portion."

Claim 1 according to the first auxiliary request reads as follows:

"A computer-implemented system for publishing text-based data, said computer-implemented system characterised by:

means for storing said text-based data;

means for receiving instructions and dividing said text-based data into a plurality of predefined portions, wherein each said predefined portion is an optimum storage unit that is identified by analyzing the naturally occurring structure of said text-based data and how said text-based data would be used by an end user;

means for modifying at least one predefined portion and storing a corresponding at least one modified predefined portion, wherein the at least one predefined portion and the corresponding modified predefined portion are both stored;

means for encoding said text-based data using a plurality of attributes (s1, s2, s3, s4, L, C, J), each attribute being a point on an axis of a multidimensional space (100); and

means for encoding each of said predefined portions and said at least one modified predefined portion with at least one linking means for connecting said plurality of attributes;

wherein said system is configured such that said plurality of predefined portions and said at least one modified predefined portion can be directly retrieved using said plurality of attributes to define the point in said multidimensional space that corresponds to one of said plurality of predefined portions or said at least one modified predefined portion; and

wherein said predefined portions do not overlap."

Both the main request and the auxiliary request comprise further independent claims directed to a computer readable recoding medium (claim 20 in the main request and claim 14 in the first auxiliary request) and to a computer-implemented method (claim 40 in the main request and claim 27 in the first auxiliary request). As these claims are not relevant to the Board's decision, their full text need not be given.

X. The arguments submitted by the appellant with the statement of grounds of appeal may be summarised as follows:

Features (c) and (f) of claim 1 of the main request (as itemised in this decision, point 3) were not known from document D5. Documents D5 to D7 related to an object-oriented document component management system. According to document D5, the system allowed for fine-grained access and version control (cf. page 1 of D5), and maintained revision information on individual elements (cf. page 2). D6 indicated that the known system included capabilities such as version control, revision tracking and component re-use. D7 showed a system that worked at the document component level as opposed to the document level of other competing systems.

None of documents D5 to D7 disclosed all the features of claim 1. In fact, before the present invention there were two methods for performing versioning: a whole document approach and a document assembly approach.

According to the document assembly approach, which was consistent with the state of the art at the time of the present invention, the system, in addition to storing the original document, also stored particular version information relevant to the document and all of its components (revision tracking). When a particular version of the document was requested, the version information for all relevant components was retrieved. These instructions were then applied to assemble the components so that the correct version of the document was displayed.

The document assembly approach involved a level of complexity difficult to administer and maintain without a large number of errors. In particular, since each change built on and had to be seen in the context of previous changes, it was necessary to apply each change in the correct order, if errors were to be avoided. Performing a stable document assembly in real-time became unfeasible very quickly because of the large number of revision instructions that might need to be retrieved to assemble the requested information.

In the system according to the present application, revision instructions were not maintained because it was not necessary to keep the document hierarchy. The approach was in fact completely different and constituted a fundamental shift from both prior art approaches. In other words, the present invention was a shift from versioning with a certain hierarchy to a novel method and approach that was primarily concerned with improving usability for a particular application.

In its relevant parts, claim 1 according to the main request recited a system for publishing that comprised means for modifying and storing at least one predefined portion so that the at least one predefined portion and a corresponding modified predefined portion were both stored.

In addition to storing modified predefined portions, claim 1 recited that the plurality of predefined portions and the at least one modified portion could be directly retrieved using the plurality of attributes to define the point on said multidimensional space that corresponded to one of said plurality of predefined portions or said at least one modified predefined portion. Direct retrieval, as recited in claim 1, was counter to the disclosure in D5 to D7 because the document reassembly approach taught away from storing data in the same manner.

Accordingly, the subject-matter of claim 1 of the main request involved an inventive step (Article 56 EPC).

Claim 1 according to the first auxiliary request recited similar features and clarified some aspects of claim 1 according to the main request. In addition, claim 1 of the first auxiliary request recited that the predefined portions did not overlap. It would be clear to a person skilled in the art that an attempt to use the present invention to do versioning at multiple levels (i.e. by allowing predefined portions to overlap) would have been inconsistent with the specification.

Reasons for the Decision

1. The appeal is admissible.

2. The present invention relates to an "electronic publishing system" and is supposed to overcome some of the drawbacks of prior art systems which publish information in electronic form "using the document or book metaphor" (see application as published, page 1, lines 12 to 15).

2.1 The alleged disadvantages of conventional publishing systems are mentioned in the "BACKGROUND" section of the description with respect to the search for legal information scattered in different documents which have undergone multiple revisions. In this particular case, a user has to consult many different documents before finding the correct answer to a query.

2.2 In particular, the application points out that the "smallest piece of information" used by conventional systems for publishing legal provisions is either (I) an Act or Regulation, or (II) the word. Typically, conventional publishing systems choose the word as the smallest piece of information when legislation is amended. However, tracking every single word involves a level of complexity that is difficult to administer (cf. application as published, page 2, line 40 to page 3, line 30). Hence, it is impractical to store the complete history of every word and phrase of a piece of legislation that has undergone many amendments.

Main request

3. Claim 1 according to the main request relates to a "computer-implemented system for publishing an electronic publication using text-based data". The features of the claimed system can be itemised as follows:

(a) means for providing a plurality of predefined portions of text-based data with each predefined portion being stored;

(b) means for providing a plurality of linking means of a markup language;

(c) means for modifying and storing at least one predefined portion so that the at least one predefined portion and a corresponding modified predefined portion are both stored;

(d) means for encoding each predefined portion of said text-based data and said at least one modified predefined portion of text-based data with at least one linking means; and

(e) means for organising and retrieving said plurality of predefined portions and said at least one modified predefined portion of said text-based data using a plurality of attributes, each attribute being a point on an axis of a multidimensional space;

(f) wherein said plurality of predefined portions and said at least one modified predefined portion can be directly retrieved using said plurality of attributes to define the point in said multidimensional space that corresponds to one of said plurality of predefined portions or said at least one modified predefined portion.

3.2 The Examining Division considered that D5 constituted the closest prior art and that the subject-matter of claim 1 differed from the known computer-implemented system for electronic publishing in that, according to the invention, a "predefined portion of text-based data" and the corresponding modified predefined portion of text-based data were stored.

3.2.1 In the Examining Division's opinion (decision under appeal, page 6, first paragraph), "storing both the original and the changed version of a changed document or document portion is a completely obvious and straightforward solution for the problem of performing document versioning administration, which would clearly have been within the common general knowledge of the skilled person, and which the skilled person starting from D5 would have chosen in accordance with circumstances and without involving an inventive step when trying to implement the system as specified in D5".

3.3 The appellant has contested that the distinction between the whole document and the document versioning approaches was trivial and that the skilled person would have been aware of both approaches and of the possible trade-offs between them.

In the appellant's opinion, the versioning system described in D5 to D7, like the document assembly approach, involved maintaining version information for all levels of hierarchy of objects (such as revision structures). Accordingly, any trade-off between the document assembly approach, adopted in documents D5 to D7, and the whole document approach would still necessarily involve storing revision instructions and maintaining the document hierarchy.

3.3.1 According to the appellant, the system recited in claim 1 stored both predefined portions and corresponding modified predefined portions and thus was essentially different from a system which performed versioning, as disclosed in D5 to D7, or relied on the whole document approach. Specifically, the system according to claim 1 divided the document into structural elements that corresponded to a unit of information a user might wish to access. The unit of information might vary for different documents or within the document. The predefined portions, therefore, could be selected as a function of the user's interests (cf. statement of grounds of appeal, point 8.).

3.3.2 Hence, if there was a change in a predefined portion, the predefined portion was modified and the modified predefined portion was stored in addition to the original predefined portion. Indeed, as acknowledged by the appellant, this method required more storage space than the document reassembly method. On the other hand, the system of the invention did not need complex real-time processing of instructions when a particular version was requested.

Interpretation of claim 1

4. In the Board's opinion, the system according to claim 1 of the appellant's main request comprises features, such as "means for providing a plurality of predefined portions", "means for modifying and storing at least one predefined portion" and "means for encoding each predetermined portion [ ... ] with at least one linking means", which in the broad context of the claim may be given different interpretations.

4.1 In particular, the predefined portions of text-data could be understood as portions which may be selected by a notional user, reflect some logical structure in the text, or which are determined by the author or publisher of the text data. As the "granularity" of the predefined portions is not specified in the claim, it could even be assumed that the words of the text-data constituted the "predetermined portions", and that the "modified predetermined portions" are words which have undergone revisions.

4.1.1 According to the wording of claim 1, the system of the invention appears to be responsible for "modifying the predefined portion". However, according to the description, feature (c) implies that the system of the invention receives a modified version of a stored predetermined portion, links it with the corresponding (unmodified) predetermined portion and stores it (cf. published application, page 9, lines 8 to 20).

4.2 Furthermore, if claim 1 is read in the light of the embodiments described in the application, it appears to comprise features which cannot be autonomously performed by a computer, i.e. without direct intervention by a user.

4.2.1 In particular, it appears that the predefined portions of text-data (such as a piece of legislation) are items which a user may wish to consult, and that the modified predefined portions are revisions of the text the user may wish to keep. In this case, the teaching of the invention would essentially consist in storing items of a piece of legislation together with their possible modifications so that the interested user may quickly retrieve the desired items in the required version.

4.3 Apart from the different interpretations that may be given to some of the claim features, the Board agrees with the appellant that claim 1 of the main request essentially relates to a system for publishing text-based data which stores both predefined portions of the text-based data and corresponding modified versions of these positions, so as to allow a simple and speedy retrieval of any version of a predefined portion (see paragraphs 7 and 9 of the statement of grounds of appeal).

Inventive step

5. For the inventive step assessment it is assumed, for the sake of argument, that the expression "predetermined portions" in claim 1 does not define single "words" of text data, since according to this interpretation of "predefined portions", prior art systems based on the document assembly approach as acknowledged in the introductory part of the description would appear to fall within the terms of claim 1.

6. Document D5, which the Examining Division considered to represent the most relevant state of the art relates to an electronic publishing system known as "Astoria" and developed by "XSoft", a Division of "Xerox". Aspects of the "Astoria" publishing system are also disclosed in documents D6 and D7.

6.1 According to the first paragraph of document D5, "Astoria is an object-oriented document component management system that enables users to easily find, use, share and manage SGML [Structured Generalised Markup Language] documents and their components, as well as unstructured documents" (cf. in this context also the application as published, page 9, lines 8 to 11).

6.1.1 As specified in the second paragraph of page 2 of document D5, "Because of its sophisticated integration with SGML editors, Astoria maintains revision information on individual elements, and past versions are always available".

As pointed out on page 2 of D5, third paragraph,"[a]ny SGML element stored in Astoria can be referenced in many different documents".

6.1.2 Furthermore, "Astoria provides a search tool that makes element reuse straightforward. Astoria's search engine lets users search on document content, SGML structure, SGML attributes, and version data such as date and author" (D5, page 2, paragraph "SEARCH").

6.2 In the chapter bridging pages 1 and 2 of document D6, it is also specified that "Astoria builds on the value of SGML with document management capabilities such as version control, revision tracking and component re-use".

6.3 In document D7 (page 1, lines 8 to 12) it is in particular specified that "Astoria deals with the concept of "document components". A document component is a piece that is designed to be maintained as a unit. Whether this be at the volume or book level, or at some finer granular point, such as paragraph or list item".

6.4 In the Board's opinion, document D5 relates, together with documents D6 to D7, to a computer-implemented system for publishing an electronic publication using text-based data which comprises or necessarily implies the following features:

- means for providing a plurality of predefined portions of text-based data with each predefined portion being stored;

- means for providing a plurality of linking means of a markup language;

- means for modifying at least one predefined portion and storing corresponding modifications;

- means for encoding each predefined portion of said text-based data with at least one linking means; and

- means for organising and retrieving said plurality of predefined portions of said text-based data using a plurality of attributes, each attribute being a point on an axis of a multidimensional space;

- wherein said plurality of predefined portions can be directly retrieved using said plurality of attributes to define the point in said multidimensional space that corresponds to one of said plurality of predefined portions.

6.4.1 As to the last two features which are considered to be necessarily implied by the Astoria system, it is specified on page 9, lines 13 to 18, of the published application that "[f]or each of the predefined portions, the system stores a copy of the predefined portion and a modified predefined portion in the first database whenever it is changed. A second (relational) database is preferably provided that comprises plural attributes for managing the information of the first database, with each attribute being a point on an axis of a multidimensional space for organising the data for publication".

6.4.2 In other words, the present application merely teaches that the predetermined portions (and the modified predetermined portions) are retrieved on the basis of (SGML) attributes and that these attributes constitute the dimensions of a multidimensional space. As the description does not disclose any particular implementation of the first and second databases, it has to be assumed that the applicant relied on what was known in the art at the priority date of the present application.

6.4.3 Similarly, document D5 refers to a search tool provided by "Astoria" which, inter alia, makes use of SGML attributes (D5, page 2, paragraph "Search"). The skilled reader of D5 would understand these attributes as dimensions of a multidimensional space such that a "predetermined portion" could be searched and retrieved by specifying a point defined by specific attributes in the corresponding multidimensional space (cf. published application, page 1, lines 12 to 15).

6.5 As pointed out above (see point 6.1.1), the "Astoria" system maintains revision information on individual elements so that past versions are always available. However, none of the documents D5 to D7 specifies how revision tracking is actually implemented and, in particular, whether different versions of a "document component" (cf. D7), such as a paragraph or a list item, are stored as modified "document components" or as "revision instructions" of the original document component.

6.5.1 As acknowledged in the description of the present application, it is known to the skilled person to keep track of different versions of a document by storing either "current information" (i.e. up-to-date information) or "historical information" (cf. published application, page 1, lines 21 to 26). Another conventional approach used in electronic publishing of documents subject to multiple revisions is to keep track of amendments at the level of "smallest piece of information" which may be represented by an Act, a Regulation or a word (cf. application as published, page 2, line 40 to page 3, line 6). As pointed out on page 3, lines 6 and 7, conventional publishing systems typically choose a word as the smallest piece when legislation is amended.

It is evident that the first approach requires more storage capacity but allows a fast retrieval of a desired version, whereas the second approach saves on storage but requires more time to fetch a particular version which possibly has to be assembled on the basis of stored tracking information.

6.5.2 It is obvious to a skilled person that a particular user who requires frequent access to different versions of some paragraphs or other items of text-based data will benefit from an electronic publishing system which provides ready access to the required versions, and that ready access can be easily achieved by storing the revised versions of the required items, instead of assembling a particular version of an item each time it is requested. In fact, this could be regarded as an application of the conventional approach of storing information as either "current or historical information" to particular "portions" of a document which are known to be required by a user.

If, on the other hand, saving storage were to be given priority and processing time were of no concern, it would be obviously preferable to store only revision information and assemble the desired version when required.

6.5.3 In other words, the Board essentially agrees with the Examining Division that the system according to claim 1 of the main request results from an obvious and straightforward trade-off between storage and runtime efficiency and, in particular, between the cost of storing redundant data and the cost of assembling document versions at retrieval time.

6.5.4 Hence, it would have been obvious to the skilled person to adapt the Astoria system known, for instance, from D5 to the particular demands of a user requiring frequent and fast access to different versions of particular portions of text-based data by storing both the original and the modified portions of the text-based data (feature (c) of claim 1).

6.5.5 In the Board's view, claim 1 also covers the case of a document which is to be published electronically and has been divided into multiple electronic files, i.e. "predefined portions" as specified in the claim. If a revision of the document were to be published, it would be obvious to replace only the modified files. Furthermore, if the user wished to keep a history of all updates, it would also be obvious to store the original files and their modified versions and to provide a link between them (feature (d) of claim 1).

6.6 As to features (e) and (f) of claim 1 (cf. point 6.4.3 of this decision), the Board considers that it would have been straightforward for the skilled person to use (SGML) attributes to identify the modified "predetermined portions" of text data and extend the use of the search tool referred to in D5 (second page) also to the "modified predetermined portions" so that they could be directly retrieved in the same way as the original "predetermined portion".

6.6.1 In summary, the Board finds that it would have been obvious to the skilled person, starting from the "Astoria" system according to D5 and wishing to adapt it for the retrieval of items of a large document in their different versions to arrive to a system falling within the terms of claim 1 of the appellant's main request (Article 56 EPC).

First auxiliary request

7. Claim 1 according to the first auxiliary request relates to a computer-implemented system for publishing text-based data which comprises the following features:

(i) means for storing said text-based data;

(j) means for receiving instructions and dividing said text-based data into a plurality of predefined portions,

(k) wherein each said predefined portion is an optimum storage unit that is identified by analyzing the naturally occurring structure of said text-based data and how said text-based data would be used by an end user,

(l) means for modifying at least one predefined portion and storing a corresponding at least one modified predefined portion,

(m) wherein the at least one predefined portion and the corresponding modified predefined portion are both stored;

(n) means for encoding said text-based data using a plurality of attributes, each attribute being a point on an axis of a multidimensional space; and

(o) means for encoding each of said predefined portions and said at least one modified predefined portion with at least one linking means for connecting said plurality of attributes;

(p) wherein said system is configured such that said plurality of predefined portions and said at least one modified predefined portion can be directly retrieved using said plurality of attributes to define the point in said multidimensional space that corresponds to one of said plurality of predefined portions or said at least one modified predefined portion; and

(q) wherein said predefined portions do not overlap.

7.2 Feature (i) represents the storing step in feature (a) of claim 1 of the main request. Features (l) and (m) correspond essentially to feature (c) according to claim 1 of the main request. Feature (p) is virtually the same as feature (f), whereas feature (o) differs from feature (d) in that the at least one linking means is for connecting "said plurality of attributes".

7.3 Apart from feature (e) of claim 1 of the main request which finds no correspondence in claim 1 of the first auxiliary request, features (j), (k), (n) and (q) constitute the essential differences between the two requests.

7.3.1 According to the appellant, features (j) and (k) found support at page 8, line 40 to page 9, line 39 of the published application and, in particular at page 9, lines 34 to 39.

7.3.2 The original application does not explicitly disclose "means for receiving instructions and dividing ...". According to the description (application as published, page 9, lines 34 to 39), the "embodiments advantageously divide information into "suitably" small pieces (or blocks) of text ... . The suitability as to size of text pieces is determined by an analysis of the information and its naturally occurring structure based on knowledge of how the information is used and consumed by the end user". However, the application does not explain how the system of the invention would carry out "an analysis of the information" and take into account how the information is "used and consumed by the end user", so as to determine the "optimum storage unit". In fact, it appears that the implementation of features (j) and (k) would, at least partly, involve some intellectual activity and thus imply direct human intervention. Furthermore, it appears that the "optimum storage unit" can only be determined with respect to specific text-data and end users. As neither the text-data nor the end user is defined in the claim, it could be questioned whether these features satisfy the requirement of clarity.

7.4 As to feature (q), the appellant has submitted that it would be clear to the skilled person that an attempt to use the present invention to do versioning at multiple levels (i.e. by allowing predefined portions to overlap) would have been inconsistent with the specification. Furthermore, the reference in the description (page 9, lines 11 to 13) to the possibility of implementing a hierarchy of divisions of legislation did not indicate that there could be an overlap between the portions.

7.4.1 The Board notes that there is no explicit disclosure of feature (q). However, it appears that in the exemplary embodiments of the invention the portions of text-data do no overlap. In particular, the indication in the description (page 11, lines 37 to 39) that "the section-level portion is preferably the predefined portion of the publication data, which is the smallest piece of information to be tracked", appears to imply non-overlapping portions. On the other hand, the same could be said for conventional publishing systems (cf. page 2, line 40 to page 3, line 7).

7.5 Feature (n) is not explicitly disclosed in the application as filed where the use of "attributes" is consistently related to the "predefined portions", as in claim 1 of the main request.

The exact meaning of this feature in the context of the claim is not clear. In particular, it is not clear which text-based data, apart from the predefined portions, should now be encoded and what purpose this would serve (Article 84 EPC).

8. As pointed out above, features (j) and (k) appear to imply the intervention of a skilled operator and essentially relate to the preparation of text-data for electronic publishing. Furthermore, as their implementation depends on the text-data and on the targeted end user, it is questionable whether they have any limiting effect on the claimed subject-matter, a computer-implemented system, other than supporting different document units as predefined portions. Besides, also the system according to D5 and D7 allows the user to select document components, such as paragraphs (D5, page 1, last paragraph, and D7, page 1, lines 6 to 12).

8.1 Apart from the above objections relating to the clarity of claim 1, the essential aspects of the system according to the first auxiliary request remain, as for the main request, the storing of unmodified and modified portions of text-data and the use of attributes to retrieve the unmodified and the modified portions. As observed above, these aspects of the invention do not appear to be directed to the solution of a specific technical problem going beyond the setting up of a database for text-data and its subsequent revisions. Rather, they reflect the particular nature of the text-data considered in the application (texts of legislation), which may undergo multiple revisions, and respond to the particular requirements of the end user, who may wish to retrieve a particular version of just a portion of text.

8.2 Hence, the Board considers that, essentially for the same reason given with respect to the main request, the subject-matter of claim 1 according to the first auxiliary request does not involve an inventive step within the meaning of Article 56 EPC.

9. The Board thus comes to the conclusion that neither the main request nor the first auxiliary request provides a basis for granting a patent. Hence, the appeal has to be dismissed.

Order

For these reasons it is decided that:

The appeal is dismissed.