T 0940/03 (Combined DVD/CD ECC decoder/SAMSUNG) of 9.11.2006

Summary of Facts and Submissions

I. This is an appeal against the decision of the examining division to refuse the European patent application 01 202 036.8, with publication number 1 132 912, an application divided from European patent application 98 305 559.1, with publication number 0 899 734, which was the subject of appeal case T 0622/02. The application was refused in a decision of the examining division announced at oral proceedings held on 5 March 2003. Written reasons were dispatched on 17 March 2003.

The application was held not to meet the requirements of Article 83 EPC, i.e. not to disclose the invention in a manner sufficiently clear and complete for it to be carried out by a person skilled in the art.

II. The applicant filed a notice of appeal and paid the appropriate fee with a letter dated 15 and received 19 May 2003. The grounds of appeal were submitted on 17 July 2003. The grounds referred to a statement from the inventor which had been filed during examination and to submissions made during examination of the parent application, which were annexed. Reference was also made to the arguments put forward in the appeal relating to the parent application. This appeal had been heard, and finally dismissed, by the board in the same composition.

III. In response to the communication accompanying the summons to oral proceedings the appellant filed a new set of claims and two documents:

D4: US 5 602 857 A

D5: US 5 996 105 A

In the course of the appeal proceedings the following document originally cited during examination of the parent application was introduced by the appellant in its submission on 9 May 2006:

D3: M.A. Hasan et al, "Algorithms and Architectures for the Design of a VLSI Reed-Solomon Codec," in S. Wicker ed., "Reed-Solomon Codes and their Applications," 1994, pages 60-107.

IV. At the oral proceedings the appellant requested that the decision under appeal be set aside and a patent granted on the basis of claims 1 to 4 submitted on 9 October 2006.

V. The single independent claim 1 reads as follows:

"A data processor apparatus for common use in a DVD (Digital Video Disk) and a CD (Compact Disk) apparatus using mode setting information provided according to whether a disk is a DVD or a CD, comprising:

a PLL for generating a clock from a pulse stream read from a disk and a demodulator for performing demodulation according to the mode setting information;

a memory unit for storing data demodulated by the demodulator in a corresponding format of EFM + for DVD and EFM for CD according to the mode setting information; and

an ECC decoder for reading and correcting the data stored in the memory unit according to the mode setting information which is stored in a mode setting unit."

VI. At the end of the oral proceedings the board announced its decision.

Reasons for the Decision

1. The invention

1.1 The application describes as prior art a combined DVD/CD player in which the data processing is implemented separately for each mode (e.g. Fig. 2). It proposes a combined data processor which can deal with both modes (Paragraph [0009] of the published application). As part of this combined data processor, it proposes a common ECC (error checking and correction) decoder (the term "demodulator" is also used in the application as a synonym) - see e.g. Paragraph [0031] and Fig. 3. This ECC decoder is included as one of the features specified in present claim 1. Hence the question whether it is sufficiently described is potentially decisive for the appeal.

2. Res judicata

2.1 Arguably, this question has already been decided by the board in T 0622/02 in the same ex parte context and it might therefore be considered to be res judicata, in which case the board would be bound by its previous decision in this respect, even if presented with new evidence.

2.2 The issue of whether a question decided with respect to a parent application can be reconsidered for a divisional of that parent could in some circumstances be critical to the outcome of an appeal. However, on the facts of the present case and in the light of the conclusions arrived at below this is not true of the present appeal and so need not be decided.

3. Sufficiency of the description of the ECC decoder

3.1 The application points out that both CDs and DVDs use Reed Solomon (RS) codes for ECC (Paragraphs [0006] and [0007]) and that the codes have the same primitive polynomial (Paragraphs [0034] and [0035]) which as the skilled person would understand follows from the fact that in both cases the basic symbol length is taken to be eight bits, so that in both cases the RS coding is generated using the Galois Field over 256 elements, GF(2**(8)). It states further that, since they share this polynomial, "merely, the code lengths and correction ranges of the DVD and CD data to be error-corrected are different from each other. Therefore, by simply controlling the code length and the correction range of the input data according to the set mode, it is possible to correct errors of the DVD and CD data with use of the single ECC decoder 62." A block diagram of modules making up the ECC decoder is shown in Fig. 6, and the functions carried out by these modules on the data are described in Paragraph [0033].

3.2 The skilled person would undoubtedly be aware that, as stated in the application at Paragraph [0033], DVDs use two RS codes called PI, which is a (182, 172) code, i.e. a block of 172 symbols is provided with 10 ECC symbols, and PO, which is a (208, 192) code. Equally, CDs use a (32, 28) ("C1") and a (28, 24) ("C2") code. It is therefore clear that a shared ECC decoder as claimed must be able to carry out decoding for more than one RS code. This is what is apparently meant by "controlling the code length and correction range according to the set mode" in the above cited passage of the application.

3.3 However, the application does not give any instruction how such a decoder might be constructed. It is merely stated that it is possible. The only details of the structure of the decoder given are in Fig. 6 and the accompanying text at Paragraphs [0032] and [0033], but these details simply relate to an apparently conventional device for ECC decoding a single code (the functions described correspond to what is called "algebraic decoding" in D3 - see page 81, line 25 to page 82, line 26). The only feature shown which relates to decoding more than one code is the "mode setting information" input, but there is no disclosure whatsoever of how the modules 66, 68, 72 and 74 are to be constructed so as each to be able to cope with more than one code.

3.4 Thus, if the provisions of Article 83 EPC are to be considered satisfied, the board must come to the conclusion that the implementation of a shared ECC decoder using algebraic decoding would have been within the reach of the skilled person using common general knowledge in the field.

3.5 The statements in Paragraphs [0034] and [0035], to the effect that the fact that the primitive polynomials for the codes are the same makes a shared ECC decoder possible, may be taken as a hint to the skilled person as to the direction to be taken in developing a shared decoder. Indeed, although the application does not explain further, the documents available to the board make it clear that decoding any RS code requires (many) multiplications which are carried out modulo a certain value, this value being directly derived from the primitive polynomial. It is therefore true that the multiplier units required for the DVD RS decoder (Fig. 2, 34) will have the same structure as those required for the CD CIRC decoder (48). The board is further willing to accept that this would be a matter of common general knowledge to the person skilled in the art. However, the application still leaves the skilled person completely in the dark as to how to make use of this fact. The various modules of the decoder evaluate different polynomials, as illustrated in Annex 1 of the statement of grounds of appeal, and these polynomials depend on the specific code. The fact that individual multiplications for the different codes in the present case require the same logic structure does not give the skilled person any indication how to share the higher-level structures which evaluate the polynomials.

3.6 Given such an evident and major gap in the disclosure of the claimed invention it was appropriate for the examining division (and later the board) to ask the appellant to explain how the skilled person would realise the missing details and why the skilled person would be able to supply them from common general knowledge of the field.

3.7 In response to the first question the appellant provided two explanations annexed to the grounds of appeal. These were originally supplied at different times during the examination of the parent application. The first, "Annex 1", includes: (1) a diagram of the structure of an ECC decoder of the same general form as Fig. 6 of the application, but nonetheless with apparently significant differences; and (2) a mathematical example of the various polynomials and algorithms shown in that diagram. It is explained with relation to the syndrome generator 66 of Fig. 6 that "cells" would be used and that, since the generator polynomials of the RS codes used have factors in common, varying numbers of the cells could be used for different RS codes. There is a further statement to the effect that the remaining blocks in the submitted diagram may be implemented in a similar way. In the alternative explanation given in Annex 2 of the grounds of appeal there is reference to an "error location polynomial", which is presumably "sigma(x)" on the diagram in Annex 1. Two generator polynomials are given, again in a form which shows their common factors, and there is reference not to "cells" but to "iterations". How the common factors of the generator polynomials influence the structure of the units which generate and apply the error location polynomial (the "Modified Euclidean Algorithm" unit and the "Chien search" unit) is not explained.

3.8 In response to the second question the appellant originally supplied, in a submission dated 9 and received 12 October 2002, a declaration by the inventor stating his belief that the skilled person would have been able to implement the invention on the basis of the information in the application. In the course of the appeal two documents, D4 and D5, were also submitted. It was argued that these, and their references to a number of co-pending and previous applications, illustrated that the field was very active and that the skilled person would be well acquainted with the techniques necessary for decoding RS codes. When asked to identify passages of direct relevance the appellant pointed to D4 Fig. 4 and the corresponding text at column 10 lines 37 to 45.

3.9 The question of how the skilled person would realise the missing details has not been answered adequately. The appellant has pointed to the fact that the generator polynomials of the different RS codes have common factors and asserted that in some unexplained way this would lead to a structure of the component units comprising cells (or "iterations") varying numbers of which would be used for the processing carried out. The structure of a cell has not been given, nor has the alleged relationship between the cells and the factors of the generator polynomial been explained. Critically the appellant has given no explanation of how cells should be employed in the evaluation and application of the polynomials actually mentioned in the application, namely the syndrome polynomial S(x), the Forney syndrome polynomial T(x), the erasure locator polynomial E(x), the errata locator polynomial W(x) and the errata evaluator polynomial Ë(x) (Paragraph [0033] of the published application). The only document available to the board which mentions cells is D3 (page 77 lines 25 to 36, page 78 Figure 5-6, page 93 lines 32 to 35 and page 100 lines 8 to 13). It identifies two implementations where cells are used, but in neither does the hardware structure reflect in any way the factors of the generator polynomial pointed to by the appellant. In fact the cell structure given in D3, Figure 5-6, page 78, is directly derived from a completely different formulation of the generator and syndrome polynomials, see page 77 equation (33) and page 93 equation (85). The second implementation is a "time domain decoder" (D3, page 100, lines 8 to 17), whereas the structure shown in Fig. 6 of the present application is that of an "algebraic decoder".

3.10 Since the appellant has not given an adequate explanation of how the skilled person would fill in the evident large gaps in the disclosure of the application, it is clear that the second step of establishing that the requisite information was common knowledge in the field also has not been satisfied. However, the board will discuss the evidence that has been provided.

3.11 The board does not consider the inventor's statement to be persuasive as to what the skilled person would have known. Primarily it consists of a declaration that the inventor believes that there is sufficient information in the description and drawings for a person skilled in the relevant technological field to build a combined DVD/CD data processor as described in the specification. Passages of the description are cited and there is a further reference to an implementation using "cells". In this the statement merely repeats arguments that have already been put forward during examination of the parent application and are discussed above. The board does not consider that the inventor's belief that the invention is sufficiently disclosed has any evidential value. In the first place the inventor, an employee of the appellant, is clearly an interested party, and would therefore find it difficult to look at the disclosure of the application in a non-partisan way. In the second place, the inventor, having been intimately acquainted with the invention over a long time, is not in a position to put himself in the position of a skilled person whose only knowledge of the invention must be derived from the application (which may of course be interpreted in the light of his background knowledge).

3.12 Documents D4 and D5 do not appear to have any evidential value in establishing what was common knowledge in the field at the priority date. The appellant stated that they were introduced to show that the field was very active and that the skilled person would be well acquainted with the techniques necessary for decoding RS codes. For this purpose they are unnecessary; document D3, which comes from a book entitled "Reed-Solomon Codes and their Applications" clearly demonstrates that the field was of considerable interest and that the skilled person would have known a number of decoding techniques. Beyond that these documents do not appear to provide any evidence. They belong to a single family of patents and patent applications; according to long-standing case law the disclosure of a patent is only in exceptional circumstances to be considered part of common general knowledge. No such exceptional circumstances have been argued or are evident in this case. Document D5 does not even qualify as prior art; indeed its filing date (14 November 1997) is after the present priority date (30 August 1997). Thus the contents of D4 and D5 are not to be considered to be common general knowledge. Neither does the specific material pointed out to the board (D4 Fig. 4 and the corresponding text at column 10 lines 37 to 45) make any contribution to answering the question of what was common knowledge; it merely discusses part of the structure of an ECC/syndrome generator subsystem, without there being any indication that this is considered to be common knowledge.

3.13 The appellant has argued that evidence that there is sufficient information in the application as filed has been provided, whereas the examining division provided no evidence to support its view that the person skilled in the art would need more information. The board does not agree. Firstly the only evidence as to the state of common knowledge to which the board attaches any weight, document D3, was in fact originally introduced by the examining division. It tends to support the view that the skilled person would not have known how to implement the ECC decoder sketched out in the application. The several different possible architectures of an RS decoder put forward and the very lack of mention of the cell structure which the appellant suggests would have been the skilled person's choice make the appellant's assertion less credible. Secondly the board has explained above why it considers it reasonable in the present case for the examining division and the board to ask for supporting arguments and/or evidence. Given the clear lack of disclosure in the application itself it was incumbent on the appellant to demonstrate that the skilled person could have filled the gaps; this it failed to do. The board is not required to provide evidence of what the person skilled in the art did not know.

3.14 The board concludes that the application does not disclose the invention in a manner sufficiently clear and complete for it to be carried out by a person skilled in the art, in violation of Article 83 EPC. The appeal must therefore be dismissed.

ORDER

For these reasons it is decided that:

The appeal is dismissed.