Monday 16 May 2011

T 789/07 – Searching Deep


The outcome of the present appeal is likely not to please some of the Examiners among the readers of this blog. The appeal was directed against the decision of an Examining Division (ED) to refuse the application under consideration.

Claims 1, 4 and 8 before the ED were drafted as follows (English translation):
1. Method of protecting a microcomputer system (10) against manipulation of data (2) stored in a memory assembly (27) of the microcomputer system (10), in particular for protecting a program stored in the memory assembly (27), characterised in that the data (2) are transmitted to the memory assembly (27) in at least partially signed or encrypted form, obtained by means of an asymmetrical encryption method, when the microcomputer system (10) is newly programmed or reprogrammed.

4. The method according to any one of claims 1 to 3, characterised in that the signed or encrypted data (2,7) are checked or decrypted in the microcomputer system (10) by means of a freely accessible public key (12).

8. The method according to claim 4 or 5, characterised in that the checking of the signed or the decrypting of the encrypted data (2,7) is performed in the microcomputer system (10) when the data (2) are being processed.
Claim 1 of the main request before the Board read:
1. Method of protecting a microcomputer system (10) against manipulation of data (2) stored in a memory assembly (27) of the microcomputer system (10), wherein the data (2) are at least partially signed by means of an asymmetrical encryption method and transmitted to the memory assembly (27) when the microcomputer system (10) is newly programmed or reprogrammed, characterised in that the checking of the signed data (2,7) is performed in a microcomputer system (10) when the data (2) are being processed, wherein the checking is a process of lower priority, in that, each time the process is called, part of the [data] of the memory assembly (27) is checked, so that after a predefinable time period the whole content of the memory assembly (27) has been checked.
and claim 1 of the auxiliary request read:
1. Method of protecting a microcomputer system (10) against manipulation of data (2) stored in a memory assembly (27) of the microcomputer system (10), wherein the data (2) are at least partially signed by means of an asymmetrical encryption method and transmitted to the memory assembly (27) when the microcomputer system (10) is newly programmed or reprogrammed, characterised in that the checking of the signed data (2,7) is performed in a microcomputer system (10) when the data (2) are being processed, by checking part of the data of the memory assembly (27) whenever the data processing leaves sufficient computing time, so that after a predefinable time period the whole content of the memory assembly (27) has been checked.

*** Translation from the German ***

[2] The invention concerns a microcomputer system which checks newly loaded data by means of a cryptographic signature so as to protect itself against unauthorised manipulation of the data.

[2.1] The characterising part of claim 1 of both requests contains the feature
(a) that the checking of the signed data is carried out when the data are processed.
Claim 1 of the main and auxiliary request define this feature more precisely by features (b) and (c), respectively, in the following way:
(b) wherein the checking is a process of lower priority, in that, each time the process is called, part of the [data] of the memory assembly is checked, so that after a predefinable time period the whole content of the memory assembly has been checked, and

(c) by checking part of the data of the memory assembly whenever the data processing leaves sufficient computing time, so that after a predefinable time period the whole content of the memory assembly has been checked, respectively.
[2.2] The appellant argues that the features (b) and (c) merely make specific feature (a) of the original claim 8 on the basis of the description […]. Therefore, they are not to be objected to under R 86(4) EPC 1973.

[2.3] On the other hand, the decision […] explains that the original claims 1 to 9 on one hand and claim 1 according to both requests on the other hand solve different and technically unrelated problems with respect to the prior art. Based on a feature of the original claim 3, claims 1 to 9 are understood to solve the problem of reducing the amount of data to be signed, whereas claim 1 of both requests are held to solve the problem of keeping with certain time requirements by using priorities when checking [the data]. Therefore, none of the new claims 1 was unitary with the original claims 1 to 9 and, as a consequence, claim 1 of both requests was inadmissible under R 86(4) EPC 1973.

[2.4] Moreover, the decision argues as follows:
i) Original claim 8 was drafted in so broad a manner that it did not necessarily refer to keeping certain time requirements. Feature (a) was equivalent to the feature “processing during the operating time of the microcomputer system”. This feature was there to avoid data manipulation during the operating time, but was not related to priorities or time requirements […].
ii) Pages 8 to 9 of the description were contradictory with the original claims because the original claims still comprised the option to use encrypted data instead of signed data. The appellant had implicitly acknowledged this contradiction by limiting the claims to signed data […].
iii) The advantages linked to the new features (b) and (c) were not related in any obvious way to avoiding an unauthorised manipulation, which was the only problem mentioned in the description, and the Guidelines for examination (C-IV 9.8.2) did not allow a corresponding reformulation of the problem […].
[3] R 86(4) EPC 1973 lays down that amended claims may not relate to (1) unsearched subject-matter which (2) does not combine with the originally claimed invention or group of inventions to form a single general inventive concept.

[4] The Board is of the opinion that the claims of the main and auxiliary request as amended do not fulfil any of the two requirements which presuppose the application of R 86(4), but that they have to be considered both searched and combining with the originally claimed invention or group of inventions to form a single general inventive concept.

Extent of the search

[5] A 92 requires the European search report to be drawn up on the basis of the claims, with due regard to the description and any drawings.

This requirement is explained in the Guidelines for examination in the EPO B-III 35 (both in their present version and in the version in force at the time of search (June 2005)), in the following way, which the Board considers appropriate:
“In principle, and insofar as possible and reasonable, the search should cover the entire subject-matter to which the claims are directed or to which they might reasonably be expected to be directed after they have been amended.”

[5.1] The European search report of November 14, 2003, lists five documents and marks them as relevant for all claims 1 to 15. A limitation of the extent of search is not mentioned. Therefore, it has to be presumed that original claim 8 was indeed searched.

[5.2] The board is of the opinion that, as a rule, in order to be able to completely search a claimed feature within the meaning of A 92 – and in line with the Guidelines – the Search Division (SD) has to establish how this feature is to be interpreted in the light of the description and the drawings. It is only by proceeding in this way that the SD can come to a reasoned assessment of which amendments may be expected in the course of the examining proceedings, within the framework of the application as filed. For example, it is to be expected that an applicant clarifies unclear features within the framework of the disclosure and specifies in more detail functional or otherwise broad features. As far as original claim 8 is concerned, this is true in particular for feature (a).

Concerning argument i) of the decision

[6.1] In a fax received on March 11, 2005, the applicant had filed a claim 1 the characterising part of which establishes “that the signed or encrypted data are checked or decrypted and if the checking or decryption is successful, an identifier is stored in a predefinable memory area of the memory assembly, and a content of the memory area is checked during the operating time of the microcomputer system using the identifier”.

[6.2] There is no doubt that these features were disclosed in the original application (see, in particular, page 7 […]). However, according to them, what takes place “during the operating time of the microcomputer system” is not the checking of the signature or the decryption of the data but the “content of the memory area” into which an identifier or a checksum has been stored before, but after successful checking of the signature or decryption of the data, and in order to provide evidence for the latter, [is checked]. In this embodiment (page 7-8 […]), the checking of the signature or the decryption, as the case may be, itself takes place immediately “following a new programming or reprogramming of the memory assembly” (page 7 […]), i.e. precisely not during the data processing.

[6.3] Feature (a) also appears in almost identical wording on page 8 as the feature of another embodiment and is discussed in more detail (page 8-9 […]). In this context features (b) and (c) are disclosed as examples for how the “check is carried out in conjunction with the processing of the data” is to be understood in detail.

[6.4] Under these circumstances the Board considers it to be obvious that feature (a) is not to be interpreted in the light of the embodiment disclosed on pages 7-8 but of the one disclosed on pages 8-9.

[6.5] Accordingly, the Board is of the opinion that at the time of the search it was to reasonable to expect that the applicant could consider to clarify or specify in more detail the original claim 8 on the base of pages 8-9 in the course of the examination, and in particular via the features (b) or (c).

Concerning argument ii) of the decision

[7.1] The ED justifies the existence of the contradiction by pointing out that a decryption of the data necessarily had to take place before their processing and that, therefore, the decryption could not be carried out with lower priority than the data processing.

[7.2] The Board concedes that in principle data have to be decrypted before they can be processed. However, as a rule, the processing will start with a small fraction of the data. As soon as this fraction of the data is decrypted, the processing can begin, and the decryption of the remaining data may be expedited by means of a process of lower priority. It is true that even then it may happen that the processing has to wait for the decryption of data but in principle this does not conflict with concurrent processing, as claimed.

[7.3] In particular – in contrast to the decision of the ED – it is possible to interpret claim 1 in a way that does not render “the check ineffective” and does not deprive the claimed priorities of “any effect” […].

[7.4] The ED does not find a contradiction if the data are signed rather than encrypted. As, therefore, one of the options of original claim – and thus also of claims 8 – according to the embodiments disclosed on pages 8-9 is considered feasible, the Board is of the opinion that this embodiment should not have been ignored when carrying out the search, even if the argument of the ED had been correct. Experience shows that during the examination of an application a plurality of originally claimed alternatives is often limited to several of those alternatives.

Concerning argument iii) of the decision

[8.1] The Board does not see how the Guidelines C-IV 9.8.2 (in their version of June 2005, which was in force at the time of the search, nowadays 11.5.2) would prohibit a reformulation of the problem. Quite to the contrary, it (appropriately) states that “what the applicant presented as ‘the problem’ in his application … may require reformulation”, and that “as a matter of principle any effect provided by the invention may be used as a basis for the reformulation of the technical problem, as long as said effect is derivable from the application as filed”.

[8.2] Applicants often legitimately react to an objection of the ED by adding features taken from the description to the original claims in order to distinguish the latter from the cited prior art, in application of A 94(3). In the present case the Board does not see how the new features would introduce a completely new inventive concept which does not technically relate to the inventive concept behind the original claims. It is true that the new features (b) and (c) do not ensure protection against manipulations as such, but they ensure the protection in such a way that simultaneously time requirements can be met. The Board is of the opinion that thereby the original inventive concept is defined with greater precision in a way that is appropriate, admissible and predictable.

[9] Therefore, the Board comes to the conclusion that the search of original claim 8 within the meaning of A 92 would have had to take into account the embodiment disclosed on pages 8-9 and that, as a consequence, claim 1 according to both the main and the auxiliary request has to be considered as searched within the meaning of R 86(4) EPC 1973.

Unity

[10] According to the indications given in the European search report all original claims have been completely searched. There was no limitation due to a lack of unity (R 46(1) EPC 1973), nor was there a statement that the search had been carried out for all claims despite a lack of unity.

[10.1] It is true that the search report cites four of the documents as being very relevant (“X”) for all claims, but there has not been a concrete objection against the original claim 8 during the examination proceedings. The first notification of the ED contains a sweeping objection of lack of novelty and inventive step, respectively, but the following justification does not mention claim 8 or discuss its features […].

[10.2] The Board is unable to see that feature (a) is known from D1 or that it is prima facie obvious with respect to D1, at least if it is correctly interpreted in the light of the discussed disclosure of pages 8-9.

[10.3] Thus feature (a) represents a contribution of the claimed invention as a whole to the state of the art. Moreover, feature (a), when correctly interpreted, has to be ascribed the same effect as features (b) and (c), i.e. to ensure the protection against data manipulation in such a way that certain time requirements can be met.

[10.4] Therefore, claim 1 of both the main and auxiliary request certainly combines with original claim 8 – and therefore also with the originally claimed invention or group of inventions – to form a single general inventive concept. The fact that the originally claimed group of inventions possibly can be ascribed a further inventive concept (in respect of original claim 3) […] does not contradict this conclusion.

Summary

[11] As a consequence, the Board reaches the conclusion that the ED would have had to admit the present claims pursuant to R 86(4) EPC 1973. After entry into force of the EPC 2000, R 137(4) is to be applied in the version that was in force from December 13, 2007, to March 31, 2010, because the latter refers to A 123 (see J 10/07 [1.3]) However, as the wording of R 137(4) is identical with that of R 86(4), it directly follows that both the main and the auxiliary request are to be admitted under R 137(4).

Additional search

[12] The ED states that, as a matter of fact, the relevant features have not been searched […]. However, as they would have had to be searched, as explained above, a corresponding additional search is to be carried out when the examining proceedings are resumed, as appropriately and explicitly foreseen in such a case in the Guidelines B-II 4.2 i). […]

The application is remitted to the ED for continuation of the examining proceedings.

To read the whole decision (in German), click here.

The file wrapper can be found here.

5 comments:

Myshkin said...

I did not check all details, but it seems the ED (or rather SD) could have raised non-unity between claims 3 and 8 during the search stage.

What seems to have stopped the SD from doing that is the broad wording of the wording of claim 8. The additional feature of claim 8 is so general that it cannot be inventive, and can therefore not be a special technical feature.

In that case the lesson to be learned by the SD is that for determining whether a feature can serve as a special technical feature it is more important that the applicant has asserted that the feature is inventive (by including it in a dependent claim) than whether, in its current wording, the feature objectively is inventive. Indeed, the applicant might in a later stage further define the feature based on the description (as happened here). It will be too late then to raise non-unity (via present R. 137(5), first sentence), so non-unity should be raised at the search stage (if at all).

Manolis said...

I don't see why examiners will not like this decision! ;-)
It was to be expected! Application of Rule 137(4) (or 86(4)) has always been problematic in the case the extend of the search had not been documented, like when a non-unity was raised or an incomplete search was made for some other reason.
An ED declaring -a posteriori - that features of a certain part of the description were not searched, has always been risky!

Myshkin said...

"Application of Rule 137(4) (or 86(4)) has always been problematic in the case the extend of the search had not been documented, like when a non-unity was raised or an incomplete search was made for some other reason."

But these two cases where there is some documentation in the search report have little to do with R. 137(4).

In case of non-unity at the search stage and amendments directed to an originally claimed invention for which no additional fee was paid, it is R. 64(1) (or R. 164(1)) that applies via G 2/92, see T 708/00, point 7, 3rd paragraph.

In case of an incomplete search, e.g. because of claims that were deemed too unclear to search, and amendments directed to those claims (let's say with clarifications taken from the description), R. 137(4) does not apply since the amended claims are still unitary with the originally filed claims. (The second sentence of present R. 137(5) might apply though, until the EBA declares this sentence as void in view of Art. 123(1) and Art. 164(2).)

In my view, the correct criterion for determining whether features taken from the description are "unsearched" is: do the features relate to a concept that was not disclosed in the originally searched claims?

In the present case, the features related to a concept disclosed in originally filed claim 8 (when read in the light of the description), so R. 137(4) did not apply.

Imposing stricter conditions on "unsearched" would seem to reward applicants for not including features in the originally filed claims directed to concepts that could at later stage be decisive for inventive step (since R. 137(4) during examination would be stricter than R. 64(1) at the search stage).

(On the other hand, an applicant relying on a too narrow interpretation of R. 137(4) runs the risk of getting into problems with the time limit on filing divisionals.)

Manolis said...

In principle I do not disagree with your analysis. Especially in the case law there is extensive discussion about Rule 86(4) and 46 (EPC 1973). And it seems that Rule 86(4) was introduced exactly to cover the cases not covered by Rule 46 (again EPC 1973), i.e. after a non-unity objection.

My only objection is about the criterion to determine whether features were searched or not being whether these features share a common concept with the searched features. If it is so, then why are there two distinct criteria in Rule 137(4)?
If it has to be non-unitary in order to be unsearched, then why make the distinction in the Rule?

Myshkin said...

"If it has to be non-unitary in order to be unsearched, then why make the distinction in the Rule?"

This is indeed the weak point of my analysis. Or maybe it's the weak point of R. 137(4)'s wording ;-)

It seems that the Board in T 708/00 was sufficiently confused by "unsearched" that it gave up trying to understand the wording of R. 137(4) and instead determined the purpose of the rule:
"8. In view of the Rule 86(4) EPC limitation to prevent the applicant from avoiding the payment of search and examination fees for successively claimed non-unitary inventions, amended claims may only be refused on the basis of this rule if the subject-matter of the claims filed originally and that of the amended claims is such that had all the claims originally been filed together, a further search fee would have been payable - on top of the search fee payable in respect of the claims actually filed at the outset - in respect of the amended claims, relating to a different invention within the meaning of Rule 46(1) EPC."

This makes sense and does not require determining whether certain features are "searched" or "unsearched".

(Of course examiners might not be too happy with the way the Board in T 708/00 then continued with the non-unity analysis.)