G3/08 on the Patentability of Computer Programs

Kluwer Patent Blog
May 20, 2010

Please refer to this post as: , ‘G3/08 on the Patentability of Computer Programs’, Kluwer Patent Blog, May 20 2010, http://patentblog.kluweriplaw.com/2010/05/20/g308-on-the-patentability-of-computer-programs/


The “nail-Byting” in the Computer IP community came to an end on 12 May 2010 when the EPO’s Enlarged Board of Appeal finally handed down its opinion in the “Programs for Computers” case. The opinion is about 4 questions which were referred to the EBA by the President (Ms. Brimelow) under Article 112(1)(b).

The decision is sobering, reassuring and both surprising and unsurprising. It is sobering, because all questions were rejected as inadmissible and reassuring because it appears to merely spell out the existing law on the question of patentability of “computer implemented inventions” (CII) – the EPO’s preferred moniker. It is unsurprising because a good third of the large number of amicus briefs filed in this case had predicted exactly this outcome.
“G3 of 08” is also surprising because it is of at least as much interest to the general EPO practitioner as to the CII practitioner.

As to the points of general interest, the opinion makes clear that the President under Article 112(b) has an unfettered right to refer questions to the EBA on a topic as many times as he or she sees fit: The President can file one set of questions on the topic, turn around on his or her heel, and come back to file yet another one arguing the exact opposite and the EBA is still required to consider them. Such repeated questions considered cannot be rejected as inadmissible simply on this ground. This needed sorting out, because a previous President (Mr. Pompidou) was of the opinion that “Dear Lord Justice Jacob, … I believe there are insufficient differences between current Board of Appeal decisions dealing with Article 52 EPC exclusions” (Letter of February 2007). Lord Justice Jacob had thought otherwise. In fact, his Lordship figured quite prominently throughout the opinion, but more on this later.

For there to be an Article 112(1)(b) issue, there must be “two different decisions” by at least “two Technical Boards of Appeals”. We learn from the instant Opinion that “two Technical Boards of Appeals” should be understood to include morphed Boards: “a single Technical Board of Appeal although in different compositions”. We are also informed that “different decisions” are to be interpreted restrictively in the sense of “conflicting decisions”.

Further, law school students may find it reassuring to hear that it is the ratio that matters: “legal rulings are characterized not by their verdicts, but by their grounds”  although reassurance comes here at the expense of being extremely unsurprising.

Of note is that the Board draws a distinction between “different decisions” being mere “legal developments” and those being “in conflict” to each other. Only the latter can admissibly attract a referral. This point appears interesting because it may make it harder in the future to get admissible questions across. Yet another EPO bash at “raising the bar”, this time for “admissibility”, the EBA giving the Administrative Council a break?

What is it then that the CII practitioner can take home from G3/08?

First, the EBA did not like the questions. Large parts of the Opinion that runs to a formidable 60 pages are taken up by “parsing” the referred questions. The EBA looked at the questions and the supporting referral through Alice’s “looking glass” and found that there surely must be something wrong with the conclusion that “tigers eat words”. Mindful of these sorts of semantic and syllogistic pitfalls, the Board straightens the logical distinction between a method carried out by a computer and the list of instructions (=”the program”) which specify that method.

Question 1 was construed to be actually asking whether, given that a claim to a “computer program comprising instructions to carry out steps x,y,z” is excluded subject matter, are then claims to a “computer system loaded with the program” or a “method for operating a computer according to the program …” or “computer program product storing the program” automatically excluded as well.

The referral claimed to have identified two flagships of the EPO’s case law on CII that diverge: T113/97 (“IBM”) and T424/03 (“Microsoft”). The EBA found no conflict here because IBM’s “it makes no difference whether the computer program is claimed by itself or the record carrier” has not been followed by any Board, and, the later “Microsoft” decision which held that “the program on a computer readable storage medium is necessarily not excluded” is in fact a consequence of the ratio in IBM. The controversial ”it makes no difference “ limb of the decision was simply a dead end; according to the EBA, a clear-cut case of “legal development” rather than “conflict”.

As a matter of courtesy, the Board answered questions 1 and 2 by summing up (see 10.13) “the present position of the case law … [which is] thus that
a claim in the area of computer programs can avoid exclusion under Articles 52(2)(c) and (3) EPC merely by explicitly mentioning the use of a computer
or a computer-readable storage medium”. It immediately adds however that if the claim has nothing else to offer it will fail to be patentable for lack of inventive step under Article 56 EPC. It is merely “the article under which a claim is rejected which is different”. The EBA confirms this shift from Article 52(2)/(3) to rejections under Article 56 CII.

In fact, towards the end of the opinion the Board hails T154/04 (“Duns”) as a “practicable” role model for this shift to Article 56 CII rejections. T154/04 did what the instant EBA did to the referred questions: T154/04’s “feature crunching” pulverized Duns’ “method for estimating sales activities” so that there was nothing left but a “processor”. Hardly surprising, it was found to be not inventive. Because of the explicit approval of T154/04, CII practitioners would probably do well to check whether their claims would survive a T154/04 attack: Unless you can tie your assumed “non-technical” and technical features into a consistent bundle which will in combination produce a “further technical fact” (this is the other limb in IBM reasoning that was followed by other Boards), the claim is unlikely to the take the Article 56 hurdle.

Question 3 asked whether a(!) claimed feature must cause a technical effect on a physical entity in the real world in order to contribute to the technical character of the claim. Again, no conflict here because, for one thing, the Technical Boards, as matter of good law for long, do not assess the claims based on one single feature in isolation but as a whole – including a feature that in and of itself would clearly be excluded subject matter.

In support of question 3, the referral purportedly identified two decisions which required there to be a technical effect in the real world. The Board disagrees. The effects in the “BBC” and “Mitsubishi” case were essentially confined to the respective computer programs. For this reason alone, the question was rejected as not admissible. Further, the EBA repeatedly stressed and so confirmed current understanding of the law that the previous Boards’ “technical contribution” approach to CIIs has been laid to rest as it involves looking at the prior art and so confounds Article 52(2) / (3) with Article 56 issues.

Question 4 concerned whether the activity of programming a computer necessarily involves technical considerations. Again, no conflict in the body of existing case law, but, helpfully, the Board recognized that there is at least a “potential for confusion” and went on to state that it is not true that any technical consideration is sufficient to confer technical character on the claimed subject-matter. The Board again goes back to IBM’s “further technical effects”: the programmer “must have had technical considerations beyond merely finding a computer algorithm to carry out some procedure”. Taking the IBM point further, “the technical character of the claimed subject-matter is guaranteed only if writing the program requires further technical consideration”. See 13.5.1.

Pervading the entire opinion is the Board’s regret that the really interesting question had not been asked: The merits of Lord Justice Jacobs’ challenge of the “intellectual honesty” of the EPO’s shift rejections under Article 56 of CII claims as exercised in “Pension benefit” and later amongst others in T154/04. This still appears to bother the EPO, which is understandable because the argument is indeed powerful. It was presented in the CII case of Aerotel/Macrossan which was before the England and Wales Court of Appeal: How can you simply declare a feature found non-technical to be state of the art if this is plainly not the case? This is to “deem the new music [on a conventional CD player] … part of the prior art”. See LJ Jacobs’ discussion at 27 in the Aerotel case. The EPO countered in T154/04 that the Lord “missed the point” because the EPO is actually nice to the CII patent applicant: rather than completely discarding the non-technical features (“the new music”), the “music” ends up on the problem-to-be-solved side of the Problem-Solution Approach, the EPO’s inventive step detection algorithm. This may explain the EBA’s fondness to T154/04 which is “… a practicable system for delimiting the innovations for which [CII] patent[s] may be granted”. So, because of G03/08, the approach in T154/04 is here to stay for some time and is likely to be knighted by inclusion into the Guidelines, which, for sure, will see many revisions. So much for the certainty afforded by the opinion.

But there is of course substantial post-G03/08 uncertainty, too: The EBA openly shuns defining what is or is not “technical”. See 9.2. Failing that, the applicant is still left in a limbo as to exactly what claim wording the EPO examiner ends up examining: is it the claim as filed by the applicant or the Examiner’s T154/04 stripped down version of what once was the applicant’s claim.