In patent law terms, the existence of a combination of features, i.e. of a combination invention, is to be viewed differently from the mere existence of partial problems, i.e. of an aggregation of features. According to current case law, partial problems exist if the features or sets of features of a claim are a mere aggregation of these features or sets of features (juxtaposition or collocation) which are not functionally interdependent, i.e. do not mutually influence each other to achieve a technical success over and above the sum of their respective individual effects, in contrast to what is assumed in the case of a combination of features. What has to be established is whether each set of features is separately obvious in the light of the prior art (T 389/86, OJ 1988, 87; T 387/87, T 294/90, T 363/94). It should also be borne in mind that solutions to partial problems in differing technical fields must be assessed on the basis of the knowledge and expertise of the person skilled in the art where the solution is found (T 32/81, OJ 1982, 225; T 324/94).
In T 389/86 (OJ 1988, 87) the relationship between the two groups of features was not one of functional reciprocity. The board ruled that in such circumstances no combinative effect could be advanced in support of inventive step; rather the question was whether each group, taken singly, was obviously derivable from the prior art. For the subject-matter of the claim to be inventive, it sufficed if one of these groups was (T 345/90, T 701/91, T 94/05, T 450/06, T 102/08, T 619/08, T 2097/10).
Similarly, in T 130/89 (OJ 1991, 514) the technical problem intended to be solved by the claimed invention also consisted of two technically independent partial problems, each solved independently by one of the claimed subject-matter's features. The board held that the independence of the claimed subject-matter's features (each producing a different effect) meant that in assessing inventive step the two closest states of the art had to be considered to enable each of the two partial problems to be defined. It concluded that since each of the partial problems was solved by means which merely performed their known functions, each partial solution was obvious and the invention thus lacked inventiveness. In T 597/93 the board again saw no inventive step in combining the claim's two features - both known per se - since they related to the solving of two entirely separate partial problems. It cited T 687/94 which held that in such cases the solutions could be assessed separately against the prior art (see also T 315/88, T 65/90, T 2110/08).
In T 711/96 the board found that characterising features (a) and (b) functioned completely independently of each other; there was no functional interplay (combination) between them. Although the setting for one value (e.g. spread) could indirectly affect that for the other (e.g. quantity), in that spread and quantity could both be adjusted upwards to maintain constant distribution, the two features were not directly related. In other words, the characterising features did not necessarily influence each other, although they could do. The board therefore assessed the inventive step of the two features separately, and concluded that both partial problems were obvious (see also T 1585/07).
In T 410/91 the board of appeal stated that no inventive step was involved since, although all the measures in claim 1 contributed to an increase in the efficiency of the plant, that contribution was based on known, different individual effects which resulted in these measures being executed in a manner expected by the skilled person. The subject-matter of claim 1 therefore involved the stringing-together of known measures which displayed their characteristic effects; no synergistic effect based on a combination of the individual measures was discernible in the sense of a mutual influence on their respective operation (see also T 144/85, T 141/87, T 407/91, T 1277/01).
In T 204/06 the board recalled that the "could-would approach" involves asking whether the skilled person would have - as opposed to could have - taken a certain step towards the invention in expectation of some improvement or advantage (T 2/83, OJ 1984, 265). This approach should not be taken to mean that inventions involving known design choices are non-obvious if only the number of choices is sufficiently great. It does imply, however, that if the skilled person expects some advantage of each feature in a claim and obtains no more than this advantage, then the claimed feature combination is obvious. It follows that any combination of features having known advantages (and disadvantages) is obvious unless it provides an unexpected effect.