The modularization paradox

Some assume that by requiring that embodiments of a patent have effects that depend on laws of nature (though excluding software as such) conceptual problems can be solved. This cannot be excluded per se though it may get paradoxical as follows:

• One may describe an invention as an application in technology rather than dealing with laws of nature.
• One may consider computer programs as software and one may also consider software as belonging to technology.
• Now consider computer programs $P$ and $Q$ where $Q$ provides a context within which $P$ may work. On the one hand $P+Q$ cannot be patented as it is 'software as such', on the other hand $P$ may be patented because of its role it may play in the context of $Q$ (which is a technical context given the above assumptions).

Taking this observation to the extreme: in a context where software as such cannot be patented and technical effects are required, one may be tempted to split a software invention into claimed components and stated components where the stated components are part of the justification of the claimed components. Interestingly this introduces a tendency to trivialize a patent description. More importantly, however, the whole state of affairs with $P$ and $P+Q$ is conceptually inconsistent. Therefore the dogma's that software as such cannot be patented and technical effects are required make sense only in a setting where one assumes beforehand that a collection of software components never represents a part of technology.