In Europe this is solved by not allowing patenting of mathematics or raw algorithms but allowing patents that cover the technical application of such things.
In certain fields the "technical effect" required [it's not called that any more, the test is more nuanced now] is, shall we say, less obvious than others. [Digital] Image processing/codecs/compression for example always seemed very close to just maths.
It's the same way in the US, one doesn't patent the mathematics or algorithm, but the application of an algorithm in a particular field. Discussions on patents are usually so low-information that important details like this are elided. However, I can get behind the point that if your worldview is limited to just software, this can be distinction without a difference.
In Europe, they tried to solve it by saying inventions which are computer implemented and only solve a business problem instead of a technical problem are considered unpatentable[1].
However, what you are saying is the same in the US its why patent applications say "A method for performing SOMETHING by REASON"
In certain fields the "technical effect" required [it's not called that any more, the test is more nuanced now] is, shall we say, less obvious than others. [Digital] Image processing/codecs/compression for example always seemed very close to just maths.