If it were only about whether or not something could be "reduced to maths" and fundamental sciences then just about nothing would be patentable. One thing to keep in mind is that back when PTOs were invented, the cost of intellectual inventions was negligible and could be taught/replicated in a matter of minutes, maybe days. Today, those inventions are a trillion-dollar industry spanning years or sometimes decades of research and development.
While I am generally against software patents due to how much they have been abused for countless stupid things, breakthrough non-obvious algorithms (stuff that is well beyond what is considered "fundamental" at the time of invention) should still be patent-worty. Why? Because if they aren't, then there is a high probability that research will remain locked behind closed doors as an industrial secret rather than get published for all to see and possibly get inspiration from for something else. That's assuming the prospect of losing your invention due to non-patentability does not discourage you from researching it in the first place. Reducing the risk of not getting rightful compensation for your inventions is the fundamental reason why patents were created in the first place.
With patents, you can try licensing new and existing technologies from the moment they are published or worst case, wait 20 years for the patents to expire instead of re-inventing/reverse-engineering them and still risk getting sued for espionage, plagiarism and any other backup infringement claims. With closed-door research, there is also no guarantee any of it will ever make it out the door.
As for historic names behind most of today's fundamental maths and sciences, a large chunk of them were bankrolled by nobles, kings, aristocrats, etc. who were mainly interested in attaching their names to upcoming bright minds for fame on relatively modest budgets if their family wasn't rich/noble to start with. Not at all the same thing as today's tens/hundreds of millions dollar research by private or publicly traded companies that need to get a reasonable return on that research to justify it and stay in business - very few people (outside academia) and businesses today can afford to bankroll major research efforts for fame only the way it used to be.
Junk/trivial/frivolous software patents may have tarnished the concept of software patents beyond repair and that would be a bad thing for people and companies who have genuine software/algorithmic inventions.
In any case, PTOs need to clarify and raise the bar on the cut-off of what is and isn't patentable.