The simplest solution is to abolish patents, the government has no business going around handing out monopolies.
Curiosity, love, boredom, responsibility, etc... There are a lot of incentives other than profit and billions already gets spent in R&D for reasons other than profit.
But if profit is your motive, and patents don't exist, you have several options. One of the interesting ones is that of monetising expertise. People will copy you, but you are also known globally as the expert, for the simple reason that you can prove you developed it. So you not only have a market made of the people who will buy the stuff, but also of those who want to learn off the expert how to build the stuff. To make the most of this requires good marketing, and an acceptance that you should be developing your next thing for when the money slows down from the current one as the market saturates, but that is the same as today anyway.
That is also the classic case for patent abuse. Take a drug that works, tack on some non-active atoms somewhere, change the name, patent it, jack up the prices on the old drug when the patent is about to expire to push buyers to your new monopoly.
(Customers usually don't care about cost, which is certainly a problem for holding down prices, but this doesn't seem anything like "patent abuse.")
I assume this is the new trade name for a variation on Viagra.
JRR did not patent 'the application of height-challenged individuals manually transported to a heated device for the purpose of destroying small amounts of jewelery', but the work is copyrighted so you cannot just find/replace the story so it reads Shrodo and Hamwise going to Mt Foomp with their trusty sidekick Bandalf.
To take a quote from your source: "Every writer owes a similar debt to those who have come before." - this is something I believe speaks more to the core of software development than the patent system we use today.
Terry Brooks' supposed imitation of The Lord of The Rings didn't stop it from generating hundreds of millions of dollars in revenue.
You have drugs at one end of the spectrum and software and business processes at the other. There is a huge cost associated with medical research due to the inherit complexity of the human body and the requirements governments put on proper testing. Any patent solution has to address that problem as well.
(And, yes, drug companies are doing a horrible job at doing real research, but that doesn't alter the substance of the argument.)
And yeah, that is the level on innovation in software patents.