This is going to be a weird question. I am an academic researcher. In my field there has been a key topic whose "accepted" solution has been known not to work since 1962. Nobody had been able to work out why. I did. There was a subtle but fundamental math error in the original scientific work. In fact, the accepted solution can be shown to be completely uncorrelated with the actual solution. I replaced the solution with a mathematically valid solution and empirically tested against the entire historical data set the field has. It works.
My problem is that there are 651 software and process patents built upon the invalid methodology. I am concerned the patents will interfere with correct scientific development in the field. Are patent claims valid if they are built on something which is mathematically or scientifically impossible? The math error was that it violated the rules of general summation. The accepted process is wrong in the same sense that 2+2=77 is wrong. These are heavily used processes. They don't work, but nobody had been able to figure out a better solution.
My second question is "is there an easy way to get these patents declared invalid without bearing tremendous expense?"