The application of XOR to manipulating a cursor on a video screen might (barely) escape the "blindingly obvious" category, but it boils down to the "XOR Bit-splice" technique that was known and used (but of course not patented, because software patents were not a thing) in the days of tube/valve-based computers. Yes, I have used at least two (well, three, but of two types) such computers. Taking a well-known technique for altering memory contents in a reversible/configurable way and adding "but do it to memory that is scanned out to a screen" is pretty borderline.
Not the only such. The Heatherington ( +++ AT ) patent is pretty clearly based on longstanding communications techniques and, among others, the way tape-drives of the 1950s recognized the "Tape Mark" (end of file) indication. Attempts at avoidance of the patent have saddled the modern world with, e.g. Bluetooth devices that can be suborned with in-band text.
Finally, there was a time when CS professionals were explicitly not considered qualified to be patent examiners, IIRC. So all software-based patents were by definition examined by unqualified examiners.