>But for a patent to be granted, a competent practitioner in the field should be able to assemble the device based on the supplied information. that's the essence of a patent: the mechanism is revealed in exchange for rights over it. So if a patent has been granted for an antigravity machine, the examiner must have believed it would be possible to build one. Or at least to build whatever is described in the patent, which may not function perfectly as an antigravity machine ;-)
Really, the patent doesn't prove anything. The article you quoted mentions some of the problems. Specifically, the patent officers may be too busy to do a detailed investigation.
Difference in opinions hath cost many millions of lives: for instance, whether flesh be bread, or bread be flesh; whether whistling be a vice or a virtue; whether it be better to kiss a post, or throw it into the fire... (from Gulliver's Travels)