I think the author is conflating "quality" of documentation with "usability" of documentation. They are probably correlated, and this might make a good filter for ignoring BAD documentation, but a documentation site can be very "usable" without containing quality content, and it can contain quality content without being "usable"