#1: The protection against data that would require "disproportionate effort" seems like a valid protection of the company. A company the size of Facebook surely has very complex databases, logs, and caches going on, so that if they had to literally fetch all the data they have about a user, I could see this becoming quite a project.
#2: I would consider data models a trade secret. So if the personal data contains information that belies the underlying structure, I'm okay with the data being "scrubbed" clean of anything that would potentially give competitors a peek into the underlying engineering.
Note that I canceled my Facebook account long ago because I think Facebook is a system that is structured so that the users' best interests are not a priority. But this doesn't mean that everything Facebook does is bad or is against users' privacy, so I find this article to be a bit sensationalist.