Liam Proven (the article author) is saying that Linus is forcing parsers to be more robust - i.e. forcing parsers to obey Jon Postel's robustness principle (which is how Jon titles it in
Section 2.10 of TCP's RFC 761; why use eponymy when descriptiveness is the same number of words).
I've noticed Liam likes writing in a way that someone with an English degree can argue is technically not incorrect, even if the audience has to read it three times to get what they are saying.