I'm not positive about this new criterion. My gut feeling is that it should be a bit nuanced on either side. For example, even if Tony doesn't write a single line for the paper, I still think he should be an author. Conversely, I don't think someone who has not thus far contributed a single line of code, issue, or tutorial should be able to jump in, write a few paragraphs, and be an author on the paper.
Having said that, I do agree that "core contributors" is too nebulous, so, for example, I'd be perfectly comfortable having François on the author list.
Juan.