But since the article isn't really about parser bugs, I don't think using a different data format will save you from most of the problems described there.