Disclaimer: I haven't read the whole article, so maybe it was mentioned there. My idea: only allowing reviews from (primary) authors of already-published, peer

Disclaimer: I haven't read the whole article, so maybe it was mentioned there. My idea: only allowing reviews from (primary) authors of already-published, peer-reviewed journals? Bootstrapping wouldn't be a problem, because there are already quite many papers published, many of them containing the email address of the authors. It would make it much harder to set up fake identities for peer reviews.