After a proposal by @montyly, it may be a good idea to allow for peer-reviewed, published papers to provide 2 links, one for the published and one for the latest version. This way the evolution of papers after publication (e.g. bitcoin backbone) will not be lost.