They're ignoring the truncation problem. If I tell someone to look at the https://w3.org/something/6287297 URL, is that really correct, or did I botch a copy/paste of the https://w3.org/something/62872975 URL? I understand that https://w3.org/something/62872975.xml isn't optimal because XML might die. Is there a future-proof AND truncation-resistant approach, like always using https://w3.org/something/62872975.end regardless of the content type?