You have a point but publication date, specially in this fast moving field, is a low grade data point. As another commenter wrote in this thread, we're reduced

You have a point but publication date, specially in this fast moving field, is a low grade data point. As another commenter wrote in this thread, we're reduced to deducing date range by looking at referenced prior work. I for one am grateful to Cornell for making ArXiv happen. They stamp the date on the preprint and their url enodes the general date [/<yy><mm>.<sernum>] as well: https://arxiv.org/pdf/1701.06020.pdf https://arxiv.org/pdf/1701.06020.pdf