magcolo said:
Btw is bad link added by bots or manually? Its wiki says "Image links that can be converted to the page containing the image, such as Pixiv image links, don't count for this tag.", does that mean we don't always have to change its source to the post link? Or is it still preferred to do so? What about images that are found in source code but not actually on the page it's "supposed" to be from?
For some sources which Danbooru supports uploading from, bad link is indeed added automatically by bots (and as such they also remove it when it isn't valid), but outside of those cases, you can still manually add the tag. The bit in the wiki refers to the fact that some sources have source normalization - taking Pixiv as an example, you'll find that the vast majority of posts sourcing from it are in fact pximg.net/
, and not pixiv.net/
, because what would be the 'bad link' in this scenario has all the data necessary in the link itself for us to figure out what the original post is supposed to be. Source normalization represents the platonic ideal, where we have the image link for direct access to it but also have access to the post link, therefore not sacrificing accessibility to the original in any capacity.
It is still preferred for you to change the source to the post link though, because most sources on Danbooru are not normalized, and they're just way more accessibility-friendly. Edge cases like the one you mentioned on images hidden in source code are difficult, because one could make an argument that the source should be the image link and not count for bad link in the same fashion that we don't count it for some older styles of Japanese blogs, where they'll only have a gallery page for all images and clicking on an image takes you to the image link and not a post, but others would say that's silly and that the tag counts, or that it should be changed to the page whose source code it was taken from.