let it leave me like a long breath

oh just like, when somebody's running a webserver, that webserver generally generates access logs -- who's looking at what. this records time of access, file requested, ip of request, stuff like that. but http requests also come with another field, the referer, that tells the server where it's requesting the thing from.

so every time you click a link, the request (generally) comes with the page you're on attached. and when your browser requests images, the same thing happens, with the referer being the page where the image is embedded. sites use this for stuff like search engine optimization, or access control (to forbid hotlinking or whatever), and so on.

so if somebody's checking their reading list through, like, http://user.dreamwidth.org/read/people_i_hate, and somebody in that filter posts an image, the server that it's hosted on will get a hit in their access logs like 'such-and-such ip requested someimage.png at so-and-so time, and the request came from http://user.dreamwidth.org/read/people_i_hate'. that's not an issue if the server is like, imgur or whatever, but if somebody's posting an image from a server they control, they could see that logfile hit and get real mad.