"The library particularly hopes that viewers will help fill out the scanty metadata attached to each image by adding learned comments and tags." . . . .