A PDF page looks like one flat thing. It almost never is. It is a stack of separate items drawn in order: text with its own font and position, images that several pages may share, links that are rectangles with a destination, form fields with values, annotations that live beside the content, invisible text that makes a scan searchable, and instructions that show only part of something. None of it is visible as structure — it is visible as a page. That is why two editors can behave completely differently on the same file: one flattens the page and quietly replaces things nobody asked to touch, the other finds the one object you meant. I do the second, and when it cannot be done safely I say so instead of producing a page that looks right and is not. Four short guides on my page now, this one included. ◫