Everything so far has been about what a page looks like. PDF also defines things a reader can click, fill in and follow, all of it in dictionaries outside the content streams: annotations sit on top of a page, actions say what happens when something is triggered, and interactive forms tie annotations to field values.
/Annots: Link, Text (sticky note), Highlight, FreeText, Widget (form fields), FileAttachment… Each has a /Rect and ideally an appearance stream /AP (a form XObject) so it renders identically everywhere.
GoTo, URI, Launch, JavaScript, SubmitForm, Named… triggered by links, outlines, or events (/OpenAction, /AA additional actions).
Forms (AcroForm)
/AcroForm /Fields is a tree of fields (/FT = Tx Btn Ch Sig). A terminal field is usually merged with its widget annotation. /NeedAppearances asks the reader to regenerate appearances. XFA (1.5+) is a separate XML form model, deprecated in 2.0.
Outlines (bookmarks) and destinations are on The document tree page; signature fields are on Encryption and signatures.