It is useful in many ways to predefine metadata types and sub-types with option lists. Scrivener allows the user to define a metadata type (say “POV”) and then restrict the kinds of within that type (say “Antagonist”, “Protagonist”, “Neutral Observer”). Great! Now future uses of that metadata tag are standardized and easier to search on and build search collections. But such affordances are currently limited to the granularity of binder object (document, folder, etc). For strings within a document, the user can write notes or footnotes, but can not apply formal metadata types and their nested kinds. There is no reason (beyond arbitrary assignment) that metadata type or kind should be chained to arbitrary content granularity chunks. Same is true of keywords and the other apparently arbitrary categories of metadata. Why should my POV metadata assignments be restricted to the document granularity when I might want to assign them to strings or lines or paragraphs or words or in the future, with project library support, to entire projects or project collections? Same for bookmarks and footnotes and notes and keywords. Why are these metadata types arbitrarily bound to equally arbitrary content granularity designations? Why can’t I apply a keyword or custom metadata to a string or other chunk within a document. Why can’t I assign a note to a document or folder or entire project? Same for the arbitrry metadata categories, “Lable” and “Status”. Why are they treated differently from other metadata and why can’t I apply them to chunks types smaller and larger than binder objects?