I would have liked to see the functionality provided years ago by the script Annotation Pane and, recently, Stack integrated into core DT.
DT3 Inspectors > Document > Annotations shows highlights and notes to highlights as links to the PDF pages. This welcomed feature still leaves that information within each pdf, unreachable as a whole in a DT database-wide search.
Another new feature in DT3, Inspectors > Annotations is also welcomed. But it doesn’t address the functionality provided by the above scripts because, if used for long pdfs, it would drive away from the concept of chunking for the storing and searching information. Although this DT3’s new feature is good for adding a short annotation about another short annotation, to store a long annotation about (let’s say, a book in pdf format) would be counter-productive.
In summary, most of my input of text into DT is done for pdfs via the one of the above scripts. I am very grateful to their developers for their effort, skill and generosity. But third-party scripts are not a future-proof way to store information that will remain consistently linked and accessible in the future. I long to see their functionality integrated into DT3 core.
How to better integrate their crucial functionality is another matter. I can think of notes in Inspectors > Document > Annotations > Details (that is, notes embedded in each pdf) being reproduced as separated markdown files, along with some metadata from the annotated file. But there might be more satisfactory implementations.