It is not really a workable solution to flip around in a 250-page book hunting for a passage when the page numbers are off by 20-30 pages. That’s how it is when you’ve got 20-30 pages of introductory material — fairly typical in the PDFs that I use.
And with all due respect, I think you have simply misunderstood my message.
The whole point is that the correct page numbering is already available inside the PDF file. There is no need to “guess” anything. MarginNote simply needs to parse the information that is already there in the file.
Adobe Acrobat, PDF Expert, Skim, Bookends, etc. — they all handle this correctly.
MarginNote doesn’t.
EDIT: This PDF functionality is referred to as “Page Labels”; for details, see “8.3.1 Page Labels” in Adobe’s PDF Reference, v. 1.7.
As for the issue of exporting citations, we have been asking about this for several years now, and the MarginNote team doesn’t seem to think it’s any sort of priority. Evidently, the “solution” offered is that we’re supposed to write our own code to export citations using the undocumented Plugin API.
If you want export, you’ll have to code it yourself.